Human Movement Summarizer
NVIDIA VILA-powered movement detection with real-time webcam/RTSP analysis and activity notifications
AI-Powered Activity Detection
Real-time human movement summarization system that watches webcam, RTSP, or video sources and generates activity summaries every few seconds. Automatically detects and notifies when notable human activity occurs.
Built on NVIDIA VILA (NIM) with OpenAI-compatible API, OpenCV for video capture, and structured JSON response parsing for robust alert detection.
Dual Operating Modes
Core Features
Real-Time Monitoring
Continuous webcam or RTSP stream analysis with configurable interval summaries and immediate activity notifications.
Report Generation
Process video files and produce structured Markdown reports of all detected movements and events with timestamps.
Activity Notifications
Desktop notifications printed for notable activities. Structured JSON parsing ensures robust alert detection.
Multi-Source Support
Works with webcam index (0, 1), RTSP streams, MP4 video files, and other OpenCV-compatible video sources.
Configurable Intervals
Adjust summary intervals (10s default), time windows (10-30s), and sampling FPS (1fps default) for optimal performance.
Strict JSON Responses
Model returns structured JSON for robust parsing of alerts and events. OpenAI-compatible API integration.
Technical Specifications
Ready to Monitor Human Activity?
Deploy NVIDIA VILA-powered movement detection with real-time notifications