🎥 Demo Video Watch the demo = https://github.com/gaurav02081/push-up-tracker-ai/blob/main/assets/demo.mp4

A computer vision-based push-up counter that uses MediaPipe for real-time pose detection and tracking. The application automatically counts your push-ups by monitoring your elbow angle and provides visual feedback.

• Real-time pose tracking using MediaPipe
• Automatic rep counting based on elbow angle
• Visual feedback with progress bars and angle display
• Session logging to CSV file with timestamps and statistics
• Optional face recognition using InsightFace
• Session summary popup at the end of each workout
• User-friendly interface with clear visual indicators

Before you begin, ensure you have:

• Windows 10/11 (tested on Windows)
• Python 3.10 or higher installed
• Webcam connected and working
• Good lighting for accurate pose detection
• 6-8 feet of space from your camera

First, verify that Python is installed on your system:

py --version

If you see a version number (e.g., "Python 3.10.0"), you're good to go. If not, download and install Python from python.org.

1. Download or clone this project to your computer
2. Open Command Prompt or PowerShell
3. Navigate to the project folder:

   cd path\to\your\project\folder

Install the required packages using these commands:

# Install core dependencies
py -m pip install opencv-python==4.8.1.78 mediapipe numpy==1.26.4

# Optional: Install face recognition (recommended)
py -m pip install onnxruntime insightface

Note: The specific versions ensure compatibility. If you encounter any errors, try installing without version numbers:

py -m pip install opencv-python mediapipe numpy

Run a quick test to ensure everything is working:

py -c "import cv2, mediapipe, numpy; print('All packages installed successfully!')"

If you see "All packages installed successfully!", you're ready to go!

1. Start the application:

   py cool.py

2. Position yourself:

   • Stand about 6-8 feet from your camera
   • Make sure your full upper body is visible
   • Ensure good lighting (avoid backlighting)
   • Wear clothing that doesn't obscure your arms

3. Start your workout:

   • The app will automatically detect your pose
   • Begin doing push-ups
   • Watch the real-time feedback on screen

4. End your session:

   • Press 'q' to quit
   • Review your session summary
   • Your stats are automatically saved

• Elbow Angle Display: Shows your current elbow angle in degrees
• Push-up Counter: Displays your total reps
• Progress Bars: Visual indicators on the sides showing your position
• Pose Overlay: Skeleton lines showing detected body parts
• User Recognition: Shows recognized user name (if face recognition is set up)

• 'q' - Quit the application and show summary
• Any other key - Continue the session

To enable automatic user recognition:

1. Create the faces folder structure:

   faces/
   ├── YourName/
   │   ├── photo1.jpg
   │   ├── photo2.jpg
   │   └── photo3.jpg
   └── AnotherUser/
       ├── photo1.jpg
       └── photo2.jpg
   

2. Add photos:

   • Create a folder with your name inside the faces/ directory
   • Add 2-3 clear photos of your face (different angles work best)
   • Supported formats: JPG, JPEG, PNG, BMP

3. Test recognition:

   • Run the application
   • Look at the camera for the first few seconds
   • Your name should appear in the top-right corner

• Down position: Elbow angle < 90°
• Up position: Elbow angle > 160°
• One rep: Complete cycle from up → down → up

The app automatically saves your workout data to workout_log.csv:

datetime,name,reps,duration_sec,pace_rpm
2024-01-15 14:30:25,John,15,120.5,7.47
2024-01-15 16:45:12,John,20,180.2,6.66

Camera not working:

• Make sure your webcam is connected and not being used by another application
• Try closing other video applications (Zoom, Teams, etc.)
• Check Windows camera permissions

Poor pose detection:

• Improve lighting (avoid shadows and backlighting)
• Ensure your full upper body is visible
• Wear clothing that doesn't obscure your arms
• Stand 6-8 feet from the camera

Package installation errors:

# If you get NumPy conflicts:
py -m pip uninstall numpy
py -m pip install numpy==1.26.4

# If OpenCV fails:
py -m pip install opencv-contrib-python

Face recognition not working:

• Ensure InsightFace is installed: py -m pip install insightface
• Check that your photos are clear and well-lit
• Make sure the folder structure is correct
• Try adding more photos of different angles

Application crashes:

• Update your graphics drivers
• Try running with a different camera
• Check Windows compatibility mode

• Close unnecessary applications to free up system resources
• Use good lighting for better detection accuracy
• Position yourself correctly - full upper body visible
• Wear contrasting clothing to your background

push-up-tracker/
├── cool.py              # Main application
├── README.md            # This file
├── requirements.txt     # Dependencies list
├── workout_log.csv      # Session history (auto-generated)
└── faces/               # Face recognition photos
    └── YourName/
        ├── photo1.jpg
        ├── photo2.jpg
        └── photo3.jpg

• OpenCV 4.8.1.78 - Computer vision and camera handling
• MediaPipe 0.10.14 - Real-time pose estimation
• NumPy 1.26.4 - Numerical computations
• InsightFace 0.7.3 - Face recognition (optional)
• ONNX Runtime - Machine learning inference

1. Pose Detection: MediaPipe analyzes each frame to detect 33 body landmarks
2. Angle Calculation: Computes elbow angle from shoulder, elbow, and wrist positions
3. Rep Counting: State machine tracks up/down positions to count complete reps
4. Face Recognition: InsightFace creates embeddings from photos and matches them in real-time
5. Data Logging: Session statistics are saved to CSV for tracking progress

• CPU: Any modern multi-core processor
• RAM: 4GB minimum, 8GB recommended
• Storage: 100MB free space
• Camera: Any USB webcam or built-in camera
• OS: Windows 10/11 (tested), should work on macOS/Linux

Feel free to improve this project by:

• Adding support for other exercises
• Improving the UI/UX
• Adding more detailed analytics
• Supporting different camera setups

This project is open source. Feel free to use and modify for personal or educational purposes.

If you encounter issues:

1. Check the troubleshooting section above
2. Ensure all prerequisites are met
3. Try reinstalling dependencies
4. Check that your camera and lighting are adequate

Happy working out! 💪