TrackForm AI MVP

A mobile app that analyzes sprint form using AI-powered pose estimation, providing real-time feedback on body angles and performance scoring.

Architecture

• Frontend: Flutter mobile app (iOS/Android)
• Backend: Python Flask server with MediaPipe pose estimation
• Analysis: Real-time biomechanics analysis with scoring

Features

• Record or upload sprint videos
• AI-powered pose estimation using MediaPipe
• Real-time angle analysis (knee, elbow, shin, armpit)
• Sprint detection and timing
• Comprehensive scoring system (0-100)
• Detailed feedback and improvement tips
• Beautiful, modern UI

Setup Instructions

Backend Setup

1. Install Python dependencies:

   cd backend
   pip install -r requirements.txt

2. Start the Flask server:

   python app.py

   The server will run on http://localhost:5000

Mobile App Setup

1. Install Flutter (if not already installed):

   • Follow the official Flutter installation guide for your platform
   • Ensure Flutter is in your PATH

2. Navigate to mobile directory:

   cd mobile

3. Install dependencies:

   flutter pub get

4. Run the app:

   flutter run

API Endpoints

• GET /health - Health check
• POST /analyze - Analyze video (basic results)
• POST /analyze/detailed - Analyze video (detailed results)

Scoring System

The app evaluates sprint form based on:

• Knee Angle (25%): Optimal drive phase angle 90-170°
• Elbow Angle (20%): 90° elbow angle for efficient arm swing
• Shin Angle (20%): Ground contact angle 35-55°
• Armpit Angle (15%): Trunk position and forward lean
• Consistency (20%): Form consistency across frames

Usage

1. Start the backend server (Python Flask)
2. Launch the Flutter app
3. Record or upload a sprint video
4. View detailed analysis results with scores and feedback

Technical Details

Backend

• Flask web framework
• MediaPipe for pose estimation
• OpenCV for video processing
• NumPy for calculations

Frontend

• Flutter framework
• Camera integration
• File picker for video uploads
• HTTP client for API communication
• Material Design 3 UI

File Structure

TrackForm AI (app)/
├── backend/
│   ├── app.py              # Flask server
│   ├── pose_analyzer.py    # MediaPipe pose analysis
│   ├── scoring.py          # Scoring algorithm
│   └── requirements.txt    # Python dependencies
├── mobile/
│   ├── lib/
│   │   ├── main.dart       # App entry point
│   │   ├── models/         # Data models
│   │   ├── screens/        # UI screens
│   │   └── services/       # API service
│   └── pubspec.yaml        # Flutter dependencies
└── README.md

Development Notes

• Backend runs on localhost:5000 for MVP
• Videos processed synchronously
• MediaPipe provides fast, accurate pose detection
• Flutter camera plugin enables cross-platform recording

Next Steps

• Add async video processing queue
• Implement user accounts and history
• Add more detailed biomechanics analysis
• Deploy to cloud infrastructure
• Add real-time analysis during recording