PhysioAI System

PhysioAI is an advanced AI-powered physiotherapy assistant designed to provide real-time posture monitoring and corrective feedback. Built using computer vision and deep learning, the system helps users perform rehabilitation exercises with precision, ensuring safety and effectiveness in a home-based setting.

🚀 Overview

The PhysioAI system addresses the challenges of remote physiotherapy by using a standard webcam to track body movements. It calculates joint angles and compares them against clinically recommended ranges to provide instant visual feedback, helping prevent injury and improve recovery outcomes.

✨ Key Features

• Real-Time Pose Estimation: Utilises MediaPipe to track 33 essential body landmarks.
• Anatomical Angle Calculation: Processes joint movements in real-time to ensure exercises are performed within correct physiological limits.
• Color-Coded Feedback: * 🟢 Green: Correct posture and alignment.
  • 🔴 Red: Postural deviation detected; immediate correction required.
• Modular Exercise Library:
  • Arms Module: Tracks shoulder, elbow, and wrist coordination.
  • Legs Module: Focuses on hip, knee, and ankle stability.
  • Shoulders Module: Monitors complex shoulder articulation and range of motion.
• Automatic Repetition Counting: Validates and counts successful reps only when the full range of motion is achieved.

🛠️ Tech Stack

• Language: Python
• Web Framework: Flask
• Computer Vision: OpenCV, MediaPipe
• Frontend: HTML5, CSS3, JavaScript
• Deployment: Designed for local and cloud-based environments (AWS EC2)

🏗️ System Architecture

1. Input Layer: Captures video stream from the user's webcam.
2. Processing Layer: * Extracts coordinates of body landmarks.
   • Calculates Euclidean distances and joint angles.
3. Logic Layer: Compares real-time data against predefined exercise thresholds.
4. Output Layer: Streams the processed video back to the web interface with visual overlays and real-time status updates.

⚙️ Installation

1. Clone the repository:

   git clone [https://github.com/yourusername/physioai.git](https://github.com/yourusername/physioai.git)
   cd physioai
   
   

2. Install required dependencies: pip install flask opencv-python mediapipe numpy

3. Run the application: python app.py

4. Access the application: Open your browser and go to http://127.0.0.1:5000

📊 Performance

• Accuracy: Achieved 93-95% accuracy in posture detection across various exercise modules.

• Latency: Optimised for low-latency feedback on standard CPU hardware.

📄 License

This project was developed as part of a Bachelor of Technology in Computer Science & Engineering at Graphic Era Hill University.