update from asus 106

algoritma-batch/batch_video_cropper.py

@@ -0,0 +1,485 @@
+"""
+Batch Video Cropper — Rekam Video RTSP per Sesi Batch Truk
+
+Program ini membaca livestream CCTV (RTSP), menjalankan algoritma penentuan batch
+menggunakan YOLO + State Machine, dan menyimpan potongan video per-batch ke folder
+yang terorganisir berdasarkan tanggal.
+
+Struktur Output:
+    ~/reTraining/data/archive/
+    ├── 2026-08-05/
+    │   ├── batch_1_09-15-30.mp4
+    │   ├── batch_2_10-22-45.mp4
+    │   └── batch_3_14-08-12.mp4
+    └── 2026-08-06/
+        └── batch_1_07-30-00.mp4
+
+Menjalankan:
+    cd ~/reTraining/algoritma-batch
+    python3 batch_video_cropper.py
+"""
+
+import os
+import cv2
+import numpy as np
+import time
+import json
+import threading
+from datetime import datetime, timedelta
+from shapely.geometry import Point, Polygon
+from ultralytics import YOLO
+
+# Import modules
+from src.tracking import ByteTrackTracker
+from src.stabilizer import BboxStabilizer
+from src.truck_roi import TruckROI
+from src.counting import LineCrossCounter
+from src.batch import BatchLifecycleManager, BatchRecord
+
+# =====================================================================
+# 1. KONFIGURASI
+# =====================================================================
+
+import platform
+IS_WINDOWS = platform.system() == "Windows"
+
+# --- Path Konfigurasi ---
+BASE_DIR = os.path.dirname(os.path.abspath(__file__))
+ZONES_JSON = os.path.join(BASE_DIR, "zones.json")
+
+if IS_WINDOWS:
+    MODEL_PATH = os.path.join(BASE_DIR, "v1-best.pt")
+    ARCHIVE_BASE = os.path.join(BASE_DIR, "archive_output")
+else:
+    MODEL_PATH = os.path.join(BASE_DIR, "v1-best.pt")
+    ARCHIVE_BASE = os.path.expanduser("~/reTraining/data/archive")
+
+# --- Sumber Video RTSP ---
+RTSP_URL = "rtsp://frigate:zenai@192.168.192.209:8554/camera_stream_640"
+
+# --- Batas Pergantian Hari (Cutoff) ---
+DAILY_CUTOFF_TIME = "20:00"
+
+# --- Parameter State Machine ---
+SACK_IDLE_TIMEOUT = 5.0      # Jeda aktivitas sebelum masuk WAITING_FOR_ACTIVITY
+MIN_BATCH_DURATION = 2.0     # Durasi minimal batch sebelum boleh masuk WAITING
+TOLERANCE_LOW_COUNT = 0.0    # Instan (0s) — batch langsung berakhir saat area truk kosong
+TOLERANCE_MED_COUNT = 0.0
+TOLERANCE_HIGH_COUNT = 0.0
+
+# --- Video Recording ---
+VIDEO_FPS = 10.0             # FPS output video (10 fps sudah cukup untuk rekaman arsip)
+VIDEO_CODEC = "mp4v"         # Codec untuk .mp4
+
+
+# =====================================================================
+# 2. THREADED RTSP READER (Menghindari Lag Buffer)
+# =====================================================================
+class RTSPStreamReader:
+    """Threaded RTSP reader yang selalu mengambil frame terbaru."""
+
+    def __init__(self, source_url):
+        self.source_url = source_url
+        self.cap = cv2.VideoCapture(source_url)
+        self.frame = None
+        self.ret = False
+        self.running = True
+        self.lock = threading.Lock()
+        self.new_frame_event = threading.Event()
+        self.thread = threading.Thread(target=self._update, daemon=True)
+        self.thread.start()
+
+    def _update(self):
+        while self.running:
+            if not self.cap.isOpened():
+                print("[RTSP] Stream terputus, mencoba reconnect dalam 5 detik...")
+                time.sleep(5)
+                self.cap = cv2.VideoCapture(self.source_url)
+                continue
+            ret, frame = self.cap.read()
+            if not ret:
+                time.sleep(0.01)
+                continue
+            with self.lock:
+                self.ret = ret
+                self.frame = frame
+            self.new_frame_event.set()
+            time.sleep(0.001)
+
+    def read(self):
+        if self.new_frame_event.wait(timeout=2.0):
+            self.new_frame_event.clear()
+            with self.lock:
+                if self.frame is None:
+                    return False, None
+                return self.ret, self.frame.copy()
+        else:
+            with self.lock:
+                if self.frame is None:
+                    return False, None
+                return self.ret, self.frame.copy()
+
+    def isOpened(self):
+        return self.cap.isOpened()
+
+    def release(self):
+        self.running = False
+        if self.cap.isOpened():
+            self.cap.release()
+
+
+# =====================================================================
+# 3. FUNGSI UTILITAS TANGGAL & FOLDER
+# =====================================================================
+def get_counting_date(dt=None):
+    """Menentukan tanggal kerja berdasarkan cutoff harian."""
+    if dt is None:
+        dt = datetime.now()
+    try:
+        cutoff = datetime.strptime(DAILY_CUTOFF_TIME, "%H:%M").time()
+    except Exception:
+        cutoff = datetime.strptime("20:00", "%H:%M").time()
+
+    if cutoff.hour == 0 and cutoff.minute == 0:
+        return dt.date().isoformat()
+
+    if dt.time() < cutoff:
+        return dt.date().isoformat()
+    return (dt.date() + timedelta(days=1)).isoformat()
+
+
+def ensure_date_folder(counting_date):
+    """Membuat folder tanggal di archive jika belum ada. Mengembalikan path folder."""
+    folder = os.path.join(ARCHIVE_BASE, counting_date)
+    os.makedirs(folder, exist_ok=True)
+    return folder
+
+
+# =====================================================================
+# 4. BATCH VIDEO RECORDER (Mengelola VideoWriter per Batch)
+# =====================================================================
+class BatchVideoRecorder:
+    """Mengelola pembukaan dan penutupan file video per sesi batch."""
+
+    def __init__(self, archive_base, video_fps=10.0, codec="mp4v"):
+        self.archive_base = archive_base
+        self.video_fps = video_fps
+        self.codec = codec
+        self.writer = None
+        self.current_path = None
+        self.frame_count = 0
+
+    def start_recording(self, batch_number, counting_date, frame_width=1280, frame_height=720):
+        """Membuka file video baru untuk batch ini."""
+        self.stop_recording()  # Pastikan writer sebelumnya ditutup
+
+        folder = ensure_date_folder(counting_date)
+        timestamp_str = datetime.now().strftime("%H-%M-%S")
+        filename = f"batch_{batch_number}_{timestamp_str}.mp4"
+        self.current_path = os.path.join(folder, filename)
+
+        fourcc = cv2.VideoWriter_fourcc(*self.codec)
+        self.writer = cv2.VideoWriter(
+            self.current_path, fourcc, self.video_fps, (frame_width, frame_height)
+        )
+        self.frame_count = 0
+
+        if self.writer.isOpened():
+            print(f"[RECORD] Mulai merekam video batch #{batch_number} -> {self.current_path}")
+        else:
+            print(f"[RECORD ERROR] Gagal membuka VideoWriter untuk: {self.current_path}")
+            self.writer = None
+
+    def write_frame(self, frame):
+        """Menulis satu frame ke video aktif."""
+        if self.writer is not None and self.writer.isOpened():
+            self.writer.write(frame)
+            self.frame_count += 1
+
+    def stop_recording(self):
+        """Menutup file video yang sedang aktif."""
+        if self.writer is not None:
+            self.writer.release()
+            self.writer = None
+            if self.current_path and self.frame_count > 0:
+                print(f"[RECORD] Video selesai disimpan: {self.current_path} ({self.frame_count} frames)")
+            elif self.current_path and self.frame_count == 0:
+                # Hapus file kosong
+                try:
+                    os.remove(self.current_path)
+                    print(f"[RECORD] File video kosong dihapus: {self.current_path}")
+                except Exception:
+                    pass
+            self.current_path = None
+            self.frame_count = 0
+
+    @property
+    def is_recording(self):
+        return self.writer is not None and self.writer.isOpened()
+
+
+# =====================================================================
+# 5. FUNGSI UTAMA — LOOP UTAMA DETEKSI & PEREKAMAN
+# =====================================================================
+def run_batch_video_cropper():
+    print("=" * 60)
+    print("  BATCH VIDEO CROPPER — RTSP → Per-Batch MP4 Recorder")
+    print("=" * 60)
+    print(f"  Model     : {MODEL_PATH}")
+    print(f"  RTSP      : {RTSP_URL}")
+    print(f"  Archive   : {ARCHIVE_BASE}")
+    print(f"  Cutoff    : {DAILY_CUTOFF_TIME}")
+    print(f"  Tolerance : Instan (0s)")
+    print("=" * 60)
+
+    # Pastikan folder archive ada
+    os.makedirs(ARCHIVE_BASE, exist_ok=True)
+
+    # --- Load Model YOLO ---
+    print("[INFO] Memuat model YOLO...")
+    model = YOLO(MODEL_PATH)
+
+    # Warm-up model
+    print("[INFO] Warm-up model...")
+    dummy = np.zeros((720, 1280, 3), dtype=np.uint8)
+    device = "cuda" if os.path.exists("/usr/local/cuda") else "cpu"
+    try:
+        import torch
+        if torch.cuda.is_available():
+            device = "cuda"
+    except ImportError:
+        pass
+    _ = model(dummy, imgsz=640, device=device, verbose=False)
+    print(f"[INFO] Model siap. Device: {device}")
+
+    # --- Setup Components ---
+    tracker = ByteTrackTracker(model, conf=0.55)
+    stabilizer = BboxStabilizer(
+        ema_alpha=0.35,
+        max_hold_frames=10,
+        max_height_ratio=1.5,
+        min_height_ratio=0.70,
+    )
+
+    # Skala koordinat (kalibrasi 1920x1080 -> 1280x720)
+    scale_x = 1280.0 / 1920.0
+    scale_y = 720.0 / 1080.0
+
+    # Detection polygon
+    detection_poly_pts = [
+        [int(574 * scale_x), int(50 * scale_y)],
+        [int(586 * scale_x), int(1077 * scale_y)],
+        [int(1418 * scale_x), int(1076 * scale_y)],
+        [int(1397 * scale_x), int(50 * scale_y)],
+    ]
+    detection_polygon = Polygon(detection_poly_pts)
+
+    # Truck polygon (untuk monitoring kehadiran karung)
+    truck_poly_pts = [
+        [int(600 * scale_x), int(385 * scale_y)],
+        [int(609 * scale_x), int(1076 * scale_y)],
+        [int(1404 * scale_x), int(1078 * scale_y)],
+        [int(1381 * scale_x), int(343 * scale_y)],
+    ]
+    truck_polygon = Polygon(truck_poly_pts)
+
+    # Line crossing
+    static_line_y = int(330 * scale_y)
+    static_line_x_start = int(577 * scale_x)
+    static_line_x_end = int(1401 * scale_x)
+
+    static_roi = TruckROI(
+        x1=int(600 * scale_x),
+        y1=int(343 * scale_y),
+        x2=int(1404 * scale_x),
+        y2=int(1078 * scale_y),
+        line_y=static_line_y,
+        confidence=1.0,
+    )
+
+    counter = LineCrossCounter(
+        line_y=static_line_y,
+        line_x_start=static_line_x_start,
+        line_x_end=static_line_x_end,
+        margin=20,
+        dedup_radius=60.0,
+    )
+
+    batch_mgr = BatchLifecycleManager(
+        stabilize_seconds=0.0,
+        stabilize_threshold_px=9999.0,
+        sack_idle_timeout=SACK_IDLE_TIMEOUT,
+        min_batch_duration=MIN_BATCH_DURATION,
+        truck_gone_tolerance=TOLERANCE_LOW_COUNT,
+    )
+
+    recorder = BatchVideoRecorder(
+        archive_base=ARCHIVE_BASE,
+        video_fps=VIDEO_FPS,
+        codec=VIDEO_CODEC,
+    )
+
+    # --- Buka RTSP Stream ---
+    print(f"[INFO] Membuka RTSP stream: {RTSP_URL}")
+    cap = RTSPStreamReader(RTSP_URL)
+    if not cap.isOpened():
+        print(f"[ERROR] Gagal membuka RTSP stream: {RTSP_URL}")
+        return
+
+    # Tracking state
+    batch_counter = 0
+    frame_idx = 0
+    last_fps_time = time.time()
+    fps_counter = 0
+
+    print("\n[INFO] Memulai loop utama... Tekan Ctrl+C untuk berhenti.\n")
+
+    try:
+        while True:
+            ret, frame = cap.read()
+            if not ret or frame is None:
+                time.sleep(0.01)
+                continue
+
+            # Resize ke 1280x720 (sesuai kalibrasi koordinat zona)
+            frame = cv2.resize(frame, (1280, 720))
+            timestamp = time.time()
+            frame_idx += 1
+
+            # FPS counter
+            fps_counter += 1
+            if fps_counter % 100 == 0:
+                elapsed = time.time() - last_fps_time
+                fps = 100.0 / elapsed if elapsed > 0 else 0
+                last_fps_time = time.time()
+                state_str = batch_mgr.state
+                rec_str = "REC" if recorder.is_recording else "---"
+                print(f"[FPS] {fps:.1f} fps | State: {state_str} | {rec_str} | Frames: {frame_idx}")
+
+            # Simpan state sebelum update
+            prev_active = batch_mgr.is_active
+            prev_state = batch_mgr.state
+
+            # --- 1. YOLO Tracking ---
+            raw_tracked_all = tracker.update(frame, [])
+            raw_tracked_sacks = [d for d in raw_tracked_all if d.class_name == "sack"]
+
+            # --- 2. Stabilizer ---
+            stable = stabilizer.update(raw_tracked_sacks)
+
+            # --- 3. Filter Detection Polygon ---
+            stable = [
+                d for d in stable
+                if detection_polygon.contains(
+                    Point((d.bbox[0] + d.bbox[2]) / 2.0, (d.bbox[1] + d.bbox[3]) / 2.0)
+                )
+            ]
+
+            # --- 4. Hitung karung di 70% area bawah truk ---
+            min_ty, max_ty = truck_polygon.bounds[1], truck_polygon.bounds[3]
+            truck_height = max_ty - min_ty
+            truck_cutoff_y = min_ty + 0.30 * truck_height
+
+            sacks_in_truck_area = 0
+            for d in stable:
+                cx = (d.bbox[0] + d.bbox[2]) / 2.0
+                cy = (d.bbox[1] + d.bbox[3]) / 2.0
+                if truck_polygon.contains(Point(cx, cy)) and cy >= truck_cutoff_y:
+                    sacks_in_truck_area += 1
+
+            # --- 5. Line Crossing ---
+            tracked_in_roi = [
+                d for d in stable
+                if static_roi.contains_x((d.bbox[0] + d.bbox[2]) / 2.0)
+            ]
+            events = counter.update(tracked_in_roi)
+            has_crossing = len(events) > 0
+
+            # =============================================================
+            # LOGIKA ALGORITMA PENENTUAN BATCH (STATE MACHINE)
+            # =============================================================
+
+            # A. Mulai Batch
+            if batch_mgr.state in ("IDLE", "TRUCK_STABILIZING"):
+                batch_mgr.update_truck(has_crossing, (0.0, 0.0), timestamp)
+
+            # B. Monitoring Batch Aktif
+            if batch_mgr.state in ("COUNTING_SACKS", "WAITING_FOR_ACTIVITY"):
+                current_count = counter.loading_count
+                if current_count < 20:
+                    batch_mgr._truck_gone_tolerance = TOLERANCE_LOW_COUNT
+                elif current_count >= 40:
+                    batch_mgr._truck_gone_tolerance = TOLERANCE_HIGH_COUNT
+                else:
+                    batch_mgr._truck_gone_tolerance = TOLERANCE_MED_COUNT
+
+                batch_mgr.update_sacks(
+                    has_crossing_event=has_crossing,
+                    sacks_in_area_count=sacks_in_truck_area,
+                    timestamp=timestamp,
+                    loading_count=counter.loading_count,
+                    unloading_count=counter.unloading_count,
+                )
+
+                if batch_mgr.state == "WAITING_FOR_ACTIVITY":
+                    batch_mgr.update_truck(sacks_in_truck_area > 0, None, timestamp)
+
+                for ev in events:
+                    now_str = datetime.now().strftime("%H:%M:%S")
+                    print(f"[{now_str}] [KARUNG] #{ev['track_id']} melintasi garis. Total: {counter.loading_count}")
+
+            # =============================================================
+            # TRANSISI BATCH — MULAI/SELESAI REKAMAN VIDEO
+            # =============================================================
+
+            # C. Batch baru saja dimulai
+            if batch_mgr.is_active and not prev_active:
+                counting_date = get_counting_date()
+                batch_counter += 1
+                now_str = datetime.now().strftime("%H:%M:%S")
+                print(f"\n>>> [{now_str}] BATCH #{batch_counter} DIMULAI (tanggal: {counting_date}) <<<")
+                recorder.start_recording(batch_counter, counting_date, 1280, 720)
+
+            # D. Batch baru saja selesai
+            elif not batch_mgr.is_active and prev_active:
+                final_count = counter.loading_count
+                now_str = datetime.now().strftime("%H:%M:%S")
+                print(f"\n>>> [{now_str}] BATCH #{batch_counter} SELESAI. Total karung: {final_count} <<<")
+                recorder.stop_recording()
+
+                # Reset counter dan stabilizer untuk batch berikutnya
+                counter.reset()
+                stabilizer.reset()
+
+            # E. Log transisi status
+            if batch_mgr.state != prev_state:
+                now_str = datetime.now().strftime("%H:%M:%S")
+                print(f"[{now_str}] [STATE] {prev_state} -> {batch_mgr.state}")
+
+            # =============================================================
+            # TULIS FRAME KE VIDEO (jika batch aktif)
+            # =============================================================
+            if batch_mgr.is_active and recorder.is_recording:
+                recorder.write_frame(frame)
+
+    except KeyboardInterrupt:
+        print("\n\n[INFO] Program dihentikan oleh pengguna (Ctrl+C).")
+
+    finally:
+        # Tutup rekaman yang masih terbuka
+        if recorder.is_recording:
+            print("[INFO] Menyimpan rekaman batch terakhir...")
+            recorder.stop_recording()
+
+        cap.release()
+
+        print("\n" + "=" * 60)
+        print("  BATCH VIDEO CROPPER SELESAI")
+        print(f"  Total batch terekam: {batch_counter}")
+        print(f"  Total frame diproses: {frame_idx}")
+        print(f"  Folder output: {ARCHIVE_BASE}")
+        print("=" * 60)
+
+
+if __name__ == "__main__":
+    run_batch_video_cropper()

algoritma-batch/cfg/tracker.yaml

@@ -0,0 +1,26 @@
+# Custom FastTrack config tuned for sack counting:
+#   - track_buffer=60: hold lost tracks for 60 frames (~2.4s at 25fps)
+#     to survive worker occlusion
+#   - new_track_thresh=0.3: harder to spawn duplicate IDs
+#   - track_low_thresh=0.05: recover faint detections behind workers
+#   - active_occ_to_lost_thresh=15: tolerate 15 occluded frames
+#   - occ_reappear_window=60: re-find tracks after long occlusion
+#   - enlarge_bbox_occ=1.15: widen search region during occlusion
+
+tracker_type: bytetrack
+track_high_thresh: 0.20
+track_low_thresh: 0.05
+new_track_thresh: 0.30
+track_buffer: 60
+match_thresh: 0.85
+fuse_score: true
+
+# Occlusion handling (FastTrack-specific)
+reset_velocity_offset_occ: 5
+reset_pos_offset_occ: 3
+enlarge_bbox_occ: 1.15
+dampen_motion_occ: 0.4
+active_occ_to_lost_thresh: 15
+occ_cover_thresh: 0.6
+occ_reappear_window: 60
+init_iou_suppress: 0.65

algoritma-batch/src/__init__.py

@@ -0,0 +1 @@
+"""Package marker."""

algoritma-batch/src/batch.py

@@ -0,0 +1,390 @@
+"""Batch lifecycle manager — 4-state machine for truck+sack sessions.
+
+State machine:
+  IDLE ──truck detected──▶ TRUCK_STABILIZING ──stable 5s──▶ COUNTING_SACKS
+   ▲                            │ truck gone                   │  ▲
+   │                            └──────▶ IDLE                  │  │
+   │                                                           │  │
+   │                               10s no sack activity        │  │ sacks resume
+   │                                                           ▼  │
+   │                                                  WAITING_FOR_ACTIVITY
+   │                                                     (batch OPEN)
+   │                                                           │
+   └──────────────── truck leaves ─────────────────────────────┘
+                     (batch finalized)
+"""
+
+from __future__ import annotations
+
+import time
+import math
+from dataclasses import dataclass, field
+from enum import Enum, auto
+
+
+class BatchState(Enum):
+    IDLE = auto()
+    TRUCK_STABILIZING = auto()
+    COUNTING_SACKS = auto()
+    WAITING_FOR_ACTIVITY = auto()  # Paused: no sacks, but truck still here
+
+
+@dataclass
+class BatchRecord:
+    """Completed batch summary."""
+
+    batch_id: int
+    start_time: float
+    end_time: float
+    loading_count: int
+    unloading_count: int
+
+    @property
+    def net_count(self) -> int:
+        return self.loading_count - self.unloading_count
+
+    @property
+    def duration_seconds(self) -> float:
+        return self.end_time - self.start_time
+
+
+class BatchLifecycleManager:
+    """Manages batch transitions based on truck stability and sack activity.
+
+    State flow:
+    - IDLE: waiting for truck to appear in ROI polygon
+    - TRUCK_STABILIZING: truck seen, tracking centroid stability
+    - COUNTING_SACKS: actively counting sacks crossing line
+    - WAITING_FOR_ACTIVITY: sacks idle, but truck still present — batch stays open
+    """
+
+    def __init__(
+        self,
+        stabilize_seconds: float = 5.0,
+        stabilize_threshold_px: float = 15.0,
+        sack_idle_timeout: float = 10.0,
+        min_batch_duration: float = 30.0,
+        truck_gone_tolerance: float = 3.0,
+        timeout_seconds: float = 30.0,  # kept for backward compat (unused)
+    ) -> None:
+        # Tunable parameters
+        self._stabilize_seconds = stabilize_seconds
+        self._stabilize_threshold_px = stabilize_threshold_px
+        self._sack_idle_timeout = sack_idle_timeout
+        self._min_batch_duration = min_batch_duration
+        self._truck_gone_tolerance = truck_gone_tolerance
+
+        # Internal state
+        self._state = BatchState.IDLE
+        self._batch_counter = 0
+        self._current_batch_id: int | None = None
+        self._batch_start_time = 0.0
+        self._history: list[BatchRecord] = []
+
+        # Truck stabilization tracking
+        self._truck_first_seen_time = 0.0
+        self._truck_last_centroid: tuple[float, float] | None = None
+        self._truck_stable_since = 0.0
+        self._truck_is_stable = False
+        self._truck_last_seen = 0.0  # timestamp when truck was last detected
+
+        # Sack activity tracking (for pause condition)
+        self._last_sack_crossing_time = 0.0
+        self._last_sack_seen_in_area_time = 0.0
+
+        # Waiting state tracking
+        self._waiting_since = 0.0
+
+        # Callbacks
+        self._on_batch_start: list = []
+        self._on_batch_end: list = []
+
+    # -- Public API: Register callbacks --
+
+    def on_batch_start(self, callback) -> None:
+        """Register callback: fn(batch_id, timestamp)."""
+        self._on_batch_start.append(callback)
+
+    def on_batch_end(self, callback) -> None:
+        """Register callback: fn(BatchRecord)."""
+        self._on_batch_end.append(callback)
+
+    # -- Public API: State update methods --
+
+    def update_truck(
+        self,
+        truck_detected: bool,
+        truck_centroid: tuple[float, float] | None,
+        timestamp: float,
+    ) -> None:
+        """Called during IDLE, TRUCK_STABILIZING, and WAITING_FOR_ACTIVITY states.
+
+        Args:
+            truck_detected: whether a truck is detected in the ROI polygon
+            truck_centroid: (cx, cy) of the truck bounding box, or None
+            timestamp: current time.time()
+        """
+        if self._state == BatchState.IDLE:
+            if truck_detected and truck_centroid is not None:
+                # Transition to STABILIZING
+                self._state = BatchState.TRUCK_STABILIZING
+                self._truck_first_seen_time = timestamp
+                self._truck_last_centroid = truck_centroid
+                self._truck_stable_since = timestamp
+                self._truck_is_stable = False
+                self._truck_last_seen = timestamp
+                print(f"[BATCH] Truk terdeteksi di area. Memantau stabilitas...")
+
+        elif self._state == BatchState.TRUCK_STABILIZING:
+            if truck_detected:
+                self._truck_last_seen = timestamp
+            
+            # Check if truck has been gone for too long (tolerance)
+            time_since_last_seen = timestamp - self._truck_last_seen
+            if not truck_detected and time_since_last_seen >= self._truck_gone_tolerance:
+                print(f"[BATCH] Truk hilang selama {time_since_last_seen:.1f}s. Kembali ke IDLE.")
+                self._state = BatchState.IDLE
+                self._truck_last_centroid = None
+                self._truck_is_stable = False
+                return
+
+            if truck_centroid is not None and self._truck_last_centroid is not None:
+                # Calculate centroid displacement
+                dx = truck_centroid[0] - self._truck_last_centroid[0]
+                dy = truck_centroid[1] - self._truck_last_centroid[1]
+                displacement = math.sqrt(dx * dx + dy * dy)
+
+                if displacement > self._stabilize_threshold_px:
+                    # Truck moved too much -> reset stability timer
+                    self._truck_stable_since = timestamp
+                    self._truck_is_stable = False
+
+                self._truck_last_centroid = truck_centroid
+
+                # Check if stable long enough
+                stable_duration = timestamp - self._truck_stable_since
+                if stable_duration >= self._stabilize_seconds:
+                    if not self._truck_is_stable:
+                        self._truck_is_stable = True
+                        print(f"[BATCH] Truk stabil selama {stable_duration:.1f}s. Memulai counting...")
+                        self._start_batch(timestamp)
+
+        elif self._state == BatchState.WAITING_FOR_ACTIVITY:
+            if truck_detected:
+                self._truck_last_seen = timestamp
+            
+            # Check if truck has been gone for tolerance period
+            time_since_last_seen = timestamp - self._truck_last_seen
+            if not truck_detected and time_since_last_seen >= self._truck_gone_tolerance:
+                # Truck has truly left! NOW we finalize the batch.
+                wait_duration = timestamp - self._waiting_since
+                print(
+                    f"[BATCH] Truk pergi setelah menunggu {wait_duration:.0f}s. "
+                    f"Batch selesai."
+                )
+                self._end_batch(timestamp, self._pending_loading, self._pending_unloading)
+
+    def update_sacks(
+        self,
+        has_crossing_event: bool,
+        sacks_in_area_count: int,
+        timestamp: float,
+        loading_count: int = 0,
+        unloading_count: int = 0,
+    ) -> None:
+        """Called during COUNTING_SACKS and WAITING_FOR_ACTIVITY states.
+
+        Args:
+            has_crossing_event: True if a sack crossed the counting line this frame
+            sacks_in_area_count: number of sacks currently detected in truck area
+            timestamp: current time.time()
+            loading_count: current cumulative loading count
+            unloading_count: current cumulative unloading count
+        """
+        # WAITING_FOR_ACTIVITY: if sacks appear again, resume counting in the SAME batch
+        if self._state == BatchState.WAITING_FOR_ACTIVITY:
+            if has_crossing_event or sacks_in_area_count > 0:
+                wait_duration = timestamp - self._waiting_since
+                print(
+                    f"[BATCH] Aktivitas karung terdeteksi setelah {wait_duration:.0f}s menunggu. "
+                    f"Melanjutkan counting batch #{self._current_batch_id}..."
+                )
+                self._state = BatchState.COUNTING_SACKS
+                self._last_sack_crossing_time = timestamp
+                self._last_sack_seen_in_area_time = timestamp
+                # Fall through to counting logic below
+            else:
+                return
+
+        if self._state != BatchState.COUNTING_SACKS:
+            return
+
+        # Update activity timers
+        if has_crossing_event:
+            self._last_sack_crossing_time = timestamp
+
+        if sacks_in_area_count > 0:
+            self._last_sack_seen_in_area_time = timestamp
+
+        # Store latest counts for when batch eventually ends
+        self._pending_loading = loading_count
+        self._pending_unloading = unloading_count
+
+        # Check pause condition: no sack activity for timeout period
+        batch_duration = timestamp - self._batch_start_time
+        time_since_last_crossing = timestamp - self._last_sack_crossing_time
+        time_since_last_sack_seen = timestamp - self._last_sack_seen_in_area_time
+
+        if (
+            batch_duration >= self._min_batch_duration
+            and time_since_last_crossing >= self._sack_idle_timeout
+            and time_since_last_sack_seen >= self._sack_idle_timeout
+        ):
+            print(
+                f"[BATCH] Tidak ada aktivitas karung selama {self._sack_idle_timeout}s. "
+                f"Menunggu truk pergi atau palet selanjutnya..."
+            )
+            self._state = BatchState.WAITING_FOR_ACTIVITY
+            self._waiting_since = timestamp
+
+    # -- Public API: Backward-compatible update (legacy) --
+
+    def update(
+        self,
+        truck_detected: bool,
+        timestamp: float,
+        loading_count: int = 0,
+        unloading_count: int = 0,
+    ) -> None:
+        """Legacy update method — kept for backward compatibility."""
+        if self._state in (BatchState.IDLE, BatchState.TRUCK_STABILIZING):
+            self.update_truck(truck_detected, None, timestamp)
+        elif self._state in (BatchState.COUNTING_SACKS, BatchState.WAITING_FOR_ACTIVITY):
+            self.update_sacks(
+                has_crossing_event=False,
+                sacks_in_area_count=1 if truck_detected else 0,
+                timestamp=timestamp,
+                loading_count=loading_count,
+                unloading_count=unloading_count,
+            )
+
+    # -- Properties --
+
+    @property
+    def state(self) -> str:
+        """Return current state as human-readable string."""
+        return self._state.name
+
+    @property
+    def current_batch_id(self) -> int | None:
+        return self._current_batch_id
+
+    @property
+    def is_active(self) -> bool:
+        """True during COUNTING or WAITING (batch is still open)."""
+        return self._state in (BatchState.COUNTING_SACKS, BatchState.WAITING_FOR_ACTIVITY)
+
+    @property
+    def is_counting(self) -> bool:
+        """True only during active sack counting."""
+        return self._state == BatchState.COUNTING_SACKS
+
+    @property
+    def is_waiting(self) -> bool:
+        """True when paused waiting for next pallet or truck departure."""
+        return self._state == BatchState.WAITING_FOR_ACTIVITY
+
+    @property
+    def is_stabilizing(self) -> bool:
+        return self._state == BatchState.TRUCK_STABILIZING
+
+    @property
+    def history(self) -> list[BatchRecord]:
+        return list(self._history)
+
+    @property
+    def batch_duration(self) -> float:
+        """Duration of current batch in seconds (0 if not active)."""
+        if not self.is_active:
+            return 0.0
+        return time.time() - self._batch_start_time
+
+    @property
+    def time_since_last_sack_activity(self) -> float:
+        """Seconds since last sack crossed line or seen in area."""
+        if not self.is_active:
+            return 0.0
+        now = time.time()
+        last_activity = max(self._last_sack_crossing_time, self._last_sack_seen_in_area_time)
+        return now - last_activity if last_activity > 0 else 0.0
+
+    @property
+    def waiting_duration(self) -> float:
+        """How long we've been in WAITING_FOR_ACTIVITY state."""
+        if self._state != BatchState.WAITING_FOR_ACTIVITY:
+            return 0.0
+        return time.time() - self._waiting_since
+
+    @property
+    def stabilize_progress(self) -> float:
+        """Progress of truck stabilization (0.0 to 1.0)."""
+        if self._state != BatchState.TRUCK_STABILIZING:
+            return 0.0
+        if self._stabilize_seconds <= 0.0:
+            return 1.0
+        elapsed = time.time() - self._truck_stable_since
+        return min(1.0, elapsed / self._stabilize_seconds)
+
+    def resume_batch(
+        self,
+        batch_id: int,
+        start_time: float,
+        loading_count: int,
+        unloading_count: int,
+    ) -> None:
+        """Resume a previously finalized batch."""
+        self._current_batch_id = batch_id
+        self._batch_counter = max(self._batch_counter, batch_id)
+        self._batch_start_time = start_time
+        self._pending_loading = loading_count
+        self._pending_unloading = unloading_count
+        self._state = BatchState.COUNTING_SACKS
+
+        # Pop from history if it was just completed
+        if self._history and self._history[-1].batch_id == batch_id:
+            self._history.pop()
+
+    # -- Private methods --
+
+    def _start_batch(self, timestamp: float) -> None:
+        self._batch_counter += 1
+        self._current_batch_id = self._batch_counter
+        self._batch_start_time = timestamp
+        self._last_sack_crossing_time = timestamp  # Grace period
+        self._last_sack_seen_in_area_time = timestamp  # Grace period
+        self._pending_loading = 0
+        self._pending_unloading = 0
+        self._state = BatchState.COUNTING_SACKS
+        for cb in self._on_batch_start:
+            cb(self._current_batch_id, timestamp)
+
+    def _end_batch(
+        self,
+        timestamp: float,
+        loading_count: int,
+        unloading_count: int,
+    ) -> None:
+        record = BatchRecord(
+            batch_id=self._current_batch_id or 0,
+            start_time=self._batch_start_time,
+            end_time=timestamp,
+            loading_count=loading_count,
+            unloading_count=unloading_count,
+        )
+        self._history.append(record)
+        self._state = BatchState.IDLE
+        self._current_batch_id = None
+        self._truck_last_centroid = None
+        self._truck_is_stable = False
+        for cb in self._on_batch_end:
+            cb(record)

algoritma-batch/src/counting.py

@@ -0,0 +1,237 @@
+"""Line-crossing counter — hybrid zone-based state tracking.
+
+Counting logic (Low-FPS robust):
+  Uses y1 (top edge) of the stabilized sack bounding box.
+
+  Each track_id goes through states:
+    UNKNOWN → ABOVE → COUNTED (when seen below line)
+    UNKNOWN → BELOW (ghost/appeared below line first → never counted)
+
+  Loading:   track had state ABOVE, now detected BELOW the zone
+  Unloading: track had state BELOW, now detected ABOVE the zone (if needed)
+
+  3-Layer deduplication:
+    Layer 1: State guard — must have been ABOVE before counting
+    Layer 2: Spatial dedup radius — same position can't trigger twice
+    Layer 3: Track ID — one track_id can only be counted once per direction
+
+  This approach is immune to low FPS because it doesn't require
+  detecting the exact frame of crossing. It only needs the track
+  to have been seen ABOVE the line at ANY point in its lifetime.
+"""
+
+from __future__ import annotations
+
+import time
+
+from src.interfaces import Detection
+
+
+class LineCrossCounter:
+    """Counts sacks crossing a horizontal zone using y1 (top edge).
+
+    The zone is a band [line_y - margin, line_y + margin].
+    A sack is "above" if y1 < line_y - margin,
+              "below" if y1 > line_y + margin.
+    While y1 is inside the band, state is held (no trigger).
+
+    Loading   = track was ever "above", now "below" (entered truck)
+    Unloading = track was ever "below", now "above" (left truck)
+    """
+
+    def __init__(
+        self,
+        line_y: int,
+        line_x_start: int,
+        line_x_end: int,
+        margin: int = 20,
+        dedup_radius: float = 60.0,
+    ) -> None:
+        self._line_y = line_y
+        self._line_x_start = line_x_start
+        self._line_x_end = line_x_end
+        self._margin = margin
+        self._dedup_radius = dedup_radius
+
+        self._loading_count = 0
+        self._unloading_count = 0
+
+        # track_id -> zone state for y1: "above" | "below" | None
+        self._state: dict[int, str | None] = {}
+        # track_id -> whether this track has EVER been in each zone
+        self._has_been_above: dict[int, bool] = {}
+        self._has_been_below: dict[int, bool] = {}
+        # track_id -> set of directions already counted
+        self._counted: dict[int, set[str]] = {}
+        # track_id -> initial coordinates (cx, y1) when first tracked
+        self._entry_points: dict[int, tuple[float, float]] = {}
+        # list of active deduplication circles
+        self._dedup_circles: list[dict] = []
+
+    @property
+    def entry_points(self) -> dict[int, tuple[float, float]]:
+        return self._entry_points
+
+    @property
+    def counted_tracks(self) -> dict[int, set[str]]:
+        return self._counted
+
+    @property
+    def line_y(self) -> int:
+        return self._line_y
+
+    @line_y.setter
+    def line_y(self, value: int) -> None:
+        self._line_y = value
+
+    @property
+    def line_x_start(self) -> int:
+        return self._line_x_start
+
+    @line_x_start.setter
+    def line_x_start(self, value: int) -> None:
+        self._line_x_start = value
+
+    @property
+    def line_x_end(self) -> int:
+        return self._line_x_end
+
+    @line_x_end.setter
+    def line_x_end(self, value: int) -> None:
+        self._line_x_end = value
+
+    def update(self, detections: list[Detection]) -> list[dict]:
+        """Process detections, return list of crossing events.
+
+        Hybrid approach:
+        - Tracks zone state per frame (above/below/in-band)
+        - BUT uses accumulated history (has_been_above) for counting decision
+        - A track counts as "loading" when:
+            1. It has been seen ABOVE the line at any previous point
+            2. Its current y1 is now BELOW the line
+            3. It hasn't been counted for loading yet
+            4. It passes spatial dedup check
+        """
+        now_t = time.time()
+        events: list[dict] = []
+        upper = self._line_y - self._margin
+        lower = self._line_y + self._margin
+
+        # Clean up expired dedup circles (older than 3.0 seconds)
+        self._dedup_circles = [c for c in self._dedup_circles if (now_t - c["time"]) <= 3.0]
+
+        for det in detections:
+            if det.track_id is None:
+                continue
+
+            x1, y1, x2, y2 = det.bbox
+            cx = (x1 + x2) / 2.0
+            tid = det.track_id
+
+            if tid not in self._entry_points:
+                self._entry_points[tid] = (cx, y1)
+
+            # Skip if centroid X outside counting bounds
+            if cx < self._line_x_start or cx > self._line_x_end:
+                continue
+
+            counted_dirs = self._counted.setdefault(tid, set())
+
+            # Determine y1 zone state (top edge of sack bbox)
+            if y1 < upper:
+                new_state = "above"
+            elif y1 > lower:
+                new_state = "below"
+            else:
+                new_state = self._state.get(tid)  # in band: hold
+
+            prev_state = self._state.get(tid)
+            self._state[tid] = new_state
+
+            # Track zone history — CRITICAL for low-FPS robustness
+            # Once a track has been seen above/below, it stays recorded forever
+            if new_state == "above":
+                self._has_been_above[tid] = True
+            elif new_state == "below":
+                self._has_been_below[tid] = True
+
+            # --- HYBRID COUNTING LOGIC ---
+            # Loading: track was EVER above, NOW below (entered truck from top)
+            # This works even if the track jumped over the line between frames
+            is_loading = (
+                new_state == "below"
+                and self._has_been_above.get(tid, False)
+                and "loading" not in counted_dirs
+            )
+
+            # Unloading: track was EVER below, NOW above (left truck)
+            is_unloading = (
+                new_state == "above"
+                and self._has_been_below.get(tid, False)
+                and "unloading" not in counted_dirs
+            )
+
+            if is_loading or is_unloading:
+                # Check spatial distance against all active dedup circles
+                is_duplicate = False
+                for circle in self._dedup_circles:
+                    dist = ((cx - circle["x"]) ** 2 + (y1 - circle["y"]) ** 2) ** 0.5
+                    if dist <= self._dedup_radius:
+                        is_duplicate = True
+                        break
+
+                if is_duplicate:
+                    continue
+
+                # Add this coordinate to the active dedup circles
+                self._dedup_circles.append({
+                    "x": cx,
+                    "y": y1,
+                    "time": now_t,
+                    "track_id": tid
+                })
+
+                if is_loading:
+                    self._loading_count += 1
+                    counted_dirs.add("loading")
+                    events.append({
+                        "track_id": tid,
+                        "direction": "loading",
+                        "cx": cx,
+                        "cy": y1
+                    })
+
+                elif is_unloading:
+                    self._unloading_count += 1
+                    counted_dirs.add("unloading")
+                    events.append({
+                        "track_id": tid,
+                        "direction": "unloading",
+                        "cx": cx,
+                        "cy": y1
+                    })
+
+        return events
+
+    @property
+    def loading_count(self) -> int:
+        return self._loading_count
+
+    @property
+    def unloading_count(self) -> int:
+        return self._unloading_count
+
+    @property
+    def net_count(self) -> int:
+        return self._loading_count - self._unloading_count
+
+    def reset(self) -> None:
+        """Reset all counters (new batch)."""
+        self._loading_count = 0
+        self._unloading_count = 0
+        self._state.clear()
+        self._has_been_above.clear()
+        self._has_been_below.clear()
+        self._counted.clear()
+        self._entry_points.clear()
+        self._dedup_circles.clear()

algoritma-batch/src/dashboard.py

@@ -0,0 +1,251 @@
+"""Dashboard overlay — draws counting info onto the video frame.
+
+Draws: truck ROI, counting zone (band), sack bounding boxes with y1
+marker (the crossing trigger edge), stats panel, batch history,
+and system state indicator.
+"""
+
+from __future__ import annotations
+
+import cv2
+import numpy as np
+
+from src.batch import BatchRecord
+from src.interfaces import Detection
+from src.truck_roi import TruckROI
+
+
+# Colors (BGR)
+GREEN = (0, 200, 0)
+RED = (0, 0, 220)
+CYAN = (220, 200, 0)
+WHITE = (255, 255, 255)
+YELLOW = (0, 230, 255)
+MAGENTA = (255, 0, 255)
+ORANGE = (0, 165, 255)
+GRAY = (140, 140, 140)
+DARK_GREEN = (0, 130, 0)
+LIGHT_BLUE = (255, 200, 100)
+
+
+# State display labels and colors
+STATE_DISPLAY = {
+    "IDLE": ("MENCARI TRUK...", ORANGE),
+    "TRUCK_STABILIZING": ("TRUK TERDETEKSI - STABILISASI", YELLOW),
+    "COUNTING_SACKS": ("MENGHITUNG KARUNG", GREEN),
+    "WAITING_FOR_ACTIVITY": ("MENUNGGU PALET / TRUK PERGI", LIGHT_BLUE),
+}
+
+
+class DashboardOverlay:
+    """Draws detection boxes, ROI, counting line, and stats onto frames."""
+
+    def draw(
+        self,
+        frame: np.ndarray,
+        detections: list[Detection],
+        roi: TruckROI | None,
+        loading_count: int,
+        unloading_count: int,
+        batch_id: int | None,
+        history: list[BatchRecord] | None = None,
+        system_state: str = "IDLE",
+        batch_duration: float = 0.0,
+        idle_timer: float = 0.0,
+        stabilize_progress: float = 0.0,
+        waiting_duration: float = 0.0,
+    ) -> np.ndarray:
+        out = frame.copy()
+        if roi is not None:
+            self._draw_roi(out, roi)
+            self._draw_counting_zone(out, roi)
+        self._draw_detections(out, detections)
+        self._draw_stats(out, loading_count, unloading_count, batch_id, history)
+        self._draw_system_state(
+            out, system_state, batch_duration, idle_timer, stabilize_progress,
+            waiting_duration,
+        )
+        return out
+
+    def _draw_roi(self, frame: np.ndarray, roi: TruckROI) -> None:
+        cv2.rectangle(
+            frame, (roi.x1, roi.y1), (roi.x2, roi.y2), ORANGE, 2,
+        )
+        cv2.putText(
+            frame, f"TRUCK ROI ({roi.confidence:.0%})",
+            (roi.x1, roi.y1 - 8),
+            cv2.FONT_HERSHEY_SIMPLEX, 0.5, ORANGE, 1,
+        )
+        # Draw truck top reference line (dashed via short segments)
+        for x in range(roi.x1, roi.x2, 20):
+            cv2.line(frame, (x, roi.y1), (min(x + 10, roi.x2), roi.y1), GRAY, 1)
+
+    def _draw_counting_zone(
+        self, frame: np.ndarray, roi: TruckROI, margin: int = 20,
+    ) -> None:
+        y = roi.line_y
+        # Draw zone band (semi-transparent)
+        overlay = frame.copy()
+        cv2.rectangle(
+            overlay, (roi.x1, y - margin), (roi.x2, y + margin),
+            MAGENTA, -1,
+        )
+        cv2.addWeighted(overlay, 0.15, frame, 0.85, 0, frame)
+        # Draw center line
+        cv2.line(frame, (roi.x1, y), (roi.x2, y), MAGENTA, 2)
+        cv2.putText(
+            frame,
+            f"COUNT LINE Y={y} (y1 trigger)",
+            (roi.x1, y - margin - 8),
+            cv2.FONT_HERSHEY_SIMPLEX, 0.5, MAGENTA, 1,
+        )
+
+    def _draw_detections(
+        self,
+        frame: np.ndarray,
+        detections: list[Detection],
+    ) -> None:
+        for det in detections:
+            x1, y1, x2, y2 = [int(v) for v in det.bbox]
+            label = "sack"
+            if det.track_id is not None:
+                label += f" #{det.track_id}"
+            label += f" {det.confidence:.0%}"
+
+            cv2.rectangle(frame, (x1, y1), (x2, y2), CYAN, 2)
+            cv2.putText(
+                frame, label, (x1, y1 - 6),
+                cv2.FONT_HERSHEY_SIMPLEX, 0.4, CYAN, 1,
+            )
+            # Mark TOP edge (y1) — the crossing trigger
+            cv2.line(frame, (x1, y1), (x2, y1), GREEN, 3)
+
+    def _draw_stats(
+        self,
+        frame: np.ndarray,
+        loading: int,
+        unloading: int,
+        batch_id: int | None,
+        history: list[BatchRecord] | None,
+    ) -> None:
+        # Panel background
+        cv2.rectangle(frame, (10, 10), (320, 160), (0, 0, 0), -1)
+        cv2.rectangle(frame, (10, 10), (320, 160), WHITE, 1)
+
+        batch_text = f"Batch #{batch_id}" if batch_id else "IDLE"
+        net = loading - unloading
+        y0 = 35
+
+        cv2.putText(
+            frame, batch_text, (20, y0),
+            cv2.FONT_HERSHEY_SIMPLEX, 0.7, YELLOW, 2,
+        )
+        cv2.putText(
+            frame, f"Loading:    {loading}", (20, y0 + 30),
+            cv2.FONT_HERSHEY_SIMPLEX, 0.6, GREEN, 2,
+        )
+        cv2.putText(
+            frame, f"Unloading:  {unloading}", (20, y0 + 60),
+            cv2.FONT_HERSHEY_SIMPLEX, 0.6, RED, 2,
+        )
+        cv2.putText(
+            frame, f"Net:        {net}", (20, y0 + 90),
+            cv2.FONT_HERSHEY_SIMPLEX, 0.6, WHITE, 2,
+        )
+
+        # History (last 3 batches)
+        if history:
+            y_h = 180
+            cv2.putText(
+                frame, "HISTORY", (20, y_h),
+                cv2.FONT_HERSHEY_SIMPLEX, 0.5, YELLOW, 1,
+            )
+            for rec in history[-3:]:
+                y_h += 22
+                txt = (
+                    f"B#{rec.batch_id}: "
+                    f"L={rec.loading_count} "
+                    f"U={rec.unloading_count} "
+                    f"Net={rec.net_count}"
+                )
+                cv2.putText(
+                    frame, txt, (20, y_h),
+                    cv2.FONT_HERSHEY_SIMPLEX, 0.4, WHITE, 1,
+                )
+
+    def _draw_system_state(
+        self,
+        frame: np.ndarray,
+        system_state: str,
+        batch_duration: float,
+        idle_timer: float,
+        stabilize_progress: float,
+        waiting_duration: float = 0.0,
+    ) -> None:
+        """Draw system state indicator bar at bottom of frame."""
+        h, w = frame.shape[:2]
+
+        # Get display info for current state
+        label, color = STATE_DISPLAY.get(system_state, ("UNKNOWN", GRAY))
+
+        # Draw state bar background
+        bar_h = 36
+        bar_y = h - bar_h
+        overlay = frame.copy()
+        cv2.rectangle(overlay, (0, bar_y), (w, h), (0, 0, 0), -1)
+        cv2.addWeighted(overlay, 0.7, frame, 0.3, 0, frame)
+
+        # Draw colored indicator dot
+        cv2.circle(frame, (20, bar_y + bar_h // 2), 8, color, -1)
+        cv2.circle(frame, (20, bar_y + bar_h // 2), 8, WHITE, 1)
+
+        # Draw state label
+        cv2.putText(
+            frame, label, (36, bar_y + bar_h // 2 + 5),
+            cv2.FONT_HERSHEY_SIMPLEX, 0.55, color, 2,
+        )
+
+        # Draw additional info based on state
+        if system_state == "TRUCK_STABILIZING":
+            # Draw stabilization progress bar
+            prog_x = 340
+            prog_w = 150
+            prog_h = 14
+            prog_y = bar_y + (bar_h - prog_h) // 2
+
+            cv2.rectangle(frame, (prog_x, prog_y), (prog_x + prog_w, prog_y + prog_h), GRAY, 1)
+            fill_w = int(prog_w * stabilize_progress)
+            if fill_w > 0:
+                cv2.rectangle(frame, (prog_x, prog_y), (prog_x + fill_w, prog_y + prog_h), YELLOW, -1)
+
+            pct_text = f"{stabilize_progress * 100:.0f}%"
+            cv2.putText(
+                frame, pct_text, (prog_x + prog_w + 8, prog_y + prog_h - 2),
+                cv2.FONT_HERSHEY_SIMPLEX, 0.4, YELLOW, 1,
+            )
+
+        elif system_state == "COUNTING_SACKS":
+            # Draw batch duration and idle timer
+            info_x = 340
+            dur_text = f"Durasi: {batch_duration:.0f}s"
+            cv2.putText(
+                frame, dur_text, (info_x, bar_y + 15),
+                cv2.FONT_HERSHEY_SIMPLEX, 0.4, WHITE, 1,
+            )
+
+            if idle_timer > 0:
+                idle_color = RED if idle_timer > 7.0 else (YELLOW if idle_timer > 4.0 else WHITE)
+                idle_text = f"Idle: {idle_timer:.1f}s / 10s"
+                cv2.putText(
+                    frame, idle_text, (info_x, bar_y + 30),
+                    cv2.FONT_HERSHEY_SIMPLEX, 0.4, idle_color, 1,
+                )
+
+        elif system_state == "WAITING_FOR_ACTIVITY":
+            # Show waiting duration and batch info
+            info_x = 360
+            wait_text = f"Menunggu: {waiting_duration:.0f}s | Batch masih terbuka"
+            cv2.putText(
+                frame, wait_text, (info_x, bar_y + bar_h // 2 + 5),
+                cv2.FONT_HERSHEY_SIMPLEX, 0.4, LIGHT_BLUE, 1,
+            )

algoritma-batch/src/detection.py

@@ -0,0 +1,82 @@
+"""YOLO-based detectors for sacks and trucks.
+
+Each detector is a single-responsibility unit (S).  New model types can be
+added as new classes without touching these (O).
+"""
+
+from __future__ import annotations
+
+import numpy as np
+from ultralytics import YOLO
+
+from src.interfaces import Detection
+
+
+class SackDetector:
+    """Detects sacks (and persons) using a YOLO segmentation model."""
+
+    def __init__(self, model_path: str, conf: float = 0.40) -> None:
+        self._model = YOLO(model_path)
+        self._conf = conf
+
+    def detect(self, frame: np.ndarray) -> list[Detection]:
+        results = self._model.predict(
+            frame, conf=self._conf, verbose=False
+        )
+        return self._parse(results[0])
+
+    def _parse(self, result) -> list[Detection]:
+        detections: list[Detection] = []
+        masks = result.masks
+        for i, box in enumerate(result.boxes):
+            cls_id = int(box.cls[0])
+            name = self._model.names[cls_id]
+            if name != "sack":
+                continue
+            x1, y1, x2, y2 = box.xyxy[0].tolist()
+            mask = None
+            if masks is not None and i < len(masks):
+                mask = masks[i].data.cpu().numpy().squeeze()
+            detections.append(
+                Detection(
+                    bbox=(x1, y1, x2, y2),
+                    confidence=float(box.conf[0]),
+                    class_id=cls_id,
+                    class_name=name,
+                    mask=mask,
+                )
+            )
+        return detections
+
+
+class TruckDetector:
+    """Detects trucks using a YOLO detection model."""
+
+    def __init__(self, model_path_or_model: str | YOLO, conf: float = 0.50) -> None:
+        if isinstance(model_path_or_model, str):
+            self._model = YOLO(model_path_or_model)
+        else:
+            self._model = model_path_or_model
+        self._conf = conf
+
+    def detect(self, frame: np.ndarray) -> list[Detection]:
+        results = self._model.predict(
+            frame, conf=self._conf, verbose=False
+        )
+        return self._parse(results[0])
+
+    def _parse(self, result) -> list[Detection]:
+        detections: list[Detection] = []
+        for box in result.boxes:
+            cls_id = int(box.cls[0])
+            name = self._model.names[cls_id]
+            x1, y1, x2, y2 = box.xyxy[0].tolist()
+            detections.append(
+                Detection(
+                    bbox=(x1, y1, x2, y2),
+                    confidence=float(box.conf[0]),
+                    class_id=cls_id,
+                    class_name=name,
+                )
+            )
+        return detections

algoritma-batch/src/interfaces.py

@@ -0,0 +1,98 @@
+"""Abstract interfaces — all components code against these, never concretions.
+
+Keeps Interface Segregation (I) and Dependency Inversion (D) satisfied.
+Each protocol is tiny and single-purpose (S).
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from typing import Protocol, runtime_checkable
+
+import numpy as np
+
+
+# ── Data transfer objects ────────────────────────────────────────────────
+
+
+@dataclass
+class Detection:
+    """Single object detection."""
+
+    bbox: tuple[float, float, float, float]  # x1, y1, x2, y2
+    confidence: float
+    class_id: int
+    class_name: str
+    track_id: int | None = None
+    mask: np.ndarray | None = None  # segmentation mask (optional)
+
+
+@dataclass
+class FrameResult:
+    """All detections for one frame."""
+
+    detections: list[Detection] = field(default_factory=list)
+    frame_index: int = 0
+    timestamp: float = 0.0
+
+
+# ── Protocols ────────────────────────────────────────────────────────────
+
+
+@runtime_checkable
+class StreamSource(Protocol):
+    """Reads frames from a video source."""
+
+    def open(self) -> bool: ...
+    def read(self) -> tuple[bool, np.ndarray | None]: ...
+    def release(self) -> None: ...
+    @property
+    def fps(self) -> float: ...
+    @property
+    def frame_size(self) -> tuple[int, int]: ...
+
+
+@runtime_checkable
+class Detector(Protocol):
+    """Runs inference on a frame and returns detections."""
+
+    def detect(self, frame: np.ndarray) -> list[Detection]: ...
+
+
+@runtime_checkable
+class Tracker(Protocol):
+    """Assigns persistent IDs to detections across frames."""
+
+    def update(
+        self, frame: np.ndarray, detections: list[Detection]
+    ) -> list[Detection]: ...
+
+    def reset(self) -> None: ...
+
+
+@runtime_checkable
+class Counter(Protocol):
+    """Counts objects crossing a virtual boundary."""
+
+    def update(self, detections: list[Detection]) -> None: ...
+
+    @property
+    def loading_count(self) -> int: ...
+
+    @property
+    def unloading_count(self) -> int: ...
+
+    def reset(self) -> None: ...
+
+
+@runtime_checkable
+class BatchManager(Protocol):
+    """Manages batch lifecycle based on truck presence."""
+
+    def update(self, truck_detected: bool, timestamp: float) -> None: ...
+
+    @property
+    def current_batch_id(self) -> int | None: ...
+
+    @property
+    def is_active(self) -> bool: ...

algoritma-batch/src/stabilizer.py

@@ -0,0 +1,126 @@
+"""Bounding-box stabilizer — EMA smoothing + dual height clamping + dropout hold.
+
+Addresses four occlusion/flickering problems:
+1. Bbox jitter: raw detections jump 10-50px between frames.
+   Fix: EMA (exponential moving average) on bbox coordinates.
+2. Bbox loss at line: worker's head/body blocks sack for 5-10 frames.
+   Fix: hold last known smoothed bbox for `max_hold_frames` (10 frames).
+3. Height expansion spike: worker body merges into sack bbox.
+   Fix: clamp height expansion (`raw_h > smooth_h * max_h_ratio`).
+4. Height shrinkage collapse: worker head/shoulder covers bottom of sack.
+   Fix: clamp height shrinkage (`raw_h < smooth_h * min_h_ratio`).
+
+All rules are applied per track ID to maintain smooth trajectories.
+"""
+
+from __future__ import annotations
+
+from src.interfaces import Detection
+
+
+class BboxStabilizer:
+    """Smooths and holds bounding boxes per track ID against worker occlusion."""
+
+    def __init__(
+        self,
+        ema_alpha: float = 0.35,
+        max_hold_frames: int = 10,
+        max_height_ratio: float = 1.5,
+        min_height_ratio: float = 0.70,
+    ) -> None:
+        self._alpha = ema_alpha
+        self._max_hold = max_hold_frames
+        self._max_h_ratio = max_height_ratio
+        self._min_h_ratio = min_height_ratio
+
+        # tid -> (smoothed_x1, smoothed_y1, smoothed_x2, smoothed_y2)
+        self._smooth: dict[int, tuple[float, float, float, float]] = {}
+        # tid -> frames since last real detection
+        self._age: dict[int, int] = {}
+        # tid -> last confidence and class info
+        self._meta: dict[int, tuple[float, int, str]] = {}
+
+    def update(self, detections: list[Detection]) -> list[Detection]:
+        """Smooth incoming detections + inject held tracks during occlusion."""
+        seen_tids: set[int] = set()
+        result: list[Detection] = []
+
+        # 1. Process real detections — apply EMA & dual height clamping
+        for det in detections:
+            tid = det.track_id
+            if tid is None:
+                result.append(det)
+                continue
+
+            seen_tids.add(tid)
+            self._age[tid] = 0
+            self._meta[tid] = (det.confidence, det.class_id, det.class_name)
+
+            x1, y1, x2, y2 = det.bbox
+            raw_h = y2 - y1
+
+            if tid in self._smooth:
+                sx1, sy1, sx2, sy2 = self._smooth[tid]
+                smooth_h = sy2 - sy1
+
+                if smooth_h > 0:
+                    # Height expansion clamp (worker body merged)
+                    if raw_h > smooth_h * self._max_h_ratio:
+                        y2 = y1 + smooth_h * self._max_h_ratio
+                    # Height shrinkage clamp (worker head/shoulder blocking bottom)
+                    elif raw_h < smooth_h * self._min_h_ratio:
+                        y2 = y1 + smooth_h * self._min_h_ratio
+
+                a = self._alpha
+                sx1 = a * x1 + (1 - a) * sx1
+                sy1 = a * y1 + (1 - a) * sy1
+                sx2 = a * x2 + (1 - a) * sx2
+                sy2 = a * y2 + (1 - a) * sy2
+            else:
+                sx1, sy1, sx2, sy2 = x1, y1, x2, y2
+
+            self._smooth[tid] = (sx1, sy1, sx2, sy2)
+
+            result.append(Detection(
+                bbox=(sx1, sy1, sx2, sy2),
+                confidence=det.confidence,
+                class_id=det.class_id,
+                class_name=det.class_name,
+                track_id=tid,
+                mask=det.mask,
+            ))
+
+        # 2. Hold tracks missing this frame (occlusion tolerance)
+        expired: list[int] = []
+        for tid in list(self._age.keys()):
+            if tid in seen_tids:
+                continue
+            self._age[tid] += 1
+            if self._age[tid] > self._max_hold:
+                expired.append(tid)
+                continue
+
+            # Inject held bbox from last smoothed position
+            sx1, sy1, sx2, sy2 = self._smooth[tid]
+            conf, cls_id, cls_name = self._meta[tid]
+            result.append(Detection(
+                bbox=(sx1, sy1, sx2, sy2),
+                confidence=conf * 0.85,  # gentle decay during occlusion hold
+                class_id=cls_id,
+                class_name=cls_name,
+                track_id=tid,
+            ))
+
+        # 3. Clean up expired tracks
+        for tid in expired:
+            del self._smooth[tid]
+            del self._age[tid]
+            del self._meta[tid]
+
+        return result
+
+    def reset(self) -> None:
+        """Clear all state (new batch)."""
+        self._smooth.clear()
+        self._age.clear()
+        self._meta.clear()

algoritma-batch/src/tracking.py

@@ -0,0 +1,83 @@
+"""FastTrack wrapper — occlusion-aware tracker with custom tuning.
+
+Uses Ultralytics FastTrack which handles:
+  - Kalman rollback on occlusion onset (restores pre-occlusion velocity)
+  - Enlarged search region during occlusion
+  - Re-identification of occluded tracks after reappearance
+
+Our custom cfg/tracker.yaml tunes:
+  - track_buffer=60 (hold lost tracks ~2.4s to survive worker occlusion)
+  - new_track_thresh=0.3 (prevent duplicate IDs from spawning)
+  - active_occ_to_lost_thresh=15 (tolerate 15 occluded frames)
+"""
+
+from __future__ import annotations
+
+import os
+
+import numpy as np
+from ultralytics import YOLO
+
+from src.interfaces import Detection
+
+_TRACKER_CFG = os.path.join(
+    os.path.dirname(os.path.dirname(__file__)), "cfg", "tracker.yaml"
+)
+
+
+class ByteTrackTracker:
+    """Tracks sacks across frames using FastTrack (occlusion-aware)."""
+
+    def __init__(self, model_path_or_model: str | YOLO, conf: float = 0.35) -> None:
+        if isinstance(model_path_or_model, str):
+            self._model = YOLO(model_path_or_model)
+            self._model_path = model_path_or_model
+        else:
+            self._model = model_path_or_model
+            self._model_path = model_path_or_model.ckpt_path if hasattr(model_path_or_model, 'ckpt_path') else ""
+        self._conf = conf
+        self._tracker_cfg = _TRACKER_CFG
+
+    def update(
+        self, frame: np.ndarray, detections: list[Detection]
+    ) -> list[Detection]:
+        """Run tracking on the frame, return detections with track IDs."""
+        results = self._model.track(
+            frame,
+            conf=self._conf,
+            persist=True,
+            tracker=self._tracker_cfg,
+            verbose=False,
+        )
+        return self._parse(results[0])
+
+    def _parse(self, result) -> list[Detection]:
+        tracked: list[Detection] = []
+        ids = result.boxes.id
+        for i, box in enumerate(result.boxes):
+            cls_id = int(box.cls[0])
+            name = self._model.names[cls_id]
+            # Retain both 'sack' and 'truck' classes
+            if name not in ("sack", "truck"):
+                continue
+            track_id = int(ids[i]) if ids is not None else None
+            x1, y1, x2, y2 = box.xyxy[0].tolist()
+
+            mask = None
+
+            tracked.append(
+                Detection(
+                    bbox=(x1, y1, x2, y2),
+                    confidence=float(box.conf[0]),
+                    class_id=cls_id,
+                    class_name=name,
+                    track_id=track_id,
+                    mask=mask,
+                )
+            )
+        return tracked
+
+    def reset(self) -> None:
+        """Reset tracker state (new batch / new truck)."""
+        if self._model_path:
+            self._model = YOLO(self._model_path)

algoritma-batch/src/truck_roi.py

@@ -0,0 +1,151 @@
+"""Truck ROI tracker — identifies the main truck and provides a stable ROI.
+
+Uses exponential moving average (EMA) to smooth the bounding box across
+frames, preventing jitter from frame-to-frame detection variance.
+
+For y2-based counting (bottom edge of sack bbox), the counting line is
+placed `LINE_OFFSET_PX` pixels relative to the truck bottom edge (`roi.y2`).
+With `offset = +20`, the line sits at `roi.y2 + 20` (~620px), cleanly
+separating sacks on the ground (`y2 > 650`) from loaded sacks (`y2 < 580`).
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+
+from src.interfaces import Detection
+
+# Offset for y1 counting line relative to truck top edge (px).
+# Positive = below truck top edge (into the truck).
+# Negative = above truck top edge (towards the camera).
+LINE_OFFSET_PX = 0
+
+
+@dataclass
+class TruckROI:
+    """Region of interest derived from the main truck bbox."""
+
+    x1: int
+    y1: int
+    x2: int
+    y2: int
+    line_y: int       # counting line Y position (pixels)
+    confidence: float
+
+    @property
+    def width(self) -> int:
+        return self.x2 - self.x1
+
+    @property
+    def height(self) -> int:
+        return self.y2 - self.y1
+
+    def contains_x(self, cx: float) -> bool:
+        """Check if a centroid X falls within the truck X bounds."""
+        return self.x1 <= cx <= self.x2
+
+
+class TruckROITracker:
+    """Tracks the main truck and provides a smoothed ROI + counting line.
+
+    Main truck = largest truck detection whose center X falls in the
+    expected lane (center region of the frame).
+
+    The counting line is placed `line_offset` pixels below the truck
+    bottom edge (`roi.y2`).
+    """
+
+    def __init__(
+        self,
+        frame_width: int,
+        frame_height: int,
+        lane_x_min: float = 0.35,
+        lane_x_max: float = 0.80,
+        ema_alpha: float = 0.15,
+        line_offset: int = LINE_OFFSET_PX,
+    ) -> None:
+        self._fw = frame_width
+        self._fh = frame_height
+        self._lane_x_min = int(lane_x_min * frame_width)
+        self._lane_x_max = int(lane_x_max * frame_width)
+        self._alpha = ema_alpha
+        self._line_offset = line_offset
+
+        # Smoothed bbox (None until first detection)
+        self._sx1: float | None = None
+        self._sy1: float | None = None
+        self._sx2: float | None = None
+        self._sy2: float | None = None
+
+        self._last_roi: TruckROI | None = None
+        self._frames_without_truck = 0
+
+    def update(self, truck_detections: list[Detection]) -> TruckROI | None:
+        """Pick the main truck, smooth its bbox, return ROI."""
+        main = self._pick_main_truck(truck_detections)
+
+        if main is None:
+            self._frames_without_truck += 1
+            if self._frames_without_truck > 5:  # Clear ROI if truck is missing for >5 updates (~3 seconds)
+                self.reset()
+                return None
+            return self._last_roi  # hold last known ROI briefly
+
+        self._frames_without_truck = 0
+        x1, y1, x2, y2 = main.bbox
+
+        # EMA smoothing
+        if self._sx1 is None:
+            self._sx1, self._sy1 = float(x1), float(y1)
+            self._sx2, self._sy2 = float(x2), float(y2)
+        else:
+            a = self._alpha
+            self._sx1 = a * x1 + (1 - a) * self._sx1
+            self._sy1 = a * y1 + (1 - a) * self._sy1
+            self._sx2 = a * x2 + (1 - a) * self._sx2
+            self._sy2 = a * y2 + (1 - a) * self._sy2
+
+        # Build ROI — line placed at truck top edge
+        roi_x1 = max(0, int(self._sx1))
+        roi_y1 = max(0, int(self._sy1))
+        roi_x2 = min(self._fw, int(self._sx2))
+        roi_y2 = min(self._fh, int(self._sy2))
+        line_y = roi_y1 + self._line_offset
+
+        self._last_roi = TruckROI(
+            x1=roi_x1, y1=roi_y1, x2=roi_x2, y2=roi_y2,
+            line_y=line_y, confidence=main.confidence,
+        )
+        return self._last_roi
+
+    @property
+    def roi(self) -> TruckROI | None:
+        return self._last_roi
+
+    @property
+    def frames_without_truck(self) -> int:
+        return self._frames_without_truck
+
+    def reset(self) -> None:
+        self._sx1 = self._sy1 = self._sx2 = self._sy2 = None
+        self._last_roi = None
+        self._frames_without_truck = 0
+
+    def _pick_main_truck(
+        self, detections: list[Detection]
+    ) -> Detection | None:
+        """Select the largest truck whose center X is in the expected lane."""
+        best: Detection | None = None
+        best_area = 0
+
+        for det in detections:
+            x1, y1, x2, y2 = det.bbox
+            cx = (x1 + x2) / 2
+            if not (self._lane_x_min <= cx <= self._lane_x_max):
+                continue
+            area = (x2 - x1) * (y2 - y1)
+            if area > best_area:
+                best = det
+                best_area = area
+
+        return best

algoritma-batch/zones.json

@@ -0,0 +1,43 @@
+{
+    "palet": [
+        [
+            612,
+            628
+        ],
+        [
+            1454,
+            630
+        ],
+        [
+            1458,
+            1074
+        ],
+        [
+            608,
+            1071
+        ]
+    ],
+    "truck": [
+        [600, 385],
+        [609, 1076],
+        [1404, 1078],
+        [1381, 343]
+    ],
+    "counting": [
+        [574, 50],
+        [586, 1077],
+        [1418, 1076],
+        [1397, 50]
+    ],
+    "left_limit": 0.27578,
+    "right_limit": 0.72578,
+    "duplicate_circle_radius": 60,
+    "min_valid_area": 15000,
+    "jarak_toleransi_duplikat": 30,
+    "max_reid_transit_distance": 400,
+    "circle_stay_timeout_sec": 10.0,
+    "inference_stride": 2,
+    "confirm_delay_sec": 0.5,
+    "exit_confirm_delay_sec": 6.0,
+    "external_stream_url": "http://192.168.192.96:8888/cam/"
+}

backend/api/batches.py

@@ -1,13 +1,14 @@
-"""Batch, frame and auto-annotation routes (REQ-020…034)."""
+import json
-
 import os
+import shutil
+import tempfile
 from typing import Optional
 
-from fastapi import APIRouter, HTTPException, Response
+from fastapi import APIRouter, File, Form, HTTPException, Response, UploadFile
 from fastapi.responses import FileResponse
 from pydantic import BaseModel
 
-from backend import autolabel, dataset, library
+from backend import autolabel, dataset, library, projects
 from backend import batches as batch_store
 from backend import review as review_store
 from backend.api.common import project_or_404, thumbnail
@@ -27,10 +28,12 @@ class AutolabelRequest(BaseModel):
     engines: Optional[list[str]] = None
     class_ids: Optional[list[int]] = None
     engine_classes: Optional[dict[str, list[str]]] = None
+    target_class_names: Optional[list[str]] = None
     threshold: float = autolabel.DEFAULT_THRESHOLD
     iou_threshold: float = autolabel.DEFAULT_IOU
     min_box_frac: float = 0.0
     resume: bool = False
+    append: bool = False
 
 
 @router.post("/api/projects/{project_id}/batches")
@@ -90,11 +93,115 @@ def start_autolabel(batch_id: int, request: AutolabelRequest) -> dict:
     try:
         engine_list = request.engines if (request.engines and len(request.engines) > 0) else [request.engine]
         return autolabel.start(batch_id, request.threshold, request.iou_threshold,
-                               request.min_box_frac, resume=request.resume,
+                               request.min_box_frac, resume=request.resume, append=request.append,
                                engines=engine_list, class_ids=request.class_ids,
-                               engine_classes=request.engine_classes)
+                               engine_classes=request.engine_classes,
+                               target_class_names=request.target_class_names)
     except batch_store.BatchError as exc:
         raise HTTPException(400, str(exc))
+@router.post("/api/batches/inspect-model")
+async def inspect_model(file: UploadFile = File(...)) -> dict:
+    if not (file.filename or "").endswith(".pt"):
+        raise HTTPException(400, "Model must be a .pt file")
+    with tempfile.NamedTemporaryFile(suffix=".pt", delete=False) as staged:
+        shutil.copyfileobj(file.file, staged)
+        staged_path = staged.name
+    await file.close()
+    try:
+        classes = projects.read_model_classes(staged_path)
+        return {"filename": file.filename, "classes": classes, "staged_path": staged_path}
+    except Exception as exc:
+        if os.path.exists(staged_path):
+            os.unlink(staged_path)
+        raise HTTPException(400, f"Could not inspect model: {exc}")
+
+
+@router.post("/api/batches/{batch_id}/autolabel-with-model")
+async def autolabel_with_model(
+    batch_id: int,
+    file: UploadFile = File(...),
+    threshold: float = Form(0.35),
+    iou_threshold: float = Form(0.8),
+    selected_classes: str = Form("[]"),
+    append: bool = Form(True),
+) -> dict:
+    if not (file.filename or "").endswith(".pt"):
+        raise HTTPException(400, "Model must be a .pt file")
+    with tempfile.NamedTemporaryFile(suffix=".pt", delete=False) as staged:
+        shutil.copyfileobj(file.file, staged)
+        staged_path = staged.name
+    await file.close()
+
+    try:
+        target_classes = json.loads(selected_classes) if selected_classes else None
+        return autolabel.start(
+            batch_id,
+            threshold=threshold,
+            iou_threshold=iou_threshold,
+            append=append,
+            custom_model_path=staged_path,
+            target_class_names=target_classes,
+        )
+    except Exception as exc:
+        if os.path.exists(staged_path):
+            os.unlink(staged_path)
+        raise HTTPException(400, f"Auto-annotation failed to start: {exc}")
+
+
+@router.post("/api/sam3/playground-test")
+async def sam3_playground_test(
+    file: UploadFile = File(...),
+    prompts: str = Form(...),
+    threshold: float = Form(0.35),
+    iou_threshold: float = Form(0.8),
+) -> dict:
+    from PIL import Image
+    from backend import labeling
+    from backend.sam3_engine import get_engine
+
+    try:
+        image = Image.open(file.file).convert("RGB")
+    except Exception as exc:
+        raise HTTPException(400, f"Could not read image: {exc}")
+
+    width, height = image.size
+    prompt_list = [p.strip() for p in prompts.split(",") if p.strip()]
+    if not prompt_list:
+        raise HTTPException(400, "At least one text prompt is required")
+
+    try:
+        engine = get_engine()
+        raw_dets = engine.detect(image, prompt_list, threshold)
+        kept_dets = labeling.deduplicate(raw_dets, iou_threshold=iou_threshold)
+    except Exception as exc:
+        raise HTTPException(500, f"SAM3 inference failed: {exc}")
+
+    results = []
+    for det in kept_dets:
+        norm_box = [
+            det.box[0] / width,
+            det.box[1] / height,
+            det.box[2] / width,
+            det.box[3] / height,
+        ]
+        polys = []
+        if det.mask is not None:
+            raw_polys = review_store.mask_to_polygons(det.mask)
+            polys = [[[float(pt[0]), float(pt[1])] for pt in poly] for poly in raw_polys]
+
+        results.append({
+            "class_id": det.class_id,
+            "prompt": det.class_name,
+            "score": round(float(det.score), 4),
+            "box": [round(v, 5) for v in norm_box],
+            "polygons": polys,
+        })
+
+    return {
+        "width": width,
+        "height": height,
+        "detections": results,
+    }
 
 
 
@@ -150,3 +257,11 @@ def clear_batch_class_annotations(batch_id: int, class_id: int) -> dict:
     deleted = review_store.clear_batch_class_annotations(batch_id, class_id)
     return {"deleted": deleted}
 
+
+@router.post("/api/batches/{batch_id}/reset-auto-annotations")
+def reset_batch_auto_annotations(batch_id: int) -> dict:
+    if batch_store.get(batch_id) is None:
+        raise HTTPException(404, "No such batch")
+    deleted = review_store.clear_batch_auto_annotations(batch_id)
+    return {"deleted": deleted}
+

backend/api/models.py

@@ -19,6 +19,7 @@ class TrainRequest(BaseModel):
     imgsz: Optional[int] = None
     device: Optional[Union[int, str]] = None
     batch_ids: Optional[list] = None
+    class_ids: Optional[list] = None
 
 
 @router.get("/api/hardware")
@@ -34,6 +35,7 @@ def start_training(project_id: int, request: TrainRequest) -> dict:
             project_id, request.epochs,
             {"batch": request.batch, "imgsz": request.imgsz, "device": request.device},
             batch_ids=request.batch_ids,
+            class_ids=request.class_ids,
         )
     except training.TrainingError as exc:
         raise HTTPException(400, str(exc))

backend/autolabel.py

@@ -18,10 +18,12 @@ DEFAULT_IOU = 0.8
 
 def start(batch_id: int, threshold: float = DEFAULT_THRESHOLD,
           iou_threshold: float = DEFAULT_IOU, min_box_frac: float = 0.0,
-          resume: bool = False, engine: str = "sam3",
+          resume: bool = False, append: bool = False, engine: str = "base_model",
           engines: Optional[List[str]] = None,
           class_ids: Optional[List[int]] = None,
-          engine_classes: Optional[dict[str, List[str]]] = None) -> dict:
+          engine_classes: Optional[dict[str, List[str]]] = None,
+          custom_model_path: Optional[str] = None,
+          target_class_names: Optional[List[str]] = None) -> dict:
     batch = batches.get(batch_id)
     if batch is None:
         raise batches.BatchError("No such batch")
@@ -34,8 +36,9 @@ def start(batch_id: int, threshold: float = DEFAULT_THRESHOLD,
         "autolabel",
         params={"batch_id": batch_id, "threshold": threshold,
                 "iou_threshold": iou_threshold, "min_box_frac": min_box_frac,
-                "resume": resume, "engine": active_engines[0], "engines": active_engines,
+                "resume": resume, "append": append, "engine": active_engines[0], "engines": active_engines,
-                "class_ids": class_ids, "engine_classes": engine_classes},
+                "class_ids": class_ids, "engine_classes": engine_classes,
+                "custom_model_path": custom_model_path, "target_class_names": target_class_names},
         project_id=batch["project_id"],
         batch_id=batch_id,
         message=f"{batch['date_label']}/{batch['batch_label']} ({'+'.join(e.upper() for e in active_engines)})",
@@ -62,16 +65,7 @@ def _run_autolabel(job) -> None:
         raise batches.BatchError("The batch disappeared before labeling started")
     project = projects.get(batch["project_id"])
 
-    raw_active = job.params.get("engines") or [job.params.get("engine", "sam3")]
+    selected_engine = job.params.get("engine", "base_model")
-    expanded_engines = []
-    for eng in raw_active:
-        if eng == "both":
-            expanded_engines.extend(["base_model", "secondary_model"])
-        elif eng == "sam3+model1":
-            expanded_engines.extend(["sam3", "base_model"])
-        else:
-            expanded_engines.append(eng)
-    expanded_engines = list(dict.fromkeys(expanded_engines))
 
     frames = batches.frames(batch["id"])
     batches.set_status(batch["id"], "labeling")
@@ -86,42 +80,34 @@ def _run_autolabel(job) -> None:
     attempted = 0
     failures = []
 
-    from ultralytics import YOLO
+    conf = job.params.get("threshold", DEFAULT_THRESHOLD)
-    m1_path = projects.training_start_point(project)
+    iou_thresh = job.params.get("iou_threshold", DEFAULT_IOU)
-    with db.cursor() as cur:
-        cur.execute("SELECT weights_path FROM model_versions WHERE project_id = ? ORDER BY version DESC LIMIT 1", (project["id"],))
-        row = cur.fetchone()
-        if row and os.path.isfile(row[0]):
-            m1_path = row[0]
-
-    yolo_models = {}
-    if "base_model" in expanded_engines or "yolo" in expanded_engines:
-        job.log(f"Loading Base/Trained Model: {os.path.basename(m1_path)}...")
-        yolo_models["base_model"] = YOLO(m1_path)
-
-    if "secondary_model" in expanded_engines:
-        m2_path = project["secondary_model_path"] if (project.get("secondary_model_path") and os.path.isfile(project["secondary_model_path"])) else m1_path
-        label_name = project.get("secondary_model_name") or os.path.basename(m2_path)
-        job.log(f"Loading Secondary Model: {label_name}...")
-        yolo_models["secondary_model"] = YOLO(m2_path)
-
-    allowed_class_ids = set(job.params["class_ids"]) if job.params.get("class_ids") is not None else None
-    engine_classes = job.params.get("engine_classes") or {}
 
+    yolo_model = None
     sam3_target_classes = []
-    if "sam3" in expanded_engines:
+
-        sam3_classes = engine_classes.get("sam3")
+    custom_path = job.params.get("custom_model_path")
-        if sam3_classes is not None:
+    target_class_names = job.params.get("target_class_names")
-            allowed_set = {c.strip().lower() for c in sam3_classes}
+
-            sam3_target_classes = [
+    if selected_engine == "sam3" and not custom_path:
-                c for c in project["classes"]
+        allowed_classes_set = {c.strip().lower() for c in target_class_names} if target_class_names else None
-                if c["name"].strip().lower() in allowed_set or c["prompt"].strip().lower() in allowed_set
+        if allowed_classes_set:
-            ]
+            sam3_target_classes = [c for c in project["classes"] if c["name"].strip().lower() in allowed_classes_set or c["prompt"].strip().lower() in allowed_classes_set]
+            # Add any new target class names that aren't in project classes yet
+            existing_names = {c["name"].strip().lower() for c in project["classes"]}
+            for name in target_class_names:
+                if name.strip().lower() not in existing_names:
+                    try:
+                        updated_proj = projects.add_class(project["id"], name=name.strip(), prompt=name.strip())
+                        project["classes"] = updated_proj["classes"]
+                        for new_c in project["classes"]:
+                            if new_c["name"].strip().lower() == name.strip().lower() and new_c not in sam3_target_classes:
+                                sam3_target_classes.append(new_c)
+                    except Exception:
+                        pass
         else:
-            sam3_target_classes = [
+            sam3_target_classes = [c for c in project["classes"]]
-                c for c in project["classes"]
+
-                if (allowed_class_ids is None or c["class_id"] in allowed_class_ids)
-            ]
         prompts = [c["prompt"] for c in sam3_target_classes]
         if prompts:
             from backend.sam3_engine import engine_is_loaded, get_engine
@@ -130,13 +116,27 @@ def _run_autolabel(job) -> None:
             engine = get_engine()
             job.log(f"SAM3 ready on {engine.device}; prompts: {', '.join(prompts)}")
         else:
-            job.log("SAM3 selected but 0 prompts match class filter.")
+            job.log("SAM3 selected but 0 prompts match project classes.")
+    else:
+        from ultralytics import YOLO
+        if custom_path and os.path.isfile(custom_path):
+            m_path = custom_path
+            job.log(f"Loading Custom Model: {os.path.basename(m_path)}...")
+        else:
+            m_path = projects.training_start_point(project)
+            with db.cursor() as cur:
+                cur.execute("SELECT weights_path FROM model_versions WHERE project_id = ? ORDER BY version DESC LIMIT 1", (project["id"],))
+                row = cur.fetchone()
+                if row and os.path.isfile(row[0]):
+                    m_path = row[0]
+            job.log(f"Loading Base Model: {os.path.basename(m_path)}...")
+
+        yolo_model = YOLO(m_path)
 
     name_to_class_id = {item["name"].strip().lower(): item["class_id"] for item in project["classes"]}
-    conf = job.params.get("threshold", DEFAULT_THRESHOLD)
+    allowed_classes_set = {c.strip().lower() for c in target_class_names} if target_class_names else None
-    iou_thresh = job.params.get("iou_threshold", DEFAULT_IOU)
 
-    job.log(f"Starting multi-engine auto-labeling ({', '.join(expanded_engines)})...")
+    job.log(f"Starting auto-labeling with {selected_engine}...")
 
     for index, frame in enumerate(frames):
         if job.cancelled:
@@ -151,29 +151,21 @@ def _run_autolabel(job) -> None:
             frame_file = os.path.join(directory, frame["filename"])
             all_raw_detections = []
 
-            for eng_key, y_model in yolo_models.items():
+            if yolo_model is not None:
-                allowed_for_eng = engine_classes.get(eng_key)
+                results = yolo_model.predict(frame_file, conf=conf, verbose=False)
-                if allowed_for_eng is not None and len(allowed_for_eng) == 0:
-                    continue
-                results = y_model.predict(frame_file, conf=conf, verbose=False)
                 if results and len(results) > 0:
                     model_names = results[0].names
                     for box in results[0].boxes:
                         cls_idx = int(box.cls[0].item())
                         cls_name = str(model_names.get(cls_idx, cls_idx)).strip().lower()
-                        if allowed_for_eng is not None and cls_name not in [c.strip().lower() for c in allowed_for_eng]:
+                        
+                        if allowed_classes_set is not None and cls_name not in allowed_classes_set:
                             continue
-                        if cls_name not in name_to_class_id:
+
-                            try:
+                        target_class_id = name_to_class_id.get(cls_name)
-                                updated_proj = projects.add_class(project["id"], {"name": cls_name, "prompt": cls_name})
+                        if target_class_id is None:
-                                project["classes"] = updated_proj["classes"]
-                                name_to_class_id = {item["name"].strip().lower(): item["class_id"] for item in project["classes"]}
-                            except Exception:
-                                pass
-                        if cls_name in name_to_class_id:
-                            target_class_id = name_to_class_id[cls_name]
-                        else:
                             continue
+
                         score = float(box.conf[0].item())
                         xyxyn = box.xyxyn[0].tolist()
                         all_raw_detections.append(labeling.Detection(
@@ -184,7 +176,7 @@ def _run_autolabel(job) -> None:
                             mask=None
                         ))
 
-            if "sam3" in expanded_engines and sam3_target_classes:
+            elif selected_engine == "sam3" and sam3_target_classes:
                 prompts = [c["prompt"] for c in sam3_target_classes]
                 res = labeling.label_image(
                     frame_file, frame["filename"], prompts, conf,
@@ -208,7 +200,10 @@ def _run_autolabel(job) -> None:
                     for geometry in _geometries(det, frame["width"], frame["height"], project["label_type"]):
                         items.append({"class_id": det.class_id, "geometry": geometry, "score": det.score})
 
-            review.replace_auto(frame["id"], items)
+            if job.params.get("append"):
+                review.append_auto(frame["id"], items)
+            else:
+                review.replace_auto(frame["id"], items)
             written += len(items)
             job.progress(index + 1, len(frames), f"{frame['filename']}: {len(items)} shape(s)")
         except Exception as exc:

backend/dataset.py

@@ -14,6 +14,7 @@ Label files are plain YOLO:
 import os
 import shutil
 import time
+from typing import List, Optional
 
 from backend import batches, config, db, jobs, projects, review
 
@@ -74,12 +75,55 @@ def _next_split(cur, project_id: int, val_every: int) -> str:
     return "val" if position % val_every == val_every - 1 else "train"
 
 
-def write_data_yaml(project: dict, batch_ids: list = None) -> str:
+def sync_labels(project_id: int, selected_class_ids: Optional[List[int]] = None) -> dict:
+    """Re-sync label files on disk for all merged frames in the project dataset."""
+    project = projects.get(project_id)
+    root = dataset_dir(project["slug"])
+    with db.cursor() as cur:
+        cur.execute(
+            "SELECT d.frame_id, d.label_rel FROM dataset_items d WHERE d.project_id = ?",
+            (project_id,),
+        )
+        items = cur.fetchall()
+
+    class_map = None
+    if selected_class_ids is not None and len(selected_class_ids) > 0:
+        class_map = {cid: idx for idx, cid in enumerate(sorted(selected_class_ids))}
+
+    synced_files = 0
+    total_lines = 0
+    for frame_id, label_rel in items:
+        annotations = review.listing(frame_id)
+        if class_map is not None:
+            annotations = [a for a in annotations if a["class_id"] in class_map]
+
+        lines = []
+        for item in annotations:
+            mapped_cid = class_map[item["class_id"]] if class_map is not None else item["class_id"]
+            lines.append(_label_line(mapped_cid, item["geometry"], project["label_type"]))
+
+        path = os.path.join(root, label_rel)
+        os.makedirs(os.path.dirname(path), exist_ok=True)
+        with open(path, "w", encoding="utf-8") as f:
+            f.write("\n".join(lines) + ("\n" if lines else ""))
+        synced_files += 1
+        total_lines += len(lines)
+
+    return {"synced_files": synced_files, "total_lines": total_lines}
+
+
+def write_data_yaml(project: dict, batch_ids: list = None, selected_class_ids: Optional[List[int]] = None) -> str:
     """Rebuild data.yaml from the project's classes (REQ-051)."""
+    sync_labels(project["id"], selected_class_ids=selected_class_ids)
     root = dataset_dir(project["slug"])
     os.makedirs(root, exist_ok=True)
     counts = summary(project["id"])["splits"]
-    names = ", ".join(f"'{item['name']}'" for item in project["classes"])
+
+    target_classes = project["classes"]
+    if selected_class_ids is not None and len(selected_class_ids) > 0:
+        target_classes = [c for c in project["classes"] if c["class_id"] in selected_class_ids]
+
+    names = ", ".join(f"'{item['name']}'" for item in target_classes)
 
     if batch_ids:
         with db.cursor() as cur:

backend/hardware.py

@@ -44,9 +44,9 @@ def defaults(epochs: int = 50) -> dict:
     if info["device"] == "cpu":
         settings = {"batch": 4, "imgsz": 512, "device": "cpu", "workers": 2}
         note = "No GPU visible — training on CPU will be very slow."
-    elif vram < 8:
+    elif vram < 6:
-        settings = {"batch": 8, "imgsz": 640, "device": 0, "workers": 2}
+        settings = {"batch": 16, "imgsz": 640, "device": 0, "workers": 4}
-        note = f"{vram} GB of VRAM: small batches, 640 px."
+        note = f"{vram} GB of VRAM: batch 16, 640 px."
     elif vram <= 16:
         settings = {"batch": 32, "imgsz": 640, "device": 0, "workers": 8}
         note = f"{vram} GB of VRAM: optimized batch 32, 640 px."

backend/labeling.py

@@ -42,11 +42,18 @@ def _iou(box_a: List[float], box_b: List[float]) -> float:
 
 
 def deduplicate(detections: List[Detection], iou_threshold: float = 0.8) -> List[Detection]:
-    """Greedy NMS across all prompts: highest score wins an overlapping region."""
+    """Greedy NMS per class: highest score wins within the SAME class."""
+    by_class: dict[int, List[Detection]] = {}
+    for det in detections:
+        by_class.setdefault(det.class_id, []).append(det)
+
     kept: List[Detection] = []
-    for det in sorted(detections, key=lambda d: d.score, reverse=True):
+    for cls_dets in by_class.values():
-        if all(_iou(det.box, k.box) < iou_threshold for k in kept):
+        cls_kept: List[Detection] = []
-            kept.append(det)
+        for det in sorted(cls_dets, key=lambda d: d.score, reverse=True):
+            if all(_iou(det.box, k.box) < iou_threshold for k in cls_kept):
+                cls_kept.append(det)
+        kept.extend(cls_kept)
     return kept
 
 

backend/projects.py

@@ -384,6 +384,11 @@ def _kept_prompts(project: dict, names: List[str]) -> List[str]:
 def training_start_point(project: dict) -> str:
     """The weights a training run should start from (REQ-060, REQ-004)."""
     path = project["base_model_path"]
+    if path and not os.path.isfile(path):
+        if path.startswith("/data/"):
+            alt_path = os.path.join(config.DATA_DIR, path[6:])
+            if os.path.isfile(alt_path):
+                path = alt_path
     if path and os.path.isfile(path):
         return path
     return PRETRAINED[project["label_type"]]

backend/review.py

@@ -234,6 +234,38 @@ def replace_auto(frame_id: int, items: List[dict]) -> int:
     return len(items)
 
 
+def append_auto(frame_id: int, items: List[dict]) -> int:
+    """Add new automatic shapes to this frame without duplicating existing ones."""
+    if not items:
+        return 0
+    existing = listing(frame_id)
+    filtered_items = []
+    for item in items:
+        is_dup = False
+        item_box = to_box(item["geometry"])
+        for ex in existing:
+            if ex["class_id"] == item["class_id"]:
+                ex_box = to_box(ex["geometry"])
+                from backend.labeling import _iou
+                if _iou(item_box, ex_box) >= 0.85:
+                    is_dup = True
+                    break
+        if not is_dup:
+            filtered_items.append(item)
+
+    if not filtered_items:
+        return 0
+
+    with db.cursor() as cur:
+        cur.executemany(
+            """INSERT INTO annotations (frame_id, class_id, geometry, score, source, created_at)
+               VALUES (?, ?, ?, ?, 'auto', ?)""",
+            [(frame_id, item["class_id"], json.dumps(item["geometry"]),
+              item.get("score", 1.0), time.time()) for item in filtered_items],
+        )
+    return len(filtered_items)
+
+
 def frames_with_auto(batch_id: int) -> set:
     """Frame ids that already carry automatic shapes — the resume skip-list
     for REQ-035."""
@@ -339,6 +371,24 @@ def clear_batch_class_annotations(batch_id: int, class_id: int) -> int:
         return cur.rowcount
 
 
+def clear_batch_auto_annotations(batch_id: int) -> int:
+    """Delete all automatic annotations (source = 'auto') for a batch and reset frame statuses."""
+    with db.cursor() as cur:
+        cur.execute(
+            """DELETE FROM annotations
+               WHERE source = 'auto' AND frame_id IN (
+                   SELECT id FROM frames WHERE batch_id = ?
+               )""",
+            (batch_id,),
+        )
+        deleted = cur.rowcount
+        cur.execute(
+            "UPDATE frames SET review_status = 'pending' WHERE batch_id = ?",
+            (batch_id,),
+        )
+        return deleted
+
+
 def _check_class(project_id: int, class_id: int) -> None:
     with db.cursor() as cur:
         cur.execute(

backend/training.py

@@ -25,7 +25,7 @@ def models_dir(project_slug: str) -> str:
     return os.path.join(config.project_dir(project_slug), "models")
 
 
-def start(project_id: int, epochs: int = 50, overrides: Optional[dict] = None, batch_ids: Optional[list] = None) -> dict:
+def start(project_id: int, epochs: int = 50, overrides: Optional[dict] = None, batch_ids: Optional[list] = None, class_ids: Optional[list] = None) -> dict:
     project = projects.get(project_id)
     if project is None:
         raise TrainingError("No such project")
@@ -38,7 +38,7 @@ def start(project_id: int, epochs: int = 50, overrides: Optional[dict] = None, b
     settings = hardware.resolve(overrides, epochs)
     job = jobs.create(
         "train",
-        params={"project_id": project_id, "settings": settings, "batch_ids": batch_ids},
+        params={"project_id": project_id, "settings": settings, "batch_ids": batch_ids, "class_ids": class_ids},
         project_id=project_id,
         message=f"{counts['train']} train / {counts['val']} val",
     )
@@ -102,7 +102,8 @@ def _run_train(job) -> None:
     project = projects.get(job.params["project_id"])
     settings = job.params["settings"]
     batch_ids = job.params.get("batch_ids")
-    data_yaml = dataset.write_data_yaml(project, batch_ids=batch_ids)
+    class_ids = job.params.get("class_ids")
+    data_yaml = dataset.write_data_yaml(project, batch_ids=batch_ids, selected_class_ids=class_ids)
 
     # SAM3 and a training run must not hold VRAM at the same time (REQ-065).
     from backend.sam3_engine import release_engine
@@ -133,6 +134,10 @@ def _run_train(job) -> None:
     model.add_callback("on_fit_epoch_end", on_epoch)
     job.progress(0, settings["epochs"])
 
+    import torch
+    if torch.cuda.is_available():
+        torch.backends.cudnn.benchmark = True
+
     keep_run_dir = False
     try:
         model.train(
@@ -142,7 +147,7 @@ def _run_train(job) -> None:
             batch=settings["batch"],
             device=settings["device"],
             workers=settings.get("workers", 8),
-            cache=False,
+            cache="ram",
             project=os.path.join(out_dir, "runs"),
             name="train",
             exist_ok=True,

frontend/src/App.jsx

@@ -6,6 +6,7 @@ import LibraryPage from './pages/LibraryPage'
 import TrimPage from './pages/TrimPage'
 import ReviewPage from './pages/ReviewPage'
 import ModelsPage from './pages/ModelsPage'
+import Sam3PlaygroundPage from './pages/Sam3PlaygroundPage'
 import './roboflow.css'
 
 function parseRoute(hash) {
@@ -14,6 +15,10 @@ function parseRoute(hash) {
   const query = new URLSearchParams(queryString || '')
   const parts = path.split('/').filter(Boolean)
 
+  if (parts[0] === 'sam3-playground') {
+    return { name: 'sam3-playground' }
+  }
+
   if (parts[0] === 'batches' && parts[1]) {
     return { name: 'review', batchId: Number(parts[1]) }
   }
@@ -135,6 +140,7 @@ export default function App() {
               onProject={(p) => setCurrentProject(p)}
             />
           )}
+          {route.name === 'sam3-playground' && <Sam3PlaygroundPage />}
         </ErrorBoundary>
       </main>
     </div>

frontend/src/api.js

@@ -60,8 +60,32 @@ export const api = {
 
   startAutolabel: (batchId, body) =>
     request(`/batches/${batchId}/autolabel`, { method: 'POST', body: body ?? {} }),
+  inspectModel: (file) => {
+    const form = new FormData()
+    form.append('file', file)
+    return request('/batches/inspect-model', { method: 'POST', form })
+  },
+  autolabelWithModel: (batchId, file, threshold, selectedClasses, iouThreshold = 0.8) => {
+    const form = new FormData()
+    form.append('file', file)
+    form.append('threshold', threshold)
+    form.append('iou_threshold', iouThreshold)
+    form.append('selected_classes', JSON.stringify(selectedClasses))
+    form.append('append', 'true')
+    return request(`/batches/${batchId}/autolabel-with-model`, { method: 'POST', form })
+  },
+  sam3PlaygroundTest: (file, prompts, threshold = 0.35, iouThreshold = 0.8) => {
+    const form = new FormData()
+    form.append('file', file)
+    form.append('prompts', prompts)
+    form.append('threshold', threshold)
+    form.append('iou_threshold', iouThreshold)
+    return request('/sam3/playground-test', { method: 'POST', form })
+  },
   clearBatchClassAnnotations: (batchId, classId) =>
     request(`/batches/${batchId}/classes/${classId}/annotations`, { method: 'DELETE' }),
+  resetBatchAutoAnnotations: (batchId) =>
+    request(`/batches/${batchId}/reset-auto-annotations`, { method: 'POST' }),
   nextPending: (batchId, afterIdx = -1) =>
     request(`/batches/${batchId}/next-pending?after_idx=${afterIdx}`),
 

frontend/src/components/Sidebar.jsx

@@ -74,11 +74,15 @@ export default function Sidebar({ route, currentProject, theme, onToggleTheme })
       </div>
 
       <div className="sidebar-section">
-        {!collapsed && <div className="sidebar-section-title">MODELS</div>}
+        {!collapsed && <div className="sidebar-section-title">MODELS & TOOLS</div>}
         <a href={`#/projects/${pId}/models`} onClick={(e) => handleNav(e, `/projects/${pId}/models`)} className={`sidebar-item ${route.name === 'models' ? 'active' : ''}`} title="Train & Select Engine">
           <span className="sidebar-icon"><RocketIcon size={16} /></span>
           {!collapsed && <span>Train & Select Engine</span>}
         </a>
+        <a href="#/sam3-playground" onClick={(e) => handleNav(e, '/sam3-playground')} className={`sidebar-item ${route.name === 'sam3-playground' ? 'active' : ''}`} title="SAM3 Playground">
+          <span className="sidebar-icon">🤖</span>
+          {!collapsed && <span>SAM3 Playground</span>}
+        </a>
       </div>
 
       <div className="sidebar-spacer" />

frontend/src/pages/LibraryPage.jsx

@@ -43,40 +43,27 @@ function ActiveJobsBanner({ jobs, onCancel }) {
 
 function BatchList({ project, batches, activeJobs, onChanged, onError }) {
   const [busyId, setBusyId] = useState(null)
-  const [selectedEngine, setSelectedEngine] = useState({})
+  const [appendChoiceBatch, setAppendChoiceBatch] = useState(null)
-  const [engineClassMap, setEngineClassMap] = useState({})
+  const [appendModalState, setAppendModalState] = useState(null)
-  const [expandedFilterBatchId, setExpandedFilterBatchId] = useState(null)
+  const [sam3AppendState, setSam3AppendState] = useState(null)
+  const [customPromptInput, setCustomPromptInput] = useState('')
+  const [baseModelModalState, setBaseModelModalState] = useState(null)
 
-  const hasSecondaryModel = Boolean(project?.secondary_model_path)
+  function openBaseModelAutolabelModal(batch) {
-  const model1Classes = project?.classes?.map((c) => c.name) || []
+    const projectClasses = project?.classes?.map((c) => c.name) || []
-  const model2Classes = project?.secondary_model_classes?.length > 0
+    setBaseModelModalState({
-    ? project.secondary_model_classes
+      batch,
-    : model1Classes
+      selectedClasses: [...projectClasses],
+      threshold: 0.35,
+      iouThreshold: 0.8,
+    })
+  }
 
-  const defaultEngines = project?.base_model_path ? ['base_model'] : ['sam3']
+  async function resetAutoAnnotations(batch) {
-  const [selectedThreshold, setSelectedThreshold] = useState({})
+    if (!window.confirm(`Reset auto-annotations for batch "${batch.batch_label}"?\n\nThis will clear all auto-generated shapes and reset frame review statuses back to pending.`)) return
-
-  async function autolabel(batch, resume = false) {
     setBusyId(batch.id)
-    const engines = selectedEngine[batch.id] || defaultEngines
-    const threshold = selectedThreshold[batch.id] ?? 0.35
-    
-    // Default class filter for each active engine if not customized
-    const currentBatchMap = engineClassMap[batch.id] || {}
-    const engine_classes = {}
-    
-    if (engines.includes('sam3')) {
-      engine_classes.sam3 = currentBatchMap.sam3 || model1Classes
-    }
-    if (engines.includes('base_model')) {
-      engine_classes.base_model = currentBatchMap.base_model || model1Classes
-    }
-    if (engines.includes('secondary_model')) {
-      engine_classes.secondary_model = currentBatchMap.secondary_model || model2Classes
-    }
-
     try {
-      await api.startAutolabel(batch.id, { resume, engines, engine_classes, threshold })
+      await api.resetBatchAutoAnnotations(batch.id)
       onChanged()
     } catch (exc) {
       onError(exc.message)
@@ -85,35 +72,45 @@ function BatchList({ project, batches, activeJobs, onChanged, onError }) {
     }
   }
 
-  const toggleEngine = (batchId, engineKey) => {
+  function openFilePickerForYolo(batch) {
-    const current = selectedEngine[batchId] || defaultEngines
+    setAppendChoiceBatch(null)
-    let next
+    const input = document.createElement('input')
-    if (current.includes(engineKey)) {
+    input.type = 'file'
-      if (current.length === 1) return // keep at least 1 engine selected
+    input.accept = '.pt'
-      next = current.filter((e) => e !== engineKey)
+    input.onchange = async (e) => {
-    } else {
+      const file = e.target.files?.[0]
-      next = [...current, engineKey]
+      if (!file) return
+      setBusyId(batch.id)
+      try {
+        const info = await api.inspectModel(file)
+        setAppendModalState({
+          batch,
+          file,
+          filename: info.filename,
+          classes: info.classes || [],
+          selectedClasses: info.classes || [],
+          threshold: 0.35,
+          iouThreshold: 0.8,
+        })
+      } catch (err) {
+        onError(err.message)
+      } finally {
+        setBusyId(null)
+      }
     }
-    setSelectedEngine({ ...selectedEngine, [batchId]: next })
+    input.click()
   }
 
-  const toggleEngineClass = (batchId, engineKey, className, defaultClasses) => {
+  function openSam3AppendModal(batch) {
-    const batchFilters = engineClassMap[batchId] || {}
+    setAppendChoiceBatch(null)
-    const currentEngClasses = batchFilters[engineKey] || defaultClasses
+    const projectClasses = project?.classes?.map((c) => c.name) || []
-    let next
+    setSam3AppendState({
-    if (currentEngClasses.includes(className)) {
+      batch,
-      if (currentEngClasses.length === 1) return // keep at least 1 class
+      selectedClasses: [...projectClasses],
-      next = currentEngClasses.filter((c) => c !== className)
+      threshold: 0.35,
-    } else {
+      iouThreshold: 0.8,
-      next = [...currentEngClasses, className]
-    }
-    setEngineClassMap({
-      ...engineClassMap,
-      [batchId]: {
-        ...batchFilters,
-        [engineKey]: next,
-      },
     })
+    setCustomPromptInput('')
   }
 
   async function deleteBatch(batch) {
@@ -151,20 +148,13 @@ function BatchList({ project, batches, activeJobs, onChanged, onError }) {
         <thead>
           <tr>
             <th>Batch</th><th>Range</th><th>Frames</th><th>Reviewed</th>
-            <th>Shapes</th><th>Model & Class Configuration</th><th>Status</th><th />
+            <th>Shapes</th><th>Status</th><th />
           </tr>
         </thead>
         <tbody>
           {batches.map((batch) => {
             const batchJob = activeJobs?.find((j) => j.batch_id === batch.id)
             const isProcessing = Boolean(batchJob || busyId === batch.id)
-            const activeEngines = selectedEngine[batch.id] || defaultEngines
-            const isFilterExpanded = expandedFilterBatchId === batch.id
-            const batchMap = engineClassMap[batch.id] || {}
-
-            const sam3Active = batchMap.sam3 || model1Classes
-            const model1Active = batchMap.base_model || model1Classes
-            const model2Active = batchMap.secondary_model || model2Classes
 
             return (
               <React.Fragment key={batch.id}>
@@ -180,96 +170,6 @@ function BatchList({ project, batches, activeJobs, onChanged, onError }) {
                   <td className="num">{batch.frame_count}</td>
                   <td className="num">{batch.reviewed}/{batch.frame_count}</td>
                   <td className="num">{batch.annotation_count}</td>
-                  <td style={{ minWidth: 260 }}>
-                    <div style={{ display: 'flex', flexDirection: 'column', gap: 6 }}>
-                      <div style={{ display: 'flex', gap: 6, flexWrap: 'wrap', alignItems: 'center' }}>
-                        <button
-                          type="button"
-                          className={`btn ${activeEngines.includes('sam3') ? 'btn-primary' : 'btn-ghost'}`}
-                          style={{ padding: '2px 6px', fontSize: '0.72rem', cursor: 'pointer', opacity: activeEngines.includes('sam3') ? 1 : 0.6 }}
-                          onClick={() => toggleEngine(batch.id, 'sam3')}
-                          disabled={isProcessing}
-                          title="SAM3 Zero-shot Text Prompt Engine"
-                        >
-                          🤖 SAM3
-                        </button>
-                        <button
-                          type="button"
-                          className={`btn ${activeEngines.includes('base_model') ? 'btn-primary' : 'btn-ghost'}`}
-                          style={{ padding: '2px 6px', fontSize: '0.72rem', cursor: 'pointer', opacity: activeEngines.includes('base_model') ? 1 : 0.6 }}
-                          onClick={() => toggleEngine(batch.id, 'base_model')}
-                          disabled={isProcessing}
-                          title="Primary Model 1 (Base / Trained)"
-                        >
-                          ⚡ Model 1
-                        </button>
-                        {hasSecondaryModel && (
-                          <button
-                            type="button"
-                            className={`btn ${activeEngines.includes('secondary_model') ? 'btn-primary' : 'btn-ghost'}`}
-                            style={{ padding: '2px 6px', fontSize: '0.72rem', cursor: 'pointer', opacity: activeEngines.includes('secondary_model') ? 1 : 0.6 }}
-                            onClick={() => toggleEngine(batch.id, 'secondary_model')}
-                            disabled={isProcessing}
-                            title={project?.secondary_model_name || "Secondary Uploaded Model 2"}
-                          >
-                            🎯 Model 2
-                          </button>
-                        )}
-                        <button
-                          type="button"
-                          className="btn btn-ghost"
-                          style={{
-                            padding: '2px 6px',
-                            fontSize: '0.72rem',
-                            display: 'flex',
-                            alignItems: 'center',
-                            gap: 3,
-                            borderColor: isFilterExpanded ? '#c084fc' : 'rgba(255,255,255,0.2)',
-                            color: isFilterExpanded ? '#c084fc' : '#e4e4e7',
-                          }}
-                          onClick={() => setExpandedFilterBatchId(isFilterExpanded ? null : batch.id)}
-                          disabled={isProcessing}
-                        >
-                          ⚙️ {isFilterExpanded ? 'Hide' : 'Per-Engine'}
-                        </button>
-                      </div>
-
-                      <div style={{ display: 'flex', flexWrap: 'wrap', gap: 4, alignItems: 'center', fontSize: '0.72rem' }}>
-                        <span className="muted" style={{ fontSize: '0.7rem' }}>Classes:</span>
-                        {model1Classes.map((clsName) => {
-                          const isModel1On = activeEngines.includes('base_model') && model1Active.includes(clsName)
-                          const isSam3On = activeEngines.includes('sam3') && sam3Active.includes(clsName)
-                          const isSecOn = activeEngines.includes('secondary_model') && hasSecondaryModel && model2Active.includes(clsName)
-                          const isActiveAny = isModel1On || isSam3On || isSecOn
-
-                          return (
-                            <button
-                              key={clsName}
-                              type="button"
-                              className="tag"
-                              style={{
-                                padding: '1px 5px',
-                                fontSize: '0.7rem',
-                                cursor: 'pointer',
-                                background: isActiveAny ? 'rgba(56, 189, 248, 0.25)' : 'rgba(255,255,255,0.05)',
-                                color: isActiveAny ? '#38bdf8' : '#71717a',
-                                border: isActiveAny ? '1px solid rgba(56, 189, 248, 0.5)' : '1px solid rgba(255,255,255,0.1)',
-                              }}
-                              title={`Toggle ${clsName} filter`}
-                              disabled={isProcessing}
-                              onClick={() => {
-                                if (activeEngines.includes('base_model')) toggleEngineClass(batch.id, 'base_model', clsName, model1Classes)
-                                if (activeEngines.includes('sam3')) toggleEngineClass(batch.id, 'sam3', clsName, model1Classes)
-                                if (activeEngines.includes('secondary_model')) toggleEngineClass(batch.id, 'secondary_model', clsName, model2Classes)
-                              }}
-                            >
-                              {isActiveAny ? '✓ ' : ''}{clsName}
-                            </button>
-                          )
-                        })}
-                      </div>
-                    </div>
-                  </td>
                   <td>
                     <span className={`tag ${batchJob ? 'info' : ''}`}>
                       {batchJob ? `${batchJob.type} (${batchJob.status})` : batch.status}
@@ -278,16 +178,20 @@ function BatchList({ project, batches, activeJobs, onChanged, onError }) {
                   <td>
                     <div className="row" style={{ gap: 6 }}>
                       <button className="btn btn-primary" disabled={isProcessing || batch.frame_count === 0}
-                              onClick={() => autolabel(batch)}>
+                              title="Annotate using project base model with target class selection"
-                        {batchJob?.type === 'autolabel' ? 'Processing…' : `Auto-annotate (${activeEngines.map(e => e === 'sam3' ? 'SAM3' : e === 'base_model' ? 'Model 1' : 'Model 2').join('+')})`}
+                              onClick={() => openBaseModelAutolabelModal(batch)}>
+                        {batchJob?.type === 'autolabel' ? 'Processing…' : 'Auto-annotate'}
+                      </button>
+                      <button className="btn" disabled={isProcessing || batch.frame_count === 0}
+                              title="Append new class detections using custom YOLO model or SAM3 text prompts"
+                              onClick={() => setAppendChoiceBatch(batch)}>
+                        + Append
+                      </button>
+                      <button className="btn" disabled={isProcessing || batch.annotation_count === 0}
+                              title="Clear all auto-generated shapes and reset frame review statuses to pending"
+                              onClick={() => resetAutoAnnotations(batch)}>
+                        🔄 Reset Auto
                       </button>
-                      {batch.annotation_count > 0 && (
-                        <button className="btn" disabled={isProcessing || batch.frame_count === 0}
-                                title="Skip frames that already have automatic shapes"
-                                onClick={() => autolabel(batch, true)}>
-                          Resume
-                        </button>
-                      )}
                       <button className="btn" disabled={batch.frame_count === 0}
                               onClick={() => navigate(`/projects/${project.id}/review?batch=${batch.id}`)}>
                         Review ({batch.reviewed}/{batch.frame_count})
@@ -299,122 +203,438 @@ function BatchList({ project, batches, activeJobs, onChanged, onError }) {
                     </div>
                   </td>
                 </tr>
-
-                {/* Inline Per-Engine Class Filter Panel */}
-                {isFilterExpanded && (
-                  <tr>
-                    <td colSpan={8} style={{ background: '#09090b', padding: '12px 16px', borderBottom: '1px solid rgba(168,85,247,0.3)' }}>
-                      <div style={{ display: 'flex', flexDirection: 'column', gap: 10 }}>
-                        <div style={{ fontSize: '0.82rem', fontWeight: 600, color: '#c084fc' }}>
-                          🛠️ Inline Per-Engine Class Filters (Select target classes per detector):
-                        </div>
-
-                        <div style={{ display: 'grid', gridTemplateColumns: 'repeat(auto-fit, minmax(240px, 1fr))', gap: 12 }}>
-                          {/* SAM3 Section */}
-                          {activeEngines.includes('sam3') && (
-                            <div style={{ background: 'rgba(24, 24, 27, 0.8)', padding: 10, borderRadius: 6, border: '1px solid rgba(255,255,255,0.1)' }}>
-                              <div style={{ fontSize: '0.78rem', fontWeight: 600, color: '#a855f7', marginBottom: 6 }}>
-                                🤖 SAM3 Text Prompts
-                              </div>
-                              <div style={{ display: 'flex', flexWrap: 'wrap', gap: 4 }}>
-                                {model1Classes.map((clsName) => {
-                                  const isActive = sam3Active.includes(clsName)
-                                  return (
-                                    <button
-                                      key={clsName}
-                                      type="button"
-                                      className="tag"
-                                      style={{
-                                        padding: '2px 6px',
-                                        fontSize: '0.72rem',
-                                        cursor: 'pointer',
-                                        background: isActive ? 'rgba(168, 85, 247, 0.3)' : 'rgba(255,255,255,0.05)',
-                                        color: isActive ? '#f3e8ff' : '#666',
-                                        border: isActive ? '1px solid rgba(168, 85, 247, 0.6)' : '1px solid transparent',
-                                      }}
-                                      onClick={() => toggleEngineClass(batch.id, 'sam3', clsName, model1Classes)}
-                                    >
-                                      {isActive ? '✓ ' : ''}{clsName}
-                                    </button>
-                                  )
-                                })}
-                              </div>
-                            </div>
-                          )}
-
-                          {/* Model 1 Section */}
-                          {activeEngines.includes('base_model') && (
-                            <div style={{ background: 'rgba(24, 24, 27, 0.8)', padding: 10, borderRadius: 6, border: '1px solid rgba(255,255,255,0.1)' }}>
-                              <div style={{ fontSize: '0.78rem', fontWeight: 600, color: '#38bdf8', marginBottom: 6 }}>
-                                ⚡ Model 1 (Primary Base) Classes
-                              </div>
-                              <div style={{ display: 'flex', flexWrap: 'wrap', gap: 4 }}>
-                                {model1Classes.map((clsName) => {
-                                  const isActive = model1Active.includes(clsName)
-                                  return (
-                                    <button
-                                      key={clsName}
-                                      type="button"
-                                      className="tag"
-                                      style={{
-                                        padding: '2px 6px',
-                                        fontSize: '0.72rem',
-                                        cursor: 'pointer',
-                                        background: isActive ? 'rgba(56, 189, 248, 0.3)' : 'rgba(255,255,255,0.05)',
-                                        color: isActive ? '#e0f2fe' : '#666',
-                                        border: isActive ? '1px solid rgba(56, 189, 248, 0.6)' : '1px solid transparent',
-                                      }}
-                                      onClick={() => toggleEngineClass(batch.id, 'base_model', clsName, model1Classes)}
-                                    >
-                                      {isActive ? '✓ ' : ''}{clsName}
-                                    </button>
-                                  )
-                                })}
-                              </div>
-                            </div>
-                          )}
-
-                          {/* Model 2 Section */}
-                          {activeEngines.includes('secondary_model') && hasSecondaryModel && (
-                            <div style={{ background: 'rgba(24, 24, 27, 0.8)', padding: 10, borderRadius: 6, border: '1px solid rgba(255,255,255,0.1)' }}>
-                              <div style={{ fontSize: '0.78rem', fontWeight: 600, color: '#f43f5e', marginBottom: 6 }}>
-                                🎯 Model 2 (Secondary Engine) Native Classes
-                              </div>
-                              <div style={{ display: 'flex', flexWrap: 'wrap', gap: 4 }}>
-                                {model2Classes.map((clsName) => {
-                                  const isActive = model2Active.includes(clsName)
-                                  return (
-                                    <button
-                                      key={clsName}
-                                      type="button"
-                                      className="tag"
-                                      style={{
-                                        padding: '2px 6px',
-                                        fontSize: '0.72rem',
-                                        cursor: 'pointer',
-                                        background: isActive ? 'rgba(244, 63, 94, 0.3)' : 'rgba(255,255,255,0.05)',
-                                        color: isActive ? '#ffe4e6' : '#666',
-                                        border: isActive ? '1px solid rgba(244, 63, 94, 0.6)' : '1px solid transparent',
-                                      }}
-                                      onClick={() => toggleEngineClass(batch.id, 'secondary_model', clsName, model2Classes)}
-                                    >
-                                      {isActive ? '✓ ' : ''}{clsName}
-                                    </button>
-                                  )
-                                })}
-                              </div>
-                            </div>
-                          )}
-                        </div>
-                      </div>
-                    </td>
-                  </tr>
-                )}
               </React.Fragment>
             )
           })}
         </tbody>
       </table>
+
+      {/* Choice Modal: SAM3 vs Custom YOLO */}
+      {appendChoiceBatch && (
+        <div style={{
+          position: 'fixed', top: 0, left: 0, right: 0, bottom: 0,
+          background: 'rgba(0,0,0,0.8)', display: 'flex', alignItems: 'center',
+          justifyContent: 'center', zIndex: 9999
+        }}>
+          <div className="panel" style={{ width: 480, maxWidth: '92vw', padding: 22, border: '1px solid rgba(168, 85, 247, 0.4)', background: '#18181b' }}>
+            <h3 style={{ margin: '0 0 6px 0', color: '#c084fc', fontSize: '1.05rem' }}>Select Engine to Append Annotations</h3>
+            <p className="hint" style={{ fontSize: '0.82rem', marginBottom: 18 }}>
+              Choose how you want to detect and append new classes to batch <strong>{appendChoiceBatch.batch_label}</strong>:
+            </p>
+
+            <div style={{ display: 'grid', gridTemplateColumns: '1fr', gap: 12, marginBottom: 20 }}>
+              {/* SAM3 Card */}
+              <div
+                style={{
+                  padding: 14, background: 'rgba(24,24,27,0.9)', borderRadius: 8,
+                  border: '1px solid rgba(168, 85, 247, 0.4)', cursor: 'pointer',
+                  transition: 'all 150ms ease'
+                }}
+                onClick={() => openSam3AppendModal(appendChoiceBatch)}
+              >
+                <div className="row" style={{ justifyContent: 'space-between', marginBottom: 4 }}>
+                  <h4 style={{ margin: 0, color: '#c084fc', fontSize: '0.92rem' }}>🤖 SAM3 Zero-Shot (Text Prompts)</h4>
+                  <span className="btn btn-ghost" style={{ padding: '2px 8px', fontSize: '0.75rem' }}>Select &gt;</span>
+                </div>
+                <p className="hint" style={{ margin: 0, fontSize: '0.78rem' }}>
+                  Detect objects by typing any text prompt (e.g. "box", "sack", "person") without needing a pre-trained model.
+                </p>
+              </div>
+
+              {/* YOLO Card */}
+              <div
+                style={{
+                  padding: 14, background: 'rgba(24,24,27,0.9)', borderRadius: 8,
+                  border: '1px solid rgba(56, 189, 248, 0.4)', cursor: 'pointer',
+                  transition: 'all 150ms ease'
+                }}
+                onClick={() => openFilePickerForYolo(appendChoiceBatch)}
+              >
+                <div className="row" style={{ justifyContent: 'space-between', marginBottom: 4 }}>
+                  <h4 style={{ margin: 0, color: '#38bdf8', fontSize: '0.92rem' }}>⚡ Custom YOLO Model (.pt)</h4>
+                  <span className="btn btn-ghost" style={{ padding: '2px 8px', fontSize: '0.75rem' }}>Upload &gt;</span>
+                </div>
+                <p className="hint" style={{ margin: 0, fontSize: '0.78rem' }}>
+                  Upload a custom trained `.pt` model file from your computer and select which classes to append.
+                </p>
+              </div>
+            </div>
+
+            <div className="row" style={{ justifyContent: 'flex-end' }}>
+              <button className="btn btn-ghost" onClick={() => setAppendChoiceBatch(null)}>Cancel</button>
+            </div>
+          </div>
+        </div>
+      )}
+
+      {/* SAM3 Append Modal */}
+      {sam3AppendState && (
+        <div style={{
+          position: 'fixed', top: 0, left: 0, right: 0, bottom: 0,
+          background: 'rgba(0,0,0,0.8)', display: 'flex', alignItems: 'center',
+          justifyContent: 'center', zIndex: 9999
+        }}>
+          <div className="panel" style={{ width: 500, maxWidth: '92vw', padding: 22, border: '1px solid rgba(168, 85, 247, 0.4)', background: '#18181b' }}>
+            <h3 style={{ margin: '0 0 8px 0', color: '#c084fc', fontSize: '1.05rem' }}>Append Annotations with SAM3</h3>
+            <p className="hint" style={{ fontSize: '0.82rem', marginBottom: 16 }}>
+              Select target text prompts to detect with SAM3:
+            </p>
+
+            <div style={{ marginBottom: 16 }}>
+              <div className="row" style={{ justifyContent: 'space-between', marginBottom: 6 }}>
+                <span className="hint" style={{ fontSize: '0.82rem' }}>Confidence Threshold:</span>
+                <strong style={{ color: '#c084fc', fontSize: '0.85rem' }}>{sam3AppendState.threshold}</strong>
+              </div>
+              <input
+                type="range"
+                min="0.05"
+                max="0.95"
+                step="0.05"
+                value={sam3AppendState.threshold}
+                onChange={(e) => setSam3AppendState({ ...sam3AppendState, threshold: parseFloat(e.target.value) })}
+                style={{ width: '100%', cursor: 'pointer' }}
+              />
+            </div>
+
+            <div style={{ marginBottom: 16 }}>
+              <div className="row" style={{ justifyContent: 'space-between', marginBottom: 6 }}>
+                <span className="hint" style={{ fontSize: '0.82rem' }}>NMS IoU Threshold:</span>
+                <strong style={{ color: '#38bdf8', fontSize: '0.85rem' }}>{sam3AppendState.iouThreshold ?? 0.8}</strong>
+              </div>
+              <input
+                type="range"
+                min="0.1"
+                max="0.9"
+                step="0.05"
+                value={sam3AppendState.iouThreshold ?? 0.8}
+                onChange={(e) => setSam3AppendState({ ...sam3AppendState, iouThreshold: parseFloat(e.target.value) })}
+                style={{ width: '100%', cursor: 'pointer' }}
+              />
+            </div>
+
+            <div style={{ marginBottom: 16 }}>
+              <span className="hint" style={{ fontSize: '0.82rem', display: 'block', marginBottom: 6 }}>Target Prompts to Detect:</span>
+              <div style={{ display: 'flex', flexWrap: 'wrap', gap: 6, maxHeight: 120, overflowY: 'auto', padding: 10, background: 'rgba(0,0,0,0.4)', borderRadius: 6, border: '1px solid rgba(255,255,255,0.1)', marginBottom: 10 }}>
+                {sam3AppendState.selectedClasses.map((clsName) => (
+                  <button
+                    key={clsName}
+                    type="button"
+                    className="tag"
+                    style={{
+                      cursor: 'pointer', padding: '4px 9px', fontSize: '0.8rem',
+                      background: 'rgba(168, 85, 247, 0.25)', color: '#f3e8ff',
+                      border: '1px solid rgba(168, 85, 247, 0.6)'
+                    }}
+                    onClick={() => {
+                      const next = sam3AppendState.selectedClasses.filter((c) => c !== clsName)
+                      setSam3AppendState({ ...sam3AppendState, selectedClasses: next })
+                    }}
+                  >
+                    ✕ {clsName}
+                  </button>
+                ))}
+              </div>
+
+              {/* Add Custom SAM3 Prompt */}
+              <div className="row" style={{ gap: 8 }}>
+                <input
+                  type="text"
+                  placeholder="Type new SAM3 text prompt (e.g. box)..."
+                  value={customPromptInput}
+                  onChange={(e) => setCustomPromptInput(e.target.value)}
+                  onKeyDown={(e) => {
+                    if (e.key === 'Enter' && customPromptInput.trim()) {
+                      const val = customPromptInput.trim().toLowerCase()
+                      if (!sam3AppendState.selectedClasses.includes(val)) {
+                        setSam3AppendState({
+                          ...sam3AppendState,
+                          selectedClasses: [...sam3AppendState.selectedClasses, val]
+                        })
+                      }
+                      setCustomPromptInput('')
+                    }
+                  }}
+                  style={{ flex: 1, padding: '6px 10px', fontSize: '0.82rem', background: '#09090b', border: '1px solid rgba(255,255,255,0.15)', borderRadius: 4, color: '#fff' }}
+                />
+                <button
+                  type="button"
+                  className="btn"
+                  style={{ fontSize: '0.8rem' }}
+                  onClick={() => {
+                    if (customPromptInput.trim()) {
+                      const val = customPromptInput.trim().toLowerCase()
+                      if (!sam3AppendState.selectedClasses.includes(val)) {
+                        setSam3AppendState({
+                          ...sam3AppendState,
+                          selectedClasses: [...sam3AppendState.selectedClasses, val]
+                        })
+                      }
+                      setCustomPromptInput('')
+                    }
+                  }}
+                >
+                  + Add Prompt
+                </button>
+              </div>
+            </div>
+
+            <div className="row" style={{ justifyContent: 'flex-end', gap: 10 }}>
+              <button className="btn btn-ghost" onClick={() => setSam3AppendState(null)}>Cancel</button>
+              <button
+                className="btn btn-primary"
+                disabled={sam3AppendState.selectedClasses.length === 0}
+                onClick={async () => {
+                  const { batch, threshold, iouThreshold, selectedClasses } = sam3AppendState
+                  setSam3AppendState(null)
+                  setBusyId(batch.id)
+                  try {
+                    await api.startAutolabel(batch.id, {
+                      resume: false,
+                      append: true,
+                      engine: 'sam3',
+                      threshold,
+                      iou_threshold: iouThreshold ?? 0.8,
+                      engine_classes: { sam3: selectedClasses },
+                      class_ids: null
+                    })
+                    onChanged()
+                  } catch (err) {
+                    onError(err.message)
+                  } finally {
+                    setBusyId(null)
+                  }
+                }}
+              >
+                Start SAM3 Append
+              </button>
+            </div>
+          </div>
+        </div>
+      )}
+
+      {/* YOLO Custom Model Append Modal */}
+      {appendModalState && (
+        <div style={{
+          position: 'fixed', top: 0, left: 0, right: 0, bottom: 0,
+          background: 'rgba(0,0,0,0.8)', display: 'flex', alignItems: 'center',
+          justifyContent: 'center', zIndex: 9999
+        }}>
+          <div className="panel" style={{ width: 500, maxWidth: '92vw', padding: 22, border: '1px solid rgba(56, 189, 248, 0.4)', background: '#18181b' }}>
+            <h3 style={{ margin: '0 0 8px 0', color: '#38bdf8', fontSize: '1.05rem' }}>Append Annotations with Custom Model</h3>
+            <p className="hint" style={{ fontSize: '0.82rem', marginBottom: 16 }}>
+              Model file: <strong style={{ color: '#fff' }}>{appendModalState.filename}</strong>
+            </p>
+
+            <div style={{ marginBottom: 16 }}>
+              <div className="row" style={{ justifyContent: 'space-between', marginBottom: 6 }}>
+                <span className="hint" style={{ fontSize: '0.82rem' }}>Confidence Threshold:</span>
+                <strong style={{ color: '#38bdf8', fontSize: '0.85rem' }}>{appendModalState.threshold}</strong>
+              </div>
+              <input
+                type="range"
+                min="0.05"
+                max="0.95"
+                step="0.05"
+                value={appendModalState.threshold}
+                onChange={(e) => setAppendModalState({ ...appendModalState, threshold: parseFloat(e.target.value) })}
+                style={{ width: '100%', cursor: 'pointer' }}
+              />
+            </div>
+
+            <div style={{ marginBottom: 16 }}>
+              <div className="row" style={{ justifyContent: 'space-between', marginBottom: 6 }}>
+                <span className="hint" style={{ fontSize: '0.82rem' }}>NMS IoU Threshold:</span>
+                <strong style={{ color: '#c084fc', fontSize: '0.85rem' }}>{appendModalState.iouThreshold ?? 0.8}</strong>
+              </div>
+              <input
+                type="range"
+                min="0.1"
+                max="0.9"
+                step="0.05"
+                value={appendModalState.iouThreshold ?? 0.8}
+                onChange={(e) => setAppendModalState({ ...appendModalState, iouThreshold: parseFloat(e.target.value) })}
+                style={{ width: '100%', cursor: 'pointer' }}
+              />
+            </div>
+
+            <div style={{ marginBottom: 20 }}>
+              <div className="row" style={{ justifyContent: 'space-between', marginBottom: 6 }}>
+                <span className="hint" style={{ fontSize: '0.82rem' }}>Select Classes to Append:</span>
+                <span style={{ fontSize: '0.78rem', color: '#a1a1aa' }}>
+                  {appendModalState.selectedClasses.length} of {appendModalState.classes.length} selected
+                </span>
+              </div>
+              
+              {appendModalState.classes.length === 0 ? (
+                <p className="hint" style={{ fontStyle: 'italic', color: '#e4e4e7' }}>No embedded class names found in model file. All predictions will be appended.</p>
+              ) : (
+                <div style={{ display: 'flex', flexWrap: 'wrap', gap: 6, maxHeight: 150, overflowY: 'auto', padding: 10, background: 'rgba(0,0,0,0.4)', borderRadius: 6, border: '1px solid rgba(255,255,255,0.1)' }}>
+                  {appendModalState.classes.map((clsName) => {
+                    const isChecked = appendModalState.selectedClasses.includes(clsName)
+                    return (
+                      <button
+                        key={clsName}
+                        type="button"
+                        className="tag"
+                        style={{
+                          cursor: 'pointer',
+                          padding: '4px 9px',
+                          fontSize: '0.8rem',
+                          background: isChecked ? 'rgba(56, 189, 248, 0.25)' : 'rgba(255,255,255,0.05)',
+                          color: isChecked ? '#e0f2fe' : '#71717a',
+                          border: isChecked ? '1px solid rgba(56, 189, 248, 0.6)' : '1px solid rgba(255,255,255,0.1)',
+                          transition: 'all 150ms ease'
+                        }}
+                        onClick={() => {
+                          const next = isChecked
+                            ? appendModalState.selectedClasses.filter((c) => c !== clsName)
+                            : [...appendModalState.selectedClasses, clsName]
+                          setAppendModalState({ ...appendModalState, selectedClasses: next })
+                        }}
+                      >
+                        {isChecked ? '✓ ' : ''}{clsName}
+                      </button>
+                    )
+                  })}
+                </div>
+              )}
+            </div>
+
+            <div className="row" style={{ justifyContent: 'flex-end', gap: 10 }}>
+              <button className="btn btn-ghost" onClick={() => setAppendModalState(null)}>Cancel</button>
+              <button
+                className="btn btn-primary"
+                disabled={appendModalState.classes.length > 0 && appendModalState.selectedClasses.length === 0}
+                onClick={async () => {
+                  const { batch, file, threshold, iouThreshold, selectedClasses } = appendModalState
+                  setAppendModalState(null)
+                  setBusyId(batch.id)
+                  try {
+                    await api.autolabelWithModel(batch.id, file, threshold, selectedClasses, iouThreshold ?? 0.8)
+                    onChanged()
+                  } catch (err) {
+                    onError(err.message)
+                  } finally {
+                    setBusyId(null)
+                  }
+                }}
+              >
+                Start Append Auto-Annotation
+              </button>
+            </div>
+          </div>
+        </div>
+      )}
+
+      {/* Base Model Auto-annotate Modal */}
+      {baseModelModalState && (
+        <div style={{
+          position: 'fixed', top: 0, left: 0, right: 0, bottom: 0,
+          background: 'rgba(0,0,0,0.8)', display: 'flex', alignItems: 'center',
+          justifyContent: 'center', zIndex: 9999
+        }}>
+          <div className="panel" style={{ width: 480, maxWidth: '92vw', padding: 22, border: '1px solid rgba(56, 189, 248, 0.4)', background: '#18181b' }}>
+            <h3 style={{ margin: '0 0 6px 0', color: '#38bdf8', fontSize: '1.05rem' }}>Auto-annotate Batch (Base Model)</h3>
+            <p className="hint" style={{ fontSize: '0.82rem', marginBottom: 16 }}>
+              Select target classes to detect using the project's base model:
+            </p>
+
+            <div style={{ marginBottom: 16 }}>
+              <div className="row" style={{ justifyContent: 'space-between', marginBottom: 6 }}>
+                <span className="hint" style={{ fontSize: '0.82rem' }}>Confidence Threshold:</span>
+                <strong style={{ color: '#38bdf8', fontSize: '0.85rem' }}>{baseModelModalState.threshold}</strong>
+              </div>
+              <input
+                type="range"
+                min="0.05"
+                max="0.95"
+                step="0.05"
+                value={baseModelModalState.threshold}
+                onChange={(e) => setBaseModelModalState({ ...baseModelModalState, threshold: parseFloat(e.target.value) })}
+                style={{ width: '100%', cursor: 'pointer' }}
+              />
+            </div>
+
+            <div style={{ marginBottom: 16 }}>
+              <div className="row" style={{ justifyContent: 'space-between', marginBottom: 6 }}>
+                <span className="hint" style={{ fontSize: '0.82rem' }}>NMS IoU Threshold:</span>
+                <strong style={{ color: '#38bdf8', fontSize: '0.85rem' }}>{baseModelModalState.iouThreshold ?? 0.8}</strong>
+              </div>
+              <input
+                type="range"
+                min="0.1"
+                max="0.9"
+                step="0.05"
+                value={baseModelModalState.iouThreshold ?? 0.8}
+                onChange={(e) => setBaseModelModalState({ ...baseModelModalState, iouThreshold: parseFloat(e.target.value) })}
+                style={{ width: '100%', cursor: 'pointer' }}
+              />
+            </div>
+
+            <div style={{ marginBottom: 18 }}>
+              <span className="hint" style={{ fontSize: '0.82rem', display: 'block', marginBottom: 6 }}>Target Classes to Detect:</span>
+              <div style={{ display: 'flex', flexWrap: 'wrap', gap: 6, padding: 10, background: 'rgba(0,0,0,0.4)', borderRadius: 6, border: '1px solid rgba(255,255,255,0.1)' }}>
+                {project?.classes?.map((cls) => {
+                  const isChecked = baseModelModalState.selectedClasses.includes(cls.name)
+                  return (
+                    <button
+                      key={cls.class_id}
+                      type="button"
+                      className="tag"
+                      style={{
+                        cursor: 'pointer',
+                        padding: '4px 9px',
+                        fontSize: '0.8rem',
+                        background: isChecked ? 'rgba(56, 189, 248, 0.25)' : 'rgba(255,255,255,0.05)',
+                        color: isChecked ? '#e0f2fe' : '#71717a',
+                        border: isChecked ? '1px solid rgba(56, 189, 248, 0.6)' : '1px solid rgba(255,255,255,0.1)',
+                      }}
+                      onClick={() => {
+                        const newSel = isChecked
+                          ? baseModelModalState.selectedClasses.filter((n) => n !== cls.name)
+                          : [...baseModelModalState.selectedClasses, cls.name]
+                        setBaseModelModalState({ ...baseModelModalState, selectedClasses: newSel })
+                      }}
+                    >
+                      {isChecked ? '✓ ' : ''}{cls.name}
+                    </button>
+                  )
+                })}
+              </div>
+            </div>
+
+            <div className="row" style={{ justifyContent: 'flex-end', gap: 10 }}>
+              <button className="btn btn-ghost" onClick={() => setBaseModelModalState(null)}>Cancel</button>
+              <button
+                className="btn btn-primary"
+                disabled={baseModelModalState.selectedClasses.length === 0}
+                onClick={async () => {
+                  const { batch, threshold, iouThreshold, selectedClasses } = baseModelModalState
+                  setBaseModelModalState(null)
+                  setBusyId(batch.id)
+                  try {
+                    await api.startAutolabel(batch.id, {
+                      resume: false,
+                      append: false,
+                      engine: 'base_model',
+                      threshold,
+                      iou_threshold: iouThreshold ?? 0.8,
+                      target_class_names: selectedClasses,
+                    })
+                    onChanged()
+                  } catch (err) {
+                    onError(err.message)
+                  } finally {
+                    setBusyId(null)
+                  }
+                }}
+              >
+                Start Auto-Annotation
+              </button>
+            </div>
+          </div>
+        </div>
+      )}
     </div>
   )
 }

frontend/src/pages/ModelsPage.jsx

@@ -91,6 +91,7 @@ export default function ModelsPage({ projectId, onProject }) {
   const [error, setError] = useState('')
 
   const [selectedBatchIds, setSelectedBatchIds] = useState([])
+  const [selectedClassIds, setSelectedClassIds] = useState([])
 
   const load = useCallback(async () => {
     const [loadedProject, loadedSummary, modelPayload, hw, jobsPayload] = await Promise.all([
@@ -103,6 +104,7 @@ export default function ModelsPage({ projectId, onProject }) {
     setModels(modelPayload.models)
     setHardware(hw)
     setSelectedBatchIds(loadedSummary.batches.map((b) => b.id))
+    setSelectedClassIds(loadedProject.classes?.map((c) => c.class_id) || [])
     const activeJob = jobsPayload.jobs?.find((j) => ['running', 'queued'].includes(j.status))
     if (activeJob) setJob(activeJob)
   }, [projectId])
@@ -126,6 +128,7 @@ export default function ModelsPage({ projectId, onProject }) {
       setJob(await api.startTraining(projectId, {
         epochs: Number(epochs),
         batch_ids: selectedBatchIds.length > 0 ? selectedBatchIds : null,
+        class_ids: selectedClassIds.length > 0 ? selectedClassIds : null,
       }))
     } catch (exc) {
       setError(exc.message)
@@ -146,6 +149,12 @@ export default function ModelsPage({ projectId, onProject }) {
     }
   }
 
+  const toggleClassSelect = (classId) => {
+    setSelectedClassIds((prev) =>
+      prev.includes(classId) ? prev.filter((cId) => cId !== classId) : [...prev, classId]
+    )
+  }
+
   if (error && !project) return <p className="error-banner"><AlertIcon size={14} /> {error}</p>
   if (!project || !summary) return <p className="empty">Loading…</p>
 
@@ -169,103 +178,40 @@ export default function ModelsPage({ projectId, onProject }) {
         <AlertIcon size={14} /> {error}
       </p>}
 
-      <div className="select-engine-section">
-        <h2>Select Engine</h2>
-        <p className="hint">
-          Select how you want to train your model. Configure custom fine-tuning parameters or use SAM3 grounding backbone.
-        </p>
-
-        <div className="engine-grid">
-          <div className="engine-card selected">
-            <div className="engine-card-header">
-              <span className="project-badge">Selected</span>
-              <span className="engine-card-title">Custom Training (YOLO11)</span>
-            </div>
-            <p className="engine-card-desc">
-              Fine-tune on the merged master dataset using pre-configured hardware batch size and image resolution.
-            </p>
-          </div>
-
-          <div className="engine-card" title="Neural Architecture Search / Rapid Auto-train">
-            <div className="engine-card-header">
-              <span className="chip-badge info">Rapid NAS</span>
-              <span className="engine-card-title">Neural Architecture Search</span>
-            </div>
-            <p className="engine-card-desc">
-              Automated model selection optimized for latency and accuracy trade-offs on your specific project dataset.
-            </p>
-          </div>
-        </div>
-      </div>
-
       <div className="panel side-panel" style={{ marginBottom: 20 }}>
-        <h2>Uploaded Models for Training & Auto-Annotation</h2>
+        <h2>Base Model Configuration</h2>
-        <p className="hint">Upload up to 2 models to use for fine-tuning baseline or auto-annotating new video batches:</p>
+        <p className="hint">Upload a base YOLO model checkpoint (.pt) to use for fine-tuning baseline and auto-annotation:</p>
         
-        <div style={{ display: 'grid', gridTemplateColumns: 'repeat(auto-fit, minmax(280px, 1fr))', gap: 16, marginTop: 12 }}>
+        <div style={{ padding: 14, background: 'rgba(0,0,0,0.3)', borderRadius: 8, border: '1px solid rgba(255,255,255,0.1)', marginTop: 12 }}>
-          <div style={{ padding: 14, background: 'rgba(0,0,0,0.3)', borderRadius: 8, border: '1px solid rgba(255,255,255,0.1)' }}>
+          <h3 style={{ fontSize: '0.9rem', color: '#c084fc', margin: '0 0 6px 0' }}>Base Model</h3>
-            <h3 style={{ fontSize: '0.9rem', color: '#c084fc', margin: '0 0 6px 0' }}>Primary Model 1 (Base Model)</h3>
+          <p className="hint" style={{ fontSize: '0.78rem', margin: '0 0 10px 0' }}>
-            <p className="hint" style={{ fontSize: '0.78rem', margin: '0 0 10px 0' }}>
+            Used as fine-tuning starting point and baseline benchmark.
-              Used as fine-tuning starting point and benchmark baseline.
+          </p>
-            </p>
+          <div style={{ fontSize: '0.8rem', color: '#a1a1aa', marginBottom: 10 }}>
-            <div style={{ fontSize: '0.8rem', color: '#a1a1aa', marginBottom: 10 }}>
+            Status: <strong style={{ color: '#fff' }}>{project.base_model_path ? 'Custom model.pt loaded' : 'Default yolo11n.pt'}</strong>
-              Status: <strong style={{ color: '#fff' }}>{project.base_model_path ? 'Custom model.pt loaded' : 'Default yolo11n.pt'}</strong>
-            </div>
-            <div style={{ fontSize: '0.78rem', color: '#38bdf8', marginBottom: 10 }}>
-              <strong>Model 1 Classes:</strong> {project.classes?.map((c) => c.name).join(', ')}
-            </div>
-            <input
-              type="file"
-              accept=".pt"
-              id="upload-primary-model"
-              style={{ display: 'none' }}
-              onChange={async (e) => {
-                const file = e.target.files?.[0]
-                if (!file) return
-                try {
-                  await api.uploadBaseModel(projectId, file)
-                  load()
-                } catch (err) {
-                  setError(err.message)
-                }
-              }}
-            />
-            <label htmlFor="upload-primary-model" className="btn btn-ghost" style={{ cursor: 'pointer', padding: '4px 10px', fontSize: '0.8rem' }}>
-              Upload Primary Model 1 (.pt)
-            </label>
           </div>
-
+          <div style={{ fontSize: '0.78rem', color: '#38bdf8', marginBottom: 10 }}>
-          <div style={{ padding: 14, background: 'rgba(0,0,0,0.3)', borderRadius: 8, border: '1px solid rgba(255,255,255,0.1)' }}>
+            <strong>Base Model Classes:</strong> {project.classes?.map((c) => c.name).join(', ')}
-            <h3 style={{ fontSize: '0.9rem', color: '#38bdf8', margin: '0 0 6px 0' }}>Secondary Model 2 (Auto-Annotate Engine)</h3>
-            <p className="hint" style={{ fontSize: '0.78rem', margin: '0 0 10px 0' }}>
-              Used as an auxiliary engine choice for fast auto-annotation.
-            </p>
-            <div style={{ fontSize: '0.8rem', color: '#a1a1aa', marginBottom: 10 }}>
-              Status: <strong style={{ color: '#fff' }}>{project.secondary_model_path ? (project.secondary_model_name || 'secondary_model.pt') : 'None uploaded'}</strong>
-            </div>
-            <div style={{ fontSize: '0.78rem', color: '#c084fc', marginBottom: 10 }}>
-              <strong>Model 2 Native Classes:</strong> {project.secondary_model_classes?.length > 0 ? project.secondary_model_classes.join(', ') : 'Extracted on upload'}
-            </div>
-            <input
-              type="file"
-              accept=".pt"
-              id="upload-secondary-model"
-              style={{ display: 'none' }}
-              onChange={async (e) => {
-                const file = e.target.files?.[0]
-                if (!file) return
-                try {
-                  await api.uploadSecondaryModel(projectId, file)
-                  load()
-                } catch (err) {
-                  setError(err.message)
-                }
-              }}
-            />
-            <label htmlFor="upload-secondary-model" className="btn btn-ghost" style={{ cursor: 'pointer', padding: '4px 10px', fontSize: '0.8rem' }}>
-              Upload Secondary Model 2 (.pt)
-            </label>
           </div>
+          <input
+            type="file"
+            accept=".pt"
+            id="upload-primary-model"
+            style={{ display: 'none' }}
+            onChange={async (e) => {
+              const file = e.target.files?.[0]
+              if (!file) return
+              try {
+                await api.uploadBaseModel(projectId, file)
+                load()
+              } catch (err) {
+                setError(err.message)
+              }
+            }}
+          />
+          <label htmlFor="upload-primary-model" className="btn btn-ghost" style={{ cursor: 'pointer', padding: '4px 10px', fontSize: '0.8rem' }}>
+            Upload Base Model (.pt)
+          </label>
         </div>
       </div>
 
@@ -291,6 +237,34 @@ export default function ModelsPage({ projectId, onProject }) {
                 : `${project.base_model_fallback} (no base model uploaded)`}{' '}
               on the selected dataset batches ({selectedBatchIds.length}/{summary.batches.length} selected).
             </p>
+
+            <div style={{ marginBottom: 12 }}>
+              <label style={{ fontSize: '0.82rem', display: 'block', marginBottom: 6 }}>Target Classes to Train:</label>
+              <div style={{ display: 'flex', flexWrap: 'wrap', gap: 6 }}>
+                {project.classes?.map((cls) => {
+                  const isChecked = selectedClassIds.includes(cls.class_id)
+                  return (
+                    <button
+                      key={cls.class_id}
+                      type="button"
+                      className="tag"
+                      style={{
+                        cursor: 'pointer',
+                        padding: '4px 9px',
+                        fontSize: '0.8rem',
+                        background: isChecked ? 'rgba(192, 132, 252, 0.25)' : 'rgba(255,255,255,0.05)',
+                        color: isChecked ? '#f3e8ff' : '#71717a',
+                        border: isChecked ? '1px solid rgba(192, 132, 252, 0.6)' : '1px solid rgba(255,255,255,0.1)',
+                      }}
+                      onClick={() => toggleClassSelect(cls.class_id)}
+                    >
+                      {isChecked ? '✓ ' : ''}{cls.name}
+                    </button>
+                  )
+                })}
+              </div>
+            </div>
+
             <div>
               <label htmlFor="epochs">Epochs</label>
               <input id="epochs" type="number" min="1" max="500" value={epochs}
@@ -303,7 +277,7 @@ export default function ModelsPage({ projectId, onProject }) {
               </p>
             )}
             <button className="btn btn-primary" onClick={train}
-                    disabled={running || summary.splits.train === 0 || selectedBatchIds.length === 0}>
+                    disabled={running || summary.splits.train === 0 || selectedBatchIds.length === 0 || selectedClassIds.length === 0}>
               {running ? 'Training…' : 'Start training'}
             </button>
             {summary.splits.train === 0 && (
@@ -312,6 +286,9 @@ export default function ModelsPage({ projectId, onProject }) {
             {selectedBatchIds.length === 0 && summary.splits.train > 0 && (
               <p className="hint" style={{ color: '#ef4444' }}>Select at least one dataset batch to train.</p>
             )}
+            {selectedClassIds.length === 0 && (
+              <p className="hint" style={{ color: '#ef4444' }}>Select at least one class to train.</p>
+            )}
           </div>
 
           {job && (

frontend/src/pages/Sam3PlaygroundPage.jsx

@@ -0,0 +1,328 @@
+import React, { useState, useRef } from 'react'
+import { api } from '../api'
+
+const COLORS = [
+  '#38bdf8', '#c084fc', '#f43f5e', '#34d399', '#fbbf24',
+  '#818cf8', '#f472b6', '#a3e635', '#2dd4bf', '#fb923c'
+]
+
+function getColor(index) {
+  return COLORS[index % COLORS.length]
+}
+
+export default function Sam3PlaygroundPage() {
+  const [imageFile, setImageFile] = useState(null)
+  const [previewUrl, setPreviewUrl] = useState(null)
+  const [promptsText, setPromptsText] = useState('white plastic bag, forklift, pallet')
+  const [threshold, setThreshold] = useState(0.35)
+  const [iouThreshold, setIouThreshold] = useState(0.8)
+  const [loading, setLoading] = useState(false)
+  const [error, setError] = useState('')
+  const [result, setResult] = useState(null)
+  const fileInputRef = useRef(null)
+
+  const handleFileSelect = (file) => {
+    if (!file) return
+    setImageFile(file)
+    setPreviewUrl(URL.createObjectURL(file))
+    setResult(null)
+    setError('')
+  }
+
+  const handleDrop = (e) => {
+    e.preventDefault()
+    const file = e.dataTransfer?.files?.[0]
+    if (file && file.type.startsWith('image/')) {
+      handleFileSelect(file)
+    }
+  }
+
+  const runTest = async () => {
+    if (!imageFile) {
+      setError('Please select or upload an image first.')
+      return
+    }
+    if (!promptsText.trim()) {
+      setError('Please enter at least one text prompt.')
+      return
+    }
+    setLoading(true)
+    setError('')
+    try {
+      const payload = await api.sam3PlaygroundTest(imageFile, promptsText, threshold, iouThreshold)
+      setResult(payload)
+    } catch (err) {
+      setError(err.message || 'SAM3 test failed.')
+    } finally {
+      setLoading(false)
+    }
+  }
+
+  return (
+    <div className="roboflow-page" style={{ padding: 24, maxWidth: 1400, margin: '0 auto' }}>
+      <div className="page-header" style={{ marginBottom: 20 }}>
+        <div>
+          <h1 style={{ margin: '0 0 6px 0', fontSize: '1.4rem', color: '#c084fc', display: 'flex', alignItems: 'center', gap: 8 }}>
+            🤖 SAM3 Global Playground
+          </h1>
+          <p className="hint" style={{ margin: 0, fontSize: '0.85rem' }}>
+            Test SAM3 zero-shot text grounding prompts on any image without affecting projects or database records.
+          </p>
+        </div>
+      </div>
+
+      {error && (
+        <div className="alert error" style={{ marginBottom: 16, padding: '10px 14px', borderRadius: 6, background: 'rgba(239, 68, 68, 0.15)', border: '1px solid rgba(239, 68, 68, 0.4)', color: '#fca5a5', fontSize: '0.85rem' }}>
+          ⚠️ {error}
+        </div>
+      )}
+
+      <div style={{ display: 'grid', gridTemplateColumns: '360px 1fr', gap: 20, alignItems: 'start' }}>
+        {/* Left Control Panel */}
+        <div className="panel" style={{ padding: 18, background: '#18181b', borderRadius: 8, border: '1px solid rgba(255,255,255,0.1)' }}>
+          <h3 style={{ margin: '0 0 14px 0', fontSize: '0.95rem', color: '#fff' }}>Configuration</h3>
+
+          {/* Image Selection / Dropzone */}
+          <div style={{ marginBottom: 16 }}>
+            <label className="hint" style={{ fontSize: '0.82rem', display: 'block', marginBottom: 6 }}>
+              Select Test Image:
+            </label>
+            <div
+              onDragOver={(e) => e.preventDefault()}
+              onDrop={handleDrop}
+              onClick={() => fileInputRef.current?.click()}
+              style={{
+                border: '2px dashed rgba(192, 132, 252, 0.4)',
+                borderRadius: 8,
+                padding: '16px 12px',
+                textAlign: 'center',
+                cursor: 'pointer',
+                background: 'rgba(0,0,0,0.3)',
+                transition: 'all 150ms ease'
+              }}
+            >
+              <input
+                ref={fileInputRef}
+                type="file"
+                accept="image/*"
+                style={{ display: 'none' }}
+                onChange={(e) => handleFileSelect(e.target.files?.[0])}
+              />
+              {imageFile ? (
+                <div style={{ fontSize: '0.82rem', color: '#38bdf8', wordBreak: 'break-all' }}>
+                  🖼️ {imageFile.name} ({(imageFile.size / 1024).toFixed(0)} KB)
+                </div>
+              ) : (
+                <div style={{ fontSize: '0.82rem', color: '#a1a1aa' }}>
+                  Click to choose image or drag & drop here (.jpg, .png)
+                </div>
+              )}
+            </div>
+          </div>
+
+          {/* Text Prompts Input */}
+          <div style={{ marginBottom: 16 }}>
+            <label className="hint" style={{ fontSize: '0.82rem', display: 'block', marginBottom: 6 }}>
+              Descriptive Text Prompts (Comma Separated):
+            </label>
+            <textarea
+              rows={3}
+              value={promptsText}
+              onChange={(e) => setPromptsText(e.target.value)}
+              placeholder="e.g. white plastic bag, forklift, pallet"
+              style={{
+                width: '100%',
+                padding: '8px 10px',
+                fontSize: '0.83rem',
+                background: '#09090b',
+                border: '1px solid rgba(255,255,255,0.15)',
+                borderRadius: 6,
+                color: '#fff',
+                resize: 'vertical'
+              }}
+            />
+            <span className="hint" style={{ fontSize: '0.75rem', color: '#71717a' }}>
+              Tip: Use specific descriptive phrases for custom objects (e.g. "woven white bag" instead of "sack").
+            </span>
+          </div>
+
+          {/* Threshold Sliders */}
+          <div style={{ marginBottom: 16 }}>
+            <div className="row" style={{ justifyContent: 'space-between', marginBottom: 4 }}>
+              <span className="hint" style={{ fontSize: '0.82rem' }}>Confidence Threshold:</span>
+              <strong style={{ color: '#c084fc', fontSize: '0.85rem' }}>{threshold}</strong>
+            </div>
+            <input
+              type="range"
+              min="0.05"
+              max="0.95"
+              step="0.05"
+              value={threshold}
+              onChange={(e) => setThreshold(parseFloat(e.target.value))}
+              style={{ width: '100%', cursor: 'pointer' }}
+            />
+          </div>
+
+          <div style={{ marginBottom: 20 }}>
+            <div className="row" style={{ justifyContent: 'space-between', marginBottom: 4 }}>
+              <span className="hint" style={{ fontSize: '0.82rem' }}>NMS IoU Threshold:</span>
+              <strong style={{ color: '#38bdf8', fontSize: '0.85rem' }}>{iouThreshold}</strong>
+            </div>
+            <input
+              type="range"
+              min="0.1"
+              max="0.9"
+              step="0.05"
+              value={iouThreshold}
+              onChange={(e) => setIouThreshold(parseFloat(e.target.value))}
+              style={{ width: '100%', cursor: 'pointer' }}
+            />
+          </div>
+
+          {/* Run Button */}
+          <button
+            className="btn btn-primary"
+            disabled={loading || !imageFile}
+            onClick={runTest}
+            style={{
+              width: '100%',
+              padding: '10px 14px',
+              fontSize: '0.9rem',
+              fontWeight: 600,
+              background: '#c084fc',
+              borderColor: '#c084fc',
+              color: '#000',
+              cursor: loading || !imageFile ? 'not-allowed' : 'pointer'
+            }}
+          >
+            {loading ? 'Running SAM3 Engine…' : '🚀 Run SAM3 Test'}
+          </button>
+
+          {/* Detection Stats List */}
+          {result && (
+            <div style={{ marginTop: 20, paddingTop: 16, borderTop: '1px solid rgba(255,255,255,0.1)' }}>
+              <h4 style={{ margin: '0 0 10px 0', fontSize: '0.88rem', color: '#34d399' }}>
+                Results: {result.detections?.length || 0} Object(s) Found
+              </h4>
+              <div style={{ display: 'flex', flexDirection: 'column', gap: 6, maxHeight: 220, overflowY: 'auto' }}>
+                {result.detections.map((det, i) => {
+                  const color = getColor(det.class_id)
+                  return (
+                    <div
+                      key={i}
+                      style={{
+                        padding: '6px 10px',
+                        background: 'rgba(0,0,0,0.4)',
+                        borderRadius: 6,
+                        borderLeft: `4px solid ${color}`,
+                        fontSize: '0.78rem',
+                        display: 'flex',
+                        justifyContent: 'space-between'
+                      }}
+                    >
+                      <strong style={{ color }}>{det.prompt}</strong>
+                      <span style={{ color: '#a1a1aa' }}>{(det.score * 100).toFixed(1)}%</span>
+                    </div>
+                  )
+                })}
+              </div>
+            </div>
+          )}
+        </div>
+
+        {/* Right Canvas Viewer */}
+        <div className="panel" style={{ padding: 18, background: '#18181b', borderRadius: 8, border: '1px solid rgba(255,255,255,0.1)', minHeight: 500 }}>
+          {!previewUrl ? (
+            <div style={{ display: 'flex', flexDirection: 'column', alignItems: 'center', justifyContent: 'center', minHeight: 450, color: '#71717a' }}>
+              <div style={{ fontSize: '2.5rem', marginBottom: 10 }}>🖼️</div>
+              <p style={{ margin: 0, fontSize: '0.9rem' }}>Upload an image on the left to start testing SAM3 text prompts.</p>
+            </div>
+          ) : (
+            <div style={{ position: 'relative', width: '100%', overflow: 'hidden', display: 'flex', justifyContent: 'center', background: '#09090b', borderRadius: 6 }}>
+              <img
+                src={previewUrl}
+                alt="SAM3 Test Target"
+                style={{ maxWidth: '100%', maxHeight: '75vh', objectFit: 'contain', display: 'block' }}
+              />
+
+              {/* Overlay Canvas SVG */}
+              {result && result.width && result.height && (
+                <svg
+                  viewBox={`0 0 ${result.width} ${result.height}`}
+                  style={{
+                    position: 'absolute',
+                    top: 0, left: 0, width: '100%', height: '100%',
+                    pointerEvents: 'none'
+                  }}
+                >
+                  {result.detections.map((det, i) => {
+                    const color = getColor(det.class_id)
+                    const [x0, y0, x1, y1] = [
+                      det.box[0] * result.width,
+                      det.box[1] * result.height,
+                      det.box[2] * result.width,
+                      det.box[3] * result.height
+                    ]
+                    const bw = x1 - x0
+                    const bh = y1 - y0
+
+                    return (
+                      <g key={i}>
+                        {/* Polygon Masks */}
+                        {det.polygons?.map((poly, pIdx) => {
+                          const ptsString = poly.map(([px, py]) => `${px * result.width},${py * result.height}`).join(' ')
+                          return (
+                            <polygon
+                              key={pIdx}
+                              points={ptsString}
+                              fill={color}
+                              fillOpacity={0.35}
+                              stroke={color}
+                              strokeWidth={2}
+                            />
+                          )
+                        })}
+
+                        {/* Bounding Box */}
+                        <rect
+                          x={x0}
+                          y={y0}
+                          width={bw}
+                          height={bh}
+                          fill="none"
+                          stroke={color}
+                          strokeWidth={2}
+                          strokeDasharray="4 2"
+                        />
+
+                        {/* Label Badge */}
+                        <rect
+                          x={x0}
+                          y={Math.max(0, y0 - 22)}
+                          width={Math.max(80, det.prompt.length * 8 + 45)}
+                          height={20}
+                          fill={color}
+                          rx={3}
+                        />
+                        <text
+                          x={x0 + 6}
+                          y={Math.max(14, y0 - 8)}
+                          fill="#000"
+                          fontSize={12}
+                          fontWeight="bold"
+                          fontFamily="sans-serif"
+                        >
+                          {det.prompt} ({Math.round(det.score * 100)}%)
+                        </text>
+                      </g>
+                    )
+                  })}
+                </svg>
+              )}
+            </div>
+          )}
+        </div>
+      </div>
+    </div>
+  )
+}

scripts/restart_app.sh

@@ -7,17 +7,18 @@ cd "$REPO_DIR"
 echo "🔄 Restarting app services..."
 
 # Kill running uvicorn and vite processes
+fuser -k -9 8000/tcp 8080/tcp 5173/tcp || true
 pkill -f "uvicorn backend.main:app" || true
 pkill -f "vite" || true
-sleep 1
+sleep 2.5
 
 export PATH="$HOME/.local/bin:$PATH"
 
 # Run backend
-nohup uv run uvicorn backend.main:app --host 0.0.0.0 --port 8000 > "$REPO_DIR/data/backend.log" 2>&1 &
+setsid uv run uvicorn backend.main:app --host 0.0.0.0 --port 8000 > "$REPO_DIR/data/backend.log" 2>&1 &
 
 # Run frontend
 cd "$REPO_DIR/frontend"
-nohup npm run dev > "$REPO_DIR/data/frontend.log" 2>&1 &
+setsid npm run dev -- --host 0.0.0.0 --port 8080 > "$REPO_DIR/data/frontend.log" 2>&1 &
 
 echo "✅ Services restarted."