Exercise Form Analysis — MVP (24‑Hour Build)

This is a minimal, working prototype that detects squats and push-ups from your webcam, counts reps, and gives simple real-time feedback using MediaPipe Pose + OpenCV.

1) Install

python -m venv .venv
# Windows
.venv\Scripts\activate
# macOS/Linux
source .venv/bin/activate

pip install -r requirements.txt

If you face issues installing mediapipe on Apple Silicon or Linux, ensure Python 3.9–3.11 and pip are up to date.

2) Run

python main.py

Keys inside the app:

• 1 → Squats mode
• 2 → Push-ups mode
• s → Show/Hide skeleton
• q → Quit

3) What it does

• Uses your webcam to detect pose landmarks in real time.
• Tracks knee angle (squats) and elbow angle (push-ups).
• Counts reps when you go down → up through thresholds.
• Shows “Good / Needs to go lower / Extend fully” messages.
• Prints a simple Workout Summary on exit.

4) Notes

• Ensure you’re fully visible to the camera.
• Good lighting improves detection accuracy.
• For push-ups, angle works best in side view.
• This is an MVP; thresholds are approximate and may need tuning for your body/mechanics.

5) Next Steps (beyond MVP)

• Add more exercises (lunges, planks, bicep curls).
• Voice feedback (pyttsx3) and sound cues.
• Per-rep form scoring; charts & session history.
• Mobile app (React Native) using a lightweight pose model or MediaPipe Tasks.
• AR overlay guides and personalized plans.