Overview
This application is used for monitoring activities through mobile phone cameras in real time. Starting from object identification and analysis such as face and movement. All activities are captured in a certain amount of time. Then the report results will be sent and stored in the database.
Interface
Workflow
- Use livekit to enable video streaming for realtime communication
- Once streaming active and participants exist, Prepare the object detection model (e.g.YOLO, then LLMs for user response and Database integration.
- During video streaming, the app performs several analysis:
- Check number of persons,
- Detect faces and
- Detect motion (e.g Movement)
- And Every 1s the system captures image from video stream and saves it to the database.
Tech Stacks
Python, Langchain, YOLOV8, Supabase, OpenAI, Nix, Google Gemini, Railway, Nixpacks