PhysioPal: AI-Powered Exercise Analysis & Physiotherapy Assistant

Bringing professional physiotherapy guidance directly to your home through AI-powered exercise analysis

🎯 Problem & Background

Canada faces a significant physiotherapy workforce crisis, requiring a 62% increase in physiotherapists just to reach the OECD average of 1.1 per 1,000 population [1]. This gap leaves patients with limited access to supervised rehabilitation, especially in remote regions. From athletes to the elderly, performing exercises at home carries the risk of incorrect form without real-time guidance. Most fitness apps only offer rep counting or workout logging, falling short of assessing exercise quality.

Therefore, there is a pressing need for an ML-powered system that brings the expertise of a physio directly into your home.

💡 Motivation

Many people want to stay healthy or recover from injury but lack access to consistent, affordable coaching. By turning everyday devices into a supportive digital coach, PhysioPal helps users exercise safely, build confidence, and stay accountable.

The goal is simple: make supervised rehab available to anyone, anywhere.

✨ What Inspired Me

The inspiration for PhysioPal came from witnessing the struggles of friends and family members who needed physiotherapy but couldn't access consistent professional guidance. Whether it was a friend recovering from a sports injury or an elderly relative needing post-surgery rehabilitation, the common thread was the same: the gap between professional care and home exercise.

I realized that while we have advanced computer vision and AI technologies, they weren't being applied to solve this critical healthcare accessibility problem. The idea of creating a "digital physiotherapist" that could provide real-time form feedback using just a webcam felt both ambitious and necessary.

🧠 What I Learned

Technical Skills

Computer Vision & Pose Estimation: Deep dive into MediaPipe's pose detection capabilities and how to extract meaningful biomechanical data
Real-time Video Processing: Building efficient pipelines for live video analysis with OpenCV and Streamlit
State Machine Design: Creating robust exercise tracking systems that can handle complex movement patterns
AI Integration: Leveraging Google's Gemini API for intelligent exercise recommendations and personalized guidance
Web Development: Building interactive dashboards with Streamlit and real-time communication with WebRTC

Domain Knowledge

Biomechanics: Understanding joint angles, movement patterns, and proper exercise form
Physiotherapy Principles: Learning about exercise progression, injury prevention, and rehabilitation protocols
Healthcare Accessibility: Researching the physiotherapy shortage and its impact on patient outcomes