نظام ذكي للتعرف على نشاط الأفراد داخل المتاجر باستخدام الذكاء الاصطناعي

AI-Based Retail Human Activity Recognition System

This project implements an intelligent video surveillance system that uses YOLO with instance segmentation and tracking to detect and recognize human activities in real-time within a retail environment. Actions such as walking, touching, and watching are automatically detected and visualized using red masks, bounding boxes, and bold class labels.

Description

This project showcases a real-time retail environment monitoring system using YOLO with instance segmentation and tracking. It detects and classifies human activities from surveillance footage in shopping aisles while overlaying bounding boxes, red masks, and bold labels for clear interpretability.

Core Objectives

• 🎯 Detect multiple human subjects in busy retail environments.
• 🎯 Classify actions such as walking, touching, and watching.
• 🎯 Visualize actions with bounding boxes and segment masks.
• 🎯 Track customer-product interactions.
• 🎯 Enable behavior analysis and anomaly detection.

Key Features

• ✅ Activity-Specific Detection for concurrent actions.
• ✅ Instance Segmentation with red masks.
• ✅ Tracking with unique IDs across frames.
• ✅ Bold uppercase action labels.
• ✅ Optimized real-time inference loop.
• ✅ Scalable and extendable architecture.

Technologies Used

• 🧠 YOLO (Ultralytics)
• 🧠 OpenCV
• 🧠 cvzone
• 🧠 Python

Use Cases

• 🏬 Retail Analytics
• 🛡️ Loss Prevention
• 🧍‍♂️ Customer Flow Mapping
• 🎥 Smart Surveillance
• 📈 Business Intelligence

Possible Enhancements

• 💾 Event Logging System
• 📊 Behavioral Heatmaps
• 🔔 Anomaly Alert System
• 🌐 Web-Based Dashboard
• ⚡ Hardware Acceleration (ONNX / TensorRT)

Conclusion

This intelligent retail behavior detection system accurately identifies and annotates human activities in real time using YOLO advanced capabilities. It delivers actionable insights for security, analytics, and customer behavior understanding through a clean and interpretable visual interface.