AI Personal Trainer

An intelligent fitness analysis application that uses computer vision (MediaPipe & OpenCV) to track exercise reps, analyze form, and provide real-time feedback.

Features

• Real-time Pose Estimation: Tracks 33 3D body landmarks.
• Exercise Support:
  • Dumbbell Curls:
    • Independent Left/Right arm tracking.
    • Counts full repetitions (< 80° flexion to > 140° extension).
    • Auto-detects active arm.
  • Squats:
    • Depth analysis.
    • Counts repetitions based on knee angles (< 100° down, > 160° up).
• Repetition Counting: Robust logic to prevent false positives and support partial reps.
• Feedback System:
  • Displays repetition count.
  • Visual overlays for joint angles and skeleton.
  • Directional feedback (e.g., "Up (Right)", "Good Depth!").
• GPU Acceleration: Experimental support for hardware acceleration via MediaPipe Tasks API.

Installation

Prerequisites

• Python 3.10 or higher.
• uv (Recommended package manager) or pip.

Setup

1. Clone the repository:

   git clone <repository-url>
   cd ai-personal-trainer

2. Install dependencies:

   uv pip install -r requirements.txt

   Or with standard pip:

   pip install -r requirements.txt

Usage

Run the application using the command line.

Basic Usage (Webcam)

uv run python main.py

Defaults to Dumbbell Curl exercise.

Select Exercise

Analyze Squats:

uv run python main.py --exercise squat

Use a Video File

Analyze a pre-recorded video instead of webcam:

uv run python main.py --video path/to/video.mp4

Enable GPU Acceleration (Experimental)

Use the modern MediaPipe Task API with GPU support:

uv run python main.py --gpu

Project Structure

• main.py: Application entry point. Handles video capture, UI drawing, and loop.
• pose_estimator.py:
  • PoseDetector: Legacy CPU-based detector wrapper.
  • GPUPoseDetector: GPU-accelerated detector using Tasks API.
• exercises.py: Contains logic for specific exercises (DumbbellCurl, Squat).
• tests/: Unit tests for exercise logic.

Technology Stack

• Python: Core language.
• MediaPipe: For pose detection and landmark extraction.
• OpenCV: For image processing and display.
• NumPy: For angle calculations.

Contributing

1. Fork the project.
2. Create your feature branch (git checkout -b feature/AmazingFeature).
3. Commit your changes (git commit -m 'Add some AmazingFeature').
4. Push to the branch (git push origin feature/AmazingFeature).
5. Open a Pull Request.

License

Distributed under the MIT License. See LICENSE for more information.