PhysioGrow: AI-Powered Physical Therapy Form Correction
Inspiration
Poor form during rehabilitation can lead to injury or ineffective treatment, and at-home sessions often lack feedback and motivation. PhysioGrow is a web application that uses real-time computer vision and AI to analyze and correct exercise form while keeping people engaged through visual progress tracking and immediate feedback.
What it does
- Analyzes exercise form in real time using pose estimation
- Provides instant audio and visual cues to correct technique
- Ties progress to a garden where a tree grows with consistent, high-quality sessions
- Rewards good form with XP, streak multipliers, and achievements
How it works
Core functionality: form analysis and correction
- Uses Google MediaPipe for real-time AI computer vision pose detection and analysis
- Tracks 33 body landmarks to produce precise skeletal data for form evaluation
Exercise form analysis
- Shoulder Abduction: analyzes elbow, wrist, and shoulder angles
- Knee Bends: tracks knee alignment, hip stability, and depth
- Arm Circles: monitors shoulder position and arm trajectory
- Compares real-time data to optimal parameters and flags deviations for correction
Real-time feedback system
- Audio cues: “Bend your knees more” and “Keep your back straight”
- Visual indicators: color-coded feedback (green = good, yellow = warning, red = poor)
- Form correction suggestions: specific, actionable instructions
Gamification integration
- Perfect-form sessions earn bonus XP and streak multipliers
- Consistently good form accelerates tree growth in the garden
- Measurable improvement over time unlocks special achievements
Innovation
Precise form analysis
Uses MediaPipe’s pose estimation for physical therapy specific analysis. The system understands rehabilitation biomechanics and gives targeted feedback.
Real-time correction
Not just tracking. The app actively corrects form in the moment to help prevent injury and improve effectiveness.
Accessibility through TTS
All feedback is delivered through text to speech for hands-free exercise and better accessibility.
Gamified motivation
A floating island garden provides visual motivation, and rewards are tied to form quality so users focus on proper technique rather than just finishing reps.
Built With
react • typescript • mediapipe • web-speech-api • tailwindcss • vite • svg
Built With
- mediapipe
- react
- svg
- tailwindcss
- typescript
- vite
- web-speech-api
Log in or sign up for Devpost to join the conversation.