# ️ Guardify-AI: Advanced Surveillance & Shoplifting Detection System    Guardify-AI is an intelligent surveillance system that leverages cutting-edge computer vision and AI to detect potential shoplifting activities in real-time. The system combines automated video recording, dual-strategy AI analysis, and an intuitive web dashboard to provide comprehensive security monitoring for retail environments. ## Problem & Solution ### The Challenge Traditional retail security systems face significant limitations: - **Manual monitoring** is labor-intensive and error-prone - **Basic motion detection** creates too many false positives - **Human oversight** is limited by attention span and fatigue - **Lack of analytics** prevents understanding of theft patterns ### Our Solution Guardify-AI provides automated, AI-powered surveillance that: - **Detects suspicious activities** with high accuracy using computer vision - **Analyzes behavior patterns** through dual AI strategies - **Provides real-time alerts** and comprehensive analytics - **Scales efficiently** across multiple retail locations ## ️ System Architecture \`\`\` ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐ │ Frontend UI │ │ Backend API │ │ Data Science │ │ React + TypeScript │◄─►│ Flask + SQLAlch │◄─►│ Pipeline │ │ Tailwind + Charts │ │ PostgreSQL+Redis│ │ Vertex AI + GPT │ └─────────────────┘ └─────────────────┘ └─────────────────┘ │ │ │ └──────────┬─────────────┼───────────────────────┘ │ │ ┌─────────────────┐ │ │ Video Processing│ │ │ Playwright Auto │ │ │ Google Storage │ │ │ Celery Tasks │◄──┘ └─────────────────┘ \`\`\` ## Quick Start ### Prerequisites - Python 3.9+ - Node.js 18+ - PostgreSQL database - Docker Desktop (for Redis) - Google Cloud account ### Installation \`\`\`bash # 1. Clone repository git clone https://github.com/your-org/guardify-ai.git cd guardify-ai # 2. Backend setup python -m venv venv source venv/bin/activate # Windows: venv\Scripts\activate pip install -r requirements.txt # 3. Frontend setup cd UI/Guardify-UI npm install cd ../.. # 4. Configure environment cp .env.example .env # Edit .env with your credentials \`\`\` ### Running the Application **Important**: For full functionality, you need Redis and Celery workers running. \`\`\`bash # 1. Start Redis (ensure Docker Desktop is running first) docker run --name redis -d -p 6379:6379 redis # 2. Start backend (Terminal 1) cd backend && python run.py # 3. Start Celery worker (Terminal 2) celery -A backend.celery_app worker --loglevel=info --pool=solo --queues=analysis # 4. Start frontend (Terminal 3) cd UI/Guardify-UI && npm run dev # 5. Run AI analysis (Terminal 4 - optional) python data_science/src/main.py --strategy unified \`\`\` **Application URLs:** - **Backend**: http://localhost:8574 - **Frontend**: http://localhost:5173 - **Celery Flower** (optional monitoring): http://localhost:5555 ## AI Analysis Strategies ### Unified Strategy (Recommended) - **Direct video→detection** analysis using few-shot learning - **Faster processing** with immediate confidence scoring - **Best for**: Real-time production environments ### Agentic Strategy - **Two-stage pipeline**: Computer Vision → LLM Analysis - **Detailed reasoning** with transparent decision making - **Best for**: Thorough analysis and model debugging \`\`\`bash # Run unified analysis (production) python data_science/src/main.py --strategy unified --max-videos 50 # Run agentic analysis (detailed) python data_science/src/main.py --strategy agentic --iterations 3 \`\`\` ## Key Features ### **Web Dashboard** - Real-time security event monitoring - Interactive analytics and charts - Shop and camera management - Responsive design for all devices ### **AI Detection** - Dual-strategy analysis pipeline - Real theft example training - Confidence scoring and reasoning - Continuous model improvement ### **Video Processing** - Automated camera feed capture - Intelligent video segmentation - Cloud storage with auto-cleanup - Real-time recording control ### **Analytics & Insights** - Event pattern analysis - Camera performance metrics - Time-based activity trends - Custom reporting capabilities ## Development ### Development Workflow \`\`\`bash # Daily development startup 1. Start Docker Desktop 2. docker run --name redis -d -p 6379:6379 redis 3. cd backend && python run.py 4. celery -A backend.celery_app worker --loglevel=info --pool=solo --queues=analysis 5. cd UI/Guardify-UI && npm run dev \`\`\` ### Project Structure \`\`\` guardify-ai/ ├── backend/ # Flask API + services │ ├── celery_tasks/ # Background job processing │ ├── services/ # Business logic │ └── video/ # Video recording automation ├── UI/Guardify-UI/ # React frontend ├── data_science/ # AI analysis pipeline ├── google_client/ # Cloud integration └── utils/ # Shared utilities \`\`\` ### API Documentation All endpoints use JWT authentication and return: \`\`\`json \{ "result": \{ /* data */ \}, "errorMessage": null \} \`\`\` **Core Endpoints:** - \`GET/POST /events\` - Event management - \`GET /stats\` - Analytics data - \`GET/POST /shops\` - Shop management - \`GET/POST /shops/\{id\}/cameras\` - Camera control ### Testing \`\`\`bash # Backend tests cd backend && python -m pytest tests/ # Frontend linting cd UI/Guardify-UI && npm run lint # AI pipeline tests cd data_science && python -m pytest tests/ \`\`\` ## ️ Configuration Create \`.env\` file with: \`\`\`env # Database DATABASE_URL=postgresql://user:pass@localhost:5432/guardify_db # Google Cloud GOOGLE_APPLICATION_CREDENTIALS=path/to/service-key.json GOOGLE_CLOUD_BUCKET_NAME=your-bucket # AI Services AZURE_OPENAI_API_KEY=your-openai-key AZURE_OPENAI_API_BASE=your-azure-endpoint # Camera System PROVISION_ISR_USERNAME=camera-username PROVISION_ISR_PASSWORD=camera-password # Application JWT_SECRET_KEY=your-jwt-secret REDIS_URL=redis://localhost:6379/0 # Optional \`\`\` ## License & Support **License**: MIT License - see [LICENSE](LICENSE) file **Support**: Create issues in the GitHub repository or check the \`logs/\` directory for debugging. ## ️ Roadmap - [ ] SMS notifications - [ ] Advanced ML model improvements --- **Built for modern retail security and loss prevention** ️