matrice-analytics 0.1.60__py3-none-any.whl

matrice_analytics/post_processing/usecases/warehouse_object_segmentation.py

@@ -0,0 +1,899 @@
+"""
+warehouse object Monitoring Use Case for Post-Processing
+
+This module provides warehouse object monitoring functionality ,
+zone analysis, and alert generation.
+
+"""
+
+from typing import Any, Dict, List, Optional
+from dataclasses import asdict
+import time
+from datetime import datetime, timezone
+import copy  # Added for deep copying detections to preserve original masks
+
+from ..core.base import BaseProcessor, ProcessingContext, ProcessingResult, ConfigProtocol, ResultFormat
+from ..utils import (
+    filter_by_confidence,
+    filter_by_categories,
+    apply_category_mapping,
+    count_objects_by_category,
+    count_objects_in_zones,
+    calculate_counting_summary,
+    match_results_structure,
+    bbox_smoothing,
+    BBoxSmoothingConfig,
+    BBoxSmoothingTracker
+)
+from dataclasses import dataclass, field
+from ..core.config import BaseConfig, AlertConfig, ZoneConfig
+
+
+@dataclass
+class WarehouseObjectConfig(BaseConfig):
+    """Configuration for Warehouse pallet detection use case in warehouse pallet monitoring."""
+    # Smoothing configuration
+    enable_smoothing: bool = True
+    smoothing_algorithm: str = "observability"  # "window" or "observability"
+    smoothing_window_size: int = 20
+    smoothing_cooldown_frames: int = 5
+    smoothing_confidence_range_factor: float = 0.5
+
+    #confidence thresholds
+    confidence_threshold: float = 0.90
+
+    usecase_categories: List[str] = field(
+        default_factory=lambda: ['pallet']
+    )
+
+    target_categories: List[str] = field(
+        default_factory=lambda: ['pallet']
+    )
+
+    alert_config: Optional[AlertConfig] = None
+
+    index_to_category: Optional[Dict[int, str]] = field(
+        default_factory=lambda: {
+            0: "pallet"
+        }
+    )
+
+
+class WarehouseObjectUseCase(BaseProcessor):
+
+    # Human-friendly display names for  categories
+    CATEGORY_DISPLAY = {
+        "pallet": "pallet"
+    }
+    def __init__(self):
+        super().__init__("warehouse_object_segmentation")
+        self.category = "retail"
+
+        # List of  categories to track
+        self.target_categories = ["pallet"]
+
+        self.CASE_TYPE: Optional[str] = 'warehouse_object_segmentation'
+        self.CASE_VERSION: Optional[str] = '1.3'
+
+        # Initialize smoothing tracker
+        self.smoothing_tracker = None
+
+        # Initialize advanced tracker (will be created on first use)
+        self.tracker = None
+
+        # Initialize tracking state variables
+        self._total_frame_counter = 0
+        self._global_frame_offset = 0
+
+        # Track start time for "TOTAL SINCE" calculation
+        self._tracking_start_time = None
+
+        # ------------------------------------------------------------------ #
+        # Canonical tracking aliasing to avoid duplicate counts              #
+        # ------------------------------------------------------------------ #
+        # Maps raw tracker-generated IDs to stable canonical IDs that persist
+        # even if the underlying tracker re-assigns a new ID after a short
+        # interruption. This mirrors the logic used in people_counting to
+        # provide accurate unique counting.
+        self._track_aliases: Dict[Any, Any] = {}
+        self._canonical_tracks: Dict[Any, Dict[str, Any]] = {}
+        # Tunable parameters – adjust if necessary for specific scenarios
+        self._track_merge_iou_threshold: float = 0.05  # IoU ≥ 0.05 →
+        self._track_merge_time_window: float = 7.0  # seconds within which to merge
+
+        self._ascending_alert_list: List[int] = []
+        self.current_incident_end_timestamp: str = "N/A"
+
+    def process(self, data: Any, config: ConfigProtocol, context: Optional[ProcessingContext] = None,
+                stream_info: Optional[Dict[str, Any]] = None) -> ProcessingResult:
+        """
+        Main entry point for post-processing.
+        Applies category mapping, smoothing, counting, alerting, and summary generation.
+        Returns a ProcessingResult with all relevant outputs.
+        """
+        start_time = time.time()
+        # Ensure config is correct type
+        if not isinstance(config, WarehouseObjectConfig):
+            return self.create_error_result("Invalid config type", usecase=self.name, category=self.category,
+                                            context=context)
+        if context is None:
+            context = ProcessingContext()
+
+        # Detect input format and store in context
+        input_format = match_results_structure(data)
+        context.input_format = input_format
+        context.confidence_threshold = config.confidence_threshold
+        
+        # Step 1: Confidence filtering
+        if config.confidence_threshold is not None:
+            processed_data = filter_by_confidence(data, config.confidence_threshold)
+        else:
+            processed_data = data
+            self.logger.debug(f"Did not apply confidence filtering with threshold since nothing was provided")
+
+        # Step 2: Apply category mapping if provided
+        if config.index_to_category:
+            processed_data = apply_category_mapping(processed_data, config.index_to_category)
+
+        # Step 3: Category filtering
+        if config.target_categories:
+            processed_data = [d for d in processed_data if d.get('category') in self.target_categories]
+
+        # Step 4: Apply bbox smoothing if enabled
+        # Deep-copy detections so that we preserve the original masks before any
+        # smoothing/tracking logic potentially removes them.
+        raw_processed_data = [copy.deepcopy(det) for det in processed_data]
+        if config.enable_smoothing:
+            if self.smoothing_tracker is None:
+                smoothing_config = BBoxSmoothingConfig(
+                    smoothing_algorithm=config.smoothing_algorithm,
+                    window_size=config.smoothing_window_size,
+                    cooldown_frames=config.smoothing_cooldown_frames,
+                    confidence_threshold=config.confidence_threshold,
+                    confidence_range_factor=config.smoothing_confidence_range_factor,
+                    enable_smoothing=True
+                )
+                self.smoothing_tracker = BBoxSmoothingTracker(smoothing_config)
+           
+            processed_data = bbox_smoothing(processed_data, self.smoothing_tracker.config, self.smoothing_tracker)
+            # Restore masks after smoothing
+
+        # Step 5: Advanced tracking (BYTETracker-like)
+        try:
+            from ..advanced_tracker import AdvancedTracker
+            from ..advanced_tracker.config import TrackerConfig
+
+            # Create tracker instance if it doesn't exist (preserves state across frames)
+            if self.tracker is None:
+                tracker_config = TrackerConfig()
+                self.tracker = AdvancedTracker(tracker_config)
+                self.logger.info("Initialized AdvancedTracker for Monitoring and tracking")
+           
+            processed_data = self.tracker.update(processed_data)
+        except Exception as e:
+            # If advanced tracker fails, fallback to unsmoothed detections
+            self.logger.warning(f"AdvancedTracker failed: {e}")
+
+        # Update tracking state for total count per label
+        self._update_tracking_state(processed_data)
+
+        # ------------------------------------------------------------------ #
+        # Re-attach segmentation masks that were present in the original input
+        # but may have been stripped during smoothing/tracking. We match each
+        # processed detection back to the raw detection with the highest IoU
+        # and copy over its "masks" field (if available).
+        # ------------------------------------------------------------------ #
+        processed_data = self._attach_masks_to_detections(processed_data, raw_processed_data)
+
+        # Update frame counter
+        self._total_frame_counter += 1
+
+        # Extract frame information from stream_info
+        frame_number = None
+        if stream_info:
+            input_settings = stream_info.get("input_settings", {})
+            start_frame = input_settings.get("start_frame")
+            end_frame = input_settings.get("end_frame")
+            # If start and end frame are the same, it's a single frame
+            if start_frame is not None and end_frame is not None and start_frame == end_frame:
+                frame_number = start_frame
+
+        # Compute summaries and alerts
+        general_counting_summary = calculate_counting_summary(data)
+        counting_summary = self._count_categories(processed_data, config)
+        # Add total unique counts after tracking using only local state
+        total_counts = self.get_total_counts()
+        counting_summary['total_counts'] = total_counts
+
+        alerts = self._check_alerts(counting_summary, frame_number, config)
+        predictions = self._extract_predictions(processed_data)
+        
+        # Step: Generate structured events and tracking stats with frame-based keys
+        incidents_list = self._generate_incidents(counting_summary, alerts, config, frame_number, stream_info)
+        tracking_stats_list = self._generate_tracking_stats(counting_summary, alerts, config, frame_number,stream_info)
+        # business_analytics_list = self._generate_business_analytics(counting_summary, alerts, config, frame_number, stream_info, is_empty=False)
+        business_analytics_list = []
+        summary_list = self._generate_summary(counting_summary, incidents_list, tracking_stats_list, business_analytics_list, alerts)
+
+        # Extract frame-based dictionaries from the lists
+        incidents = incidents_list[0] if incidents_list else {}
+        tracking_stats = tracking_stats_list[0] if tracking_stats_list else {}
+        business_analytics = business_analytics_list[0] if business_analytics_list else {}
+        summary = summary_list[0] if summary_list else {}
+        agg_summary = {str(frame_number): {
+                        "incidents": incidents,
+                        "tracking_stats": tracking_stats,
+                        "business_analytics": business_analytics,
+                        "alerts": alerts,
+                        "human_text": summary}
+                    }
+       
+        context.mark_completed()
+
+        # Build result object following the new pattern
+
+        result = self.create_result(
+            data={"agg_summary": agg_summary},
+            usecase=self.name,
+            category=self.category,
+            context=context
+        )
+        
+        return result
+
+    def _check_alerts(self, summary: dict, frame_number: Any, config: WarehouseObjectConfig) -> List[Dict]:
+        """
+        Check if any alert thresholds are exceeded and return alert dicts.
+        """
+        def get_trend(data, lookback=900, threshold=0.6):
+            '''
+            Determine if the trend is ascending or descending based on actual value progression.
+            Now works with values 0,1,2,3 (not just binary).
+            '''
+            window = data[-lookback:] if len(data) >= lookback else data
+            if len(window) < 2:
+                return True  # not enough data to determine trend
+            increasing = 0
+            total = 0
+            for i in range(1, len(window)):
+                if window[i] >= window[i - 1]:
+                    increasing += 1
+                total += 1
+            ratio = increasing / total
+            if ratio >= threshold:
+                return True
+            elif ratio <= (1 - threshold):
+                return False
+
+        frame_key = str(frame_number) if frame_number is not None else "current_frame"
+        alerts = []
+        total_detections = summary.get("total_count", 0) #CURRENT combined total count of all classes
+        total_counts_dict = summary.get("total_counts", {}) #TOTAL cumulative counts per class
+        cumulative_total = sum(total_counts_dict.values()) if total_counts_dict else 0 #TOTAL combined cumulative count
+        per_category_count = summary.get("per_category_count", {}) #CURRENT count per class
+
+        if not config.alert_config:
+            return alerts
+
+        total = summary.get("total_count", 0)
+        #self._ascending_alert_list
+        if hasattr(config.alert_config, 'count_thresholds') and config.alert_config.count_thresholds:
+
+            for category, threshold in config.alert_config.count_thresholds.items():
+                if category == "all" and total > threshold:  
+
+                    alerts.append({
+                        "alert_type": getattr(config.alert_config, 'alert_type', ['Default']) if hasattr(config.alert_config, 'alert_type') else ['Default'],
+                        "alert_id": "alert_"+category+'_'+frame_key,
+                        "incident_category": self.CASE_TYPE,
+                        "threshold_level": threshold,
+                        "ascending": get_trend(self._ascending_alert_list, lookback=900, threshold=0.8),
+                        "settings": {t: v for t, v in zip(getattr(config.alert_config, 'alert_type', ['Default']) if hasattr(config.alert_config, 'alert_type') else ['Default'],
+                                    getattr(config.alert_config, 'alert_value', ['JSON']) if hasattr(config.alert_config, 'alert_value') else ['JSON'])
+                                }                    
+                    })
+                elif category in summary.get("per_category_count", {}):
+                    count = summary.get("per_category_count", {})[category]
+                    if count > threshold:  # Fixed logic: alert when EXCEEDING threshold
+                        alerts.append({
+                            "alert_type": getattr(config.alert_config, 'alert_type', ['Default']) if hasattr(config.alert_config, 'alert_type') else ['Default'],
+                            "alert_id": "alert_"+category+'_'+frame_key,
+                            "incident_category": self.CASE_TYPE,
+                            "threshold_level": threshold,
+                            "ascending": get_trend(self._ascending_alert_list, lookback=900, threshold=0.8),
+                            "settings": {t: v for t, v in zip(getattr(config.alert_config, 'alert_type', ['Default']) if hasattr(config.alert_config, 'alert_type') else ['Default'],
+                                    getattr(config.alert_config, 'alert_value', ['JSON']) if hasattr(config.alert_config, 'alert_value') else ['JSON'])
+                                }       
+                        })
+        else:
+            pass
+        return alerts
+
+    def _generate_incidents(self, counting_summary: Dict, alerts: List, config: WarehouseObjectConfig,
+                         frame_number: Optional[int] = None, stream_info: Optional[Dict[str, Any]] = None) -> List[
+        Dict]:
+        """Generate structured events for the output format with frame-based keys."""
+
+        # Use frame number as key, fallback to 'current_frame' if not available
+        frame_key = str(frame_number) if frame_number is not None else "current_frame"
+        incidents=[]
+        total_detections = counting_summary.get("total_count", 0)
+        current_timestamp = self._get_current_timestamp_str(stream_info)
+        camera_info = self.get_camera_info_from_stream(stream_info)
+        
+        self._ascending_alert_list = self._ascending_alert_list[-900:] if len(self._ascending_alert_list) > 900 else self._ascending_alert_list
+
+        if total_detections > 0:
+            # Determine event level based on thresholds
+            level = "low"
+            intensity = 5.0
+            start_timestamp = self._get_start_timestamp_str(stream_info)
+            if start_timestamp and self.current_incident_end_timestamp=='N/A':
+                self.current_incident_end_timestamp = 'Incident still active'
+            elif start_timestamp and self.current_incident_end_timestamp=='Incident still active':
+                if len(self._ascending_alert_list) >= 15 and sum(self._ascending_alert_list[-15:]) / 15 < 1.5: 
+                    self.current_incident_end_timestamp = current_timestamp
+            elif self.current_incident_end_timestamp!='Incident still active' and self.current_incident_end_timestamp!='N/A':
+                self.current_incident_end_timestamp = 'N/A'
+                
+            if config.alert_config and config.alert_config.count_thresholds:
+                threshold = config.alert_config.count_thresholds.get("all", 15)
+                intensity = min(10.0, (total_detections / threshold) * 10)
+
+                if intensity >= 9:
+                    level = "critical"
+                    self._ascending_alert_list.append(3)
+                elif intensity >= 7:
+                    level = "significant"
+                    self._ascending_alert_list.append(2)
+                elif intensity >= 5:
+                    level = "medium"
+                    self._ascending_alert_list.append(1)
+                else:
+                    level = "low"
+                    self._ascending_alert_list.append(0)
+            else:
+                if total_detections > 30:
+                    level = "critical"
+                    intensity = 10.0
+                    self._ascending_alert_list.append(3)
+                elif total_detections > 25:
+                    level = "significant"
+                    intensity = 9.0
+                    self._ascending_alert_list.append(2)
+                elif total_detections > 15:
+                    level = "medium"
+                    intensity = 7.0
+                    self._ascending_alert_list.append(1)
+                else:
+                    level = "low"
+                    intensity = min(10.0, total_detections / 3.0)
+                    self._ascending_alert_list.append(0)
+
+            # Generate human text in new format
+            human_text_lines = [f"INCIDENTS DETECTED @ {current_timestamp}:"]
+            human_text_lines.append(f"\tSeverity Level: {(self.CASE_TYPE,level)}")
+            human_text = "\n".join(human_text_lines)
+
+            alert_settings = []
+            if config.alert_config and hasattr(config.alert_config, 'alert_type'):
+                alert_settings.append({
+                    "alert_type": getattr(config.alert_config, 'alert_type', ['Default']) if hasattr(config.alert_config, 'alert_type') else ['Default'],
+                    "incident_category": self.CASE_TYPE,
+                    "threshold_level": config.alert_config.count_thresholds if hasattr(config.alert_config, 'count_thresholds') else {},
+                    "ascending": True,
+                    "settings": {t: v for t, v in zip(getattr(config.alert_config, 'alert_type', ['Default']) if hasattr(config.alert_config, 'alert_type') else ['Default'],
+                                        getattr(config.alert_config, 'alert_value', ['JSON']) if hasattr(config.alert_config, 'alert_value') else ['JSON'])
+                                }
+                })
+        
+            event= self.create_incident(incident_id=self.CASE_TYPE+'_'+str(frame_number), incident_type=self.CASE_TYPE,
+                        severity_level=level, human_text=human_text, camera_info=camera_info, alerts=alerts, alert_settings=alert_settings,
+                        start_time=start_timestamp, end_time=self.current_incident_end_timestamp,
+                        level_settings= {"low": 1, "medium": 3, "significant":4, "critical": 7})
+            incidents.append(event)
+
+        else:
+            self._ascending_alert_list.append(0)
+            incidents.append({})
+
+        return incidents
+
+    def _generate_tracking_stats(
+            self,
+            counting_summary: Dict,
+            alerts: List,
+            config: WarehouseObjectConfig,
+            frame_number: Optional[int] = None,
+            stream_info: Optional[Dict[str, Any]] = None
+    ) -> List[Dict]:
+        """Generate structured tracking stats matching eg.json format."""
+        camera_info = self.get_camera_info_from_stream(stream_info)
+        
+        tracking_stats = []
+        # frame_key = str(frame_number) if frame_number is not None else "current_frame"
+        # tracking_stats = [{frame_key: []}]
+        # frame_tracking_stats = tracking_stats[0][frame_key]
+        
+        total_detections = counting_summary.get("total_count", 0) #CURRENT total count of all classes
+        total_counts_dict = counting_summary.get("total_counts", {}) #TOTAL cumulative counts per class
+        cumulative_total = sum(total_counts_dict.values()) if total_counts_dict else 0 #TOTAL combined cumulative count
+        per_category_count = counting_summary.get("per_category_count", {}) #CURRENT count per class
+
+        track_ids_info = self._get_track_ids_info(counting_summary.get("detections", []))
+        
+        current_timestamp = self._get_current_timestamp_str(stream_info, precision=False)
+        start_timestamp = self._get_start_timestamp_str(stream_info, precision=False)
+        
+        # Create high precision timestamps for input_timestamp and reset_timestamp
+        high_precision_start_timestamp = self._get_current_timestamp_str(stream_info, precision=True)
+        high_precision_reset_timestamp = self._get_start_timestamp_str(stream_info, precision=True)
+
+    
+        # Build total_counts array in expected format
+        total_counts = []
+        for cat, count in total_counts_dict.items():
+            if count > 0:
+                total_counts.append({
+                    "category": cat,
+                    "count": count
+                })
+        print(total_counts)
+        # Build current_counts array in expected format  
+        current_counts = []
+        for cat, count in per_category_count.items():
+            if count > 0 or total_detections > 0:  # Include even if 0 when there are detections
+                current_counts.append({
+                    "category": cat,
+                    "count": count
+                })
+        print(current_counts)
+        # Prepare detections without confidence scores (as per eg.json)
+        detections = []
+        for detection in counting_summary.get("detections", []):
+            bbox = detection.get("bounding_box", {})
+            category = detection.get("category", "person")
+            # Include segmentation if available (like in eg.json)
+            if detection.get("masks"):
+                segmentation= detection.get("masks", [])
+                detection_obj = self.create_detection_object(category, bbox, segmentation=segmentation)
+            elif detection.get("segmentation"):
+                segmentation= detection.get("segmentation")
+                detection_obj = self.create_detection_object(category, bbox, segmentation=segmentation)
+            elif detection.get("mask"):
+                segmentation= detection.get("mask")
+                detection_obj = self.create_detection_object(category, bbox, segmentation=segmentation)
+            else:
+                detection_obj = self.create_detection_object(category, bbox)
+            detections.append(detection_obj)
+        print(detections)
+        # Build alert_settings array in expected format
+        alert_settings = []
+        if config.alert_config and hasattr(config.alert_config, 'alert_type'):
+            alert_settings.append({
+                "alert_type": getattr(config.alert_config, 'alert_type', ['Default']) if hasattr(config.alert_config, 'alert_type') else ['Default'],
+                "incident_category": self.CASE_TYPE,
+                "threshold_level": config.alert_config.count_thresholds if hasattr(config.alert_config, 'count_thresholds') else {},
+                "ascending": True,
+                "settings": {t: v for t, v in zip(getattr(config.alert_config, 'alert_type', ['Default']) if hasattr(config.alert_config, 'alert_type') else ['Default'],
+                                    getattr(config.alert_config, 'alert_value', ['JSON']) if hasattr(config.alert_config, 'alert_value') else ['JSON'])
+                            }
+            })
+        print(alert_settings)
+        # Generate human_text in expected format
+        human_text_lines = [f"Tracking Statistics:"]
+        human_text_lines.append(f"CURRENT FRAME @ {current_timestamp}")
+        
+        for cat, count in per_category_count.items():
+            human_text_lines.append(f"\t{cat}: {count}")
+            
+        human_text_lines.append(f"TOTAL SINCE {start_timestamp}")
+        for cat, count in total_counts_dict.items():
+            if count > 0:
+                human_text_lines.append(f"\t{cat}: {count}")
+            
+        if alerts:
+            for alert in alerts:
+                human_text_lines.append(f"Alerts: {alert.get('settings', {})} sent @ {current_timestamp}")
+        else:
+            human_text_lines.append("Alerts: None")
+
+        human_text = "\n".join(human_text_lines)
+        reset_settings = [
+                {
+                    "interval_type": "daily",
+                    "reset_time": {
+                        "value": 9,
+                        "time_unit": "hour"
+                    }
+                }
+            ]
+        print(human_text)
+        tracking_stat=self.create_tracking_stats(total_counts=total_counts, current_counts=current_counts,
+                            detections=detections, human_text=human_text, camera_info=camera_info, alerts=alerts, alert_settings=alert_settings,
+                            reset_settings=reset_settings, start_time=high_precision_start_timestamp ,
+                            reset_time=high_precision_reset_timestamp)
+
+        tracking_stats.append(tracking_stat)
+        print(tracking_stats)
+        return tracking_stats
+
+    def _generate_business_analytics(self, counting_summary: Dict, zone_analysis: Dict, config: WarehouseObjectConfig, stream_info: Optional[Dict[str, Any]] = None, is_empty=False) -> List[Dict]:
+        """Generate standardized business analytics for the agg_summary structure."""
+        if is_empty:
+            return []
+
+        #-----IF YOUR USECASE NEEDS BUSINESS ANALYTICS, YOU CAN USE THIS FUNCTION------#
+        #camera_info = self.get_camera_info_from_stream(stream_info)
+        # business_analytics = self.create_business_analytics(nalysis_name, statistics,
+        #                          human_text, camera_info=camera_info, alerts=alerts, alert_settings=alert_settings,
+        #                          reset_settings)
+        # return business_analytics
+
+    def _generate_summary(self, summary: dict, incidents: List, tracking_stats: List, business_analytics: List, alerts: List) -> List[str]:
+        """
+        Generate a human_text string for the tracking_stat, incident, business analytics and alerts.
+        """
+        lines = {}
+        lines["Application Name"] = self.CASE_TYPE
+        lines["Application Version"] = self.CASE_VERSION
+        if len(incidents) > 0:
+            lines["Incidents:"]=f"\n\t{incidents[0].get('human_text', 'No incidents detected')}\n"
+        if len(tracking_stats) > 0:
+            lines["Tracking Statistics:"]=f"\t{tracking_stats[0].get('human_text', 'No tracking statistics detected')}\n"
+        if len(business_analytics) > 0:
+            lines["Business Analytics:"]=f"\t{business_analytics[0].get('human_text', 'No business analytics detected')}\n"
+
+        if len(incidents) == 0 and len(tracking_stats) == 0 and len(business_analytics) == 0:
+            lines["Summary"] = "No Summary Data"
+
+        return [lines]
+
+
+    def _count_categories(self, detections: list, config: WarehouseObjectConfig) -> dict:
+        """
+        Count the number of detections per category and return a summary dict.
+        The detections list is expected to have 'track_id' (from tracker), 'category', 'bounding_box', 'masks', etc.
+        Output structure will include 'track_id' and 'masks' for each detection as per AdvancedTracker output.
+        """
+        counts = {}
+        valid_detections = []
+        for det in detections:
+            cat = det.get('category', 'unknown')
+            if not all(k in det for k in ['category', 'confidence', 'bounding_box']):  # Validate required fields
+                self.logger.warning(f"Skipping invalid detection: {det}")
+                continue
+            counts[cat] = counts.get(cat, 0) + 1
+            valid_detections.append({
+                "bounding_box": det.get("bounding_box"),
+                "category": det.get("category"),
+                "confidence": det.get("confidence"),
+                "track_id": det.get("track_id"),
+                "frame_id": det.get("frame_id"),
+                "masks": det.get("masks", det.get("mask", []))  # Include masks, fallback to empty list
+            })
+        self.logger.debug(f"Valid detections after filtering: {len(valid_detections)}")
+        return {
+            "total_count": sum(counts.values()),
+            "per_category_count": counts,
+            "detections": valid_detections
+        }
+
+    def _get_track_ids_info(self, detections: list) -> Dict[str, Any]:
+        """
+        Get detailed information about track IDs (per frame).
+        """
+        # Collect all track_ids in this frame
+        frame_track_ids = set()
+        for det in detections:
+            tid = det.get('track_id')
+            if tid is not None:
+                frame_track_ids.add(tid)
+        # Use persistent total set for unique counting
+        total_track_ids = set()
+        for s in getattr(self, '_per_category_total_track_ids', {}).values():
+            total_track_ids.update(s)
+        return {
+            "total_count": len(total_track_ids),
+            "current_frame_count": len(frame_track_ids),
+            "total_unique_track_ids": len(total_track_ids),
+            "current_frame_track_ids": list(frame_track_ids),
+            "last_update_time": time.time(),
+            "total_frames_processed": getattr(self, '_total_frame_counter', 0)
+        }
+
+    def _update_tracking_state(self, detections: list):
+        """
+        Track unique categories track_ids per category for total count after tracking.
+        Applies canonical ID merging to avoid duplicate counting when the underlying
+        tracker loses an object temporarily and assigns a new ID.
+        """
+        # Lazily initialise storage dicts
+        if not hasattr(self, "_per_category_total_track_ids"):
+            self._per_category_total_track_ids = {cat: set() for cat in self.target_categories}
+        self._current_frame_track_ids = {cat: set() for cat in self.target_categories}
+
+        for det in detections:
+            cat = det.get("category")
+            raw_track_id = det.get("track_id")
+            if cat not in self.target_categories or raw_track_id is None:
+                continue
+            bbox = det.get("bounding_box", det.get("bbox"))
+            canonical_id = self._merge_or_register_track(raw_track_id, bbox)
+            # Propagate canonical ID back to detection so downstream logic uses it
+            det["track_id"] = canonical_id
+
+            self._per_category_total_track_ids.setdefault(cat, set()).add(canonical_id)
+            self._current_frame_track_ids[cat].add(canonical_id)
+
+    def get_total_counts(self):
+        """
+        Return total unique track_id count for each category.
+        """
+        return {cat: len(ids) for cat, ids in getattr(self, '_per_category_total_track_ids', {}).items()}
+
+    def _format_timestamp_for_video(self, timestamp: float) -> str:
+        """Format timestamp for video chunks (HH:MM:SS.ms format)."""
+        hours = int(timestamp // 3600)
+        minutes = int((timestamp % 3600) // 60)
+        seconds = round(float(timestamp % 60),2)
+        return f"{hours:02d}:{minutes:02d}:{seconds:.1f}"
+
+    def _format_timestamp_for_stream(self, timestamp: float) -> str:
+        """Format timestamp for streams (YYYY:MM:DD HH:MM:SS format)."""
+        dt = datetime.fromtimestamp(timestamp, tz=timezone.utc)
+        return dt.strftime('%Y:%m:%d %H:%M:%S')
+
+    def _get_current_timestamp_str(self, stream_info: Optional[Dict[str, Any]], precision=False, frame_id: Optional[str]=None) -> str:
+        """Get formatted current timestamp based on stream type."""
+        if not stream_info:
+            return "00:00:00.00"
+        # is_video_chunk = stream_info.get("input_settings", {}).get("is_video_chunk", False)
+        if precision:
+            if stream_info.get("input_settings", {}).get("start_frame", "na") != "na":
+                if frame_id:
+                    start_time = int(frame_id)/stream_info.get("input_settings", {}).get("original_fps", 30)
+                else:
+                    start_time = stream_info.get("input_settings", {}).get("start_frame", 30)/stream_info.get("input_settings", {}).get("original_fps", 30)
+                stream_time_str = self._format_timestamp_for_video(start_time)
+                return stream_time_str
+            else:
+                return datetime.now(timezone.utc).strftime("%Y-%m-%d-%H:%M:%S.%f UTC")
+
+        if stream_info.get("input_settings", {}).get("start_frame", "na") != "na":
+                if frame_id:
+                    start_time = int(frame_id)/stream_info.get("input_settings", {}).get("original_fps", 30)
+                else:
+                    start_time = stream_info.get("input_settings", {}).get("start_frame", 30)/stream_info.get("input_settings", {}).get("original_fps", 30)
+                stream_time_str = self._format_timestamp_for_video(start_time)
+                return stream_time_str
+        else:
+            # For streams, use stream_time from stream_info
+            stream_time_str = stream_info.get("input_settings", {}).get("stream_info", {}).get("stream_time", "")
+            if stream_time_str:
+                # Parse the high precision timestamp string to get timestamp
+                try:
+                    # Remove " UTC" suffix and parse
+                    timestamp_str = stream_time_str.replace(" UTC", "")
+                    dt = datetime.strptime(timestamp_str, "%Y-%m-%d-%H:%M:%S.%f")
+                    timestamp = dt.replace(tzinfo=timezone.utc).timestamp()
+                    return self._format_timestamp_for_stream(timestamp)
+                except:
+                    # Fallback to current time if parsing fails
+                    return self._format_timestamp_for_stream(time.time())
+            else:
+                return self._format_timestamp_for_stream(time.time())
+
+    def _get_start_timestamp_str(self, stream_info: Optional[Dict[str, Any]], precision=False) -> str:
+        """Get formatted start timestamp for 'TOTAL SINCE' based on stream type."""
+        if not stream_info:
+            return "00:00:00"
+        if precision:
+            if stream_info.get("input_settings", {}).get("start_frame", "na") != "na":
+                return "00:00:00"
+            else:
+                return datetime.now(timezone.utc).strftime("%Y-%m-%d-%H:%M:%S.%f UTC")
+
+        if stream_info.get("input_settings", {}).get("start_frame", "na") != "na":
+            # If video format, start from 00:00:00
+            return "00:00:00"
+        else:
+            # For streams, use tracking start time or current time with minutes/seconds reset
+            if self._tracking_start_time is None:
+                # Try to extract timestamp from stream_time string
+                stream_time_str = stream_info.get("input_settings", {}).get("stream_info", {}).get("stream_time", "")
+                if stream_time_str:
+                    try:
+                        # Remove " UTC" suffix and parse
+                        timestamp_str = stream_time_str.replace(" UTC", "")
+                        dt = datetime.strptime(timestamp_str, "%Y-%m-%d-%H:%M:%S.%f")
+                        self._tracking_start_time = dt.replace(tzinfo=timezone.utc).timestamp()
+                    except:
+                        # Fallback to current time if parsing fails
+                        self._tracking_start_time = time.time()
+                else:
+                    self._tracking_start_time = time.time()
+
+            dt = datetime.fromtimestamp(self._tracking_start_time, tz=timezone.utc)
+            # Reset minutes and seconds to 00:00 for "TOTAL SINCE" format
+            dt = dt.replace(minute=0, second=0, microsecond=0)
+            return dt.strftime('%Y:%m:%d %H:%M:%S')
+
+    # ------------------------------------------------------------------ #
+    # Helper to merge masks back into detections                        #
+    # ------------------------------------------------------------------ #
+    def _attach_masks_to_detections(
+        self,
+        processed_detections: List[Dict[str, Any]],
+        raw_detections: List[Dict[str, Any]],
+        iou_threshold: float = 0.5,
+    ) -> List[Dict[str, Any]]:
+        """
+        Attach segmentation masks from the original `raw_detections` list to the
+        `processed_detections` list returned after smoothing/tracking.
+
+        Matching between detections is performed using Intersection-over-Union
+        (IoU) of the bounding boxes. For each processed detection we select the
+        raw detection with the highest IoU above `iou_threshold` and copy its
+        `masks` (or `mask`) field. If no suitable match is found, the detection
+        keeps an empty list for `masks` to maintain a consistent schema.
+        """
+
+        if not processed_detections or not raw_detections:
+            # Nothing to do – ensure masks key exists for downstream logic.
+            for det in processed_detections:
+                det.setdefault("masks", [])
+            return processed_detections
+
+        # Track which raw detections have already been matched to avoid
+        # assigning the same mask to multiple processed detections.
+        used_raw_indices = set()
+
+        for det in processed_detections:
+            best_iou = 0.0
+            best_idx = None
+
+            for idx, raw_det in enumerate(raw_detections):
+                if idx in used_raw_indices:
+                    continue
+
+                iou = self._compute_iou(det.get("bounding_box"), raw_det.get("bounding_box"))
+                if iou > best_iou:
+                    best_iou = iou
+                    best_idx = idx
+
+            if best_idx is not None and best_iou >= iou_threshold:
+                raw_det = raw_detections[best_idx]
+                masks = raw_det.get("masks", raw_det.get("mask"))
+                if masks is not None:
+                    det["masks"] = masks
+                used_raw_indices.add(best_idx)
+            else:
+                # No adequate match – default to empty list to keep schema consistent.
+                det.setdefault("masks", ["EMPTY"])
+
+        return processed_detections
+
+    def _extract_predictions(self, detections: list) -> List[Dict[str, Any]]:
+        """
+        Extract prediction details for output (category, confidence, bounding box).
+        """
+        return [
+            {
+                "category": det.get("category", "unknown"),
+                "confidence": det.get("confidence", 0.0),
+                "bounding_box": det.get("bounding_box", {}),
+                "mask": det.get("mask", det.get("masks", None))  # Accept either key
+            }
+            for det in detections
+        ]
+
+
+    # ------------------------------------------------------------------ #
+    # Canonical ID helpers                                               #
+    # ------------------------------------------------------------------ #
+    def _compute_iou(self, box1: Any, box2: Any) -> float:
+        """Compute IoU between two bounding boxes which may be dicts or lists.
+        Falls back to 0 when insufficient data is available."""
+
+        # Helper to convert bbox (dict or list) to [x1, y1, x2, y2]
+        def _bbox_to_list(bbox):
+            if bbox is None:
+                return []
+            if isinstance(bbox, list):
+                return bbox[:4] if len(bbox) >= 4 else []
+            if isinstance(bbox, dict):
+                if "xmin" in bbox:
+                    return [bbox["xmin"], bbox["ymin"], bbox["xmax"], bbox["ymax"]]
+                if "x1" in bbox:
+                    return [bbox["x1"], bbox["y1"], bbox["x2"], bbox["y2"]]
+                # Fallback: first four numeric values
+                values = [v for v in bbox.values() if isinstance(v, (int, float))]
+                return values[:4] if len(values) >= 4 else []
+            return []
+
+        l1 = _bbox_to_list(box1)
+        l2 = _bbox_to_list(box2)
+        if len(l1) < 4 or len(l2) < 4:
+            return 0.0
+        x1_min, y1_min, x1_max, y1_max = l1
+        x2_min, y2_min, x2_max, y2_max = l2
+
+        # Ensure correct order
+        x1_min, x1_max = min(x1_min, x1_max), max(x1_min, x1_max)
+        y1_min, y1_max = min(y1_min, y1_max), max(y1_min, y1_max)
+        x2_min, x2_max = min(x2_min, x2_max), max(x2_min, x2_max)
+        y2_min, y2_max = min(y2_min, y2_max), max(y2_min, y2_max)
+
+        inter_x_min = max(x1_min, x2_min)
+        inter_y_min = max(y1_min, y2_min)
+        inter_x_max = min(x1_max, x2_max)
+        inter_y_max = min(y1_max, y2_max)
+
+        inter_w = max(0.0, inter_x_max - inter_x_min)
+        inter_h = max(0.0, inter_y_max - inter_y_min)
+        inter_area = inter_w * inter_h
+
+        area1 = (x1_max - x1_min) * (y1_max - y1_min)
+        area2 = (x2_max - x2_min) * (y2_max - y2_min)
+        union_area = area1 + area2 - inter_area
+
+        return (inter_area / union_area) if union_area > 0 else 0.0
+
+    def _merge_or_register_track(self, raw_id: Any, bbox: Any) -> Any:
+        """Return a stable canonical ID for a raw tracker ID, merging fragmented
+        tracks when IoU and temporal constraints indicate they represent the
+        same physical."""
+        if raw_id is None or bbox is None:
+            # Nothing to merge
+            return raw_id
+
+        now = time.time()
+
+        # Fast path – raw_id already mapped
+        if raw_id in self._track_aliases:
+            canonical_id = self._track_aliases[raw_id]
+            track_info = self._canonical_tracks.get(canonical_id)
+            if track_info is not None:
+                track_info["last_bbox"] = bbox
+                track_info["last_update"] = now
+                track_info["raw_ids"].add(raw_id)
+            return canonical_id
+
+        # Attempt to merge with an existing canonical track
+        for canonical_id, info in self._canonical_tracks.items():
+            # Only consider recently updated tracks
+            if now - info["last_update"] > self._track_merge_time_window:
+                continue
+            iou = self._compute_iou(bbox, info["last_bbox"])
+            if iou >= self._track_merge_iou_threshold:
+                # Merge
+                self._track_aliases[raw_id] = canonical_id
+                info["last_bbox"] = bbox
+                info["last_update"] = now
+                info["raw_ids"].add(raw_id)
+                return canonical_id
+
+        # No match – register new canonical track
+        canonical_id = raw_id
+        self._track_aliases[raw_id] = canonical_id
+        self._canonical_tracks[canonical_id] = {
+            "last_bbox": bbox,
+            "last_update": now,
+            "raw_ids": {raw_id},
+        }
+        return canonical_id
+
+    def _format_timestamp(self, timestamp: float) -> str:
+        """Format a timestamp for human-readable output."""
+        return datetime.fromtimestamp(timestamp, timezone.utc).strftime('%Y-%m-%d %H:%M:%S UTC')
+
+    def _get_tracking_start_time(self) -> str:
+        """Get the tracking start time, formatted as a string."""
+        if self._tracking_start_time is None:
+            return "N/A"
+        return self._format_timestamp(self._tracking_start_time)
+
+    def _set_tracking_start_time(self) -> None:
+        """Set the tracking start time to the current time."""
+        self._tracking_start_time = time.time()