Many people face injuries during workouts due to incorrect exercise form, which can make their fitness progress difficult and lead to long-term health issues. Fiteligence: Real-time AI-Driven Workout Assistant is designed to address this problem by providing instant, accurate feedback on exercise form using Artificial Intelligence (AI) and Computer Vision. The system detects improper movements and offers real-time corrections, helping users improve their performance, prevent injuries, and optimize their workouts. We have collected video data for four exercises, applied pose estimation using MediaPipe BlazePose, calculated and normalized joint angles, and saved the results in a CSV file., trained and tested a Machine Learning model (LSTM) to detect incorrect postures of exercises and provide feedback. We are developing a user-friendly website with real-time feedback integration.
Tools: Vs Code, Google Collab, OpenCV, MediaPipe, TensorFlow,NextJS,Python,FastAPI
Department: Department of Electrical Engineering
Project Poster
