This project implements a real-time multi-person posture detection system using the MoveNet model for pose estimation. It processes video input to detect and visualize keypoint connections, making it suitable for applications in sports analysis, human movement studies, and more.
- Real-time multi-person posture detection.
- High frame rate (>30 FPS) on modern laptops.
- Precise pose estimation using the MoveNet model.
- Image preprocessing for improved model input.
- Keypoint connection visualization.
- Ability to customize confidence thresholds.
- Integration with TensorFlow and OpenCV for image processing.
- Python 3.11
- TensorFlow
- OpenCV
- Clone the repository:
https://github.com/Tanmaykul29/Multi-Person_PostureDetection.git