GitHubHelp

asrira428/PosturePal

PosturePal is a posture‑tracking app built with Mediapipe Pose, OpenCV and Mistral AI. It captures your webcam feed to compute a live posture score, logs your sessions, and uses Mistral models to generate clear, personalized tips right in the GUI—helping you spot slouching patterns and improve over time.

★ 0Forks 0TypeScriptGitHub ↗Compare

README

🧍‍♂️ PosturePal

🌟 Inspiration

Good posture is essential for long-term health, yet many people struggle to maintain it, especially while working at a desk for extended periods. We wanted to build a solution that provides real-time feedback and actionable insights to help users develop better posture habits.

🎯 What It Does

PosturePal analyzes a user's posture using their webcam and provides instant feedback. It detects key body landmarks, calculates posture metrics, and assigns a score based on alignment.

🚀 Features:

  • Detects and evaluates posture in real-time using machine learning.
  • Generates summaries and personalized improvement strategies using AWS Bedrock and Mistral AI (Mistral Large 24.02).
  • Tracks posture history with visualizations for long-term insights.

🛠️ How We Built It

We developed PosturePal using a full-stack approach:

📌 Frontend

  • Built with React, TypeScript, and Tailwind CSS for a responsive and intuitive UI.

🔧 Backend

  • Flask API for authentication, posture data processing, and AWS integration.

🧠 Machine Learning

  • TensorFlow.js and MediaPipe power real-time posture analysis directly in the browser.
  • Trained a custom Neural Network for enhanced accuracy.

☁️ AWS Bedrock Integration

  • Runs Mistral Large to summarize posture data and generate personalized improvement strategies.

🗄️ Database

  • Supabase (PostgreSQL) stores user data, posture scores, and session history.

🚧 Challenges We Ran Into

🔹 Fine-tuning the posture detection model to minimize inaccuracies.
🔹 Ensuring real-time performance while integrating TensorFlow.js, MediaPipe, and AWS Bedrock.
🔹 Optimizing Mistral AI's recommendations by testing different prompt structures.

🏆 Accomplishments That We're Proud Of

✅ Real-time posture tracking with machine learning.
✅ AWS Bedrock + Mistral AI for personalized posture improvement strategies.
✅ Seamless user experience with authentication, tracking, and analytics.
✅ Google OAuth integration for secure login.

📚 What We Learned

🔹 Real-time computer vision and AI-powered summarization.
🔹 Optimizing ML models for web applications.
🔹 Hands-on experience with AWS Bedrock and deploying scalable AI solutions.

🚀 What's Next for PosturePal

🚀 Refining the posture scoring system with better data and feedback.
📱 Expanding device compatibility, including mobile support.
💡 Enhancing Mistral’s recommendations for more personalized guidance.

PosturePal is more than just an app—it’s a lifestyle change powered by AI. 🚀

🎯 Tracks

Health

Contributors

asrira428

Issues