Yoga-Pose-Landmarker

Yoga-Pose-Landmarker is a real-time pose extraction and skeletal tracking system that analyzes human posture from live video.
Built using MediaPipe Pose, this project converts raw body movement into structured landmark data and replays it on a dynamic skeletal model with low latency and high accuracy.

Why this project matters

Incorrect posture is one of the biggest reasons for yoga injuries and ineffective workouts.
This system bridges the gap between human motion and machine understanding by providing a foundation for real-time posture analysis and intelligent feedback systems.

Key Features

• Real-time human pose detection from live video
• Extraction and processing of 33 pose landmarks
• Skeletal pose tracking and visual replay
• Optimized for low-latency and smooth performance
• Lightweight and scalable design

System Workflow

1. Capture live video through the webcam.
2. Detect human pose landmarks using MediaPipe.
3. Process and normalize landmark coordinates.
4. Map landmarks to a skeletal structure.
5. Replay the pose smoothly for analysis and visualization.

🛠 Technology Stack

• Python
• MediaPipe
• OpenCV
• NumPy