Student project / computer vision / AI coaching
A web-based sport trainer that sees, evaluates, and coaches movement.
The team built an interactive fitness assistant that uses camera-based pose detection to recognize exercises, count repetitions, and give immediate feedback through a modern web interface.
How it works
From camera input to meaningful coaching feedback.
The system captures movement through a camera, extracts body key points with pose detection, and evaluates the current exercise state. The web app turns that analysis into clear signals: correct and incorrect repetitions, end-position instructions, timers, scores, and post-workout summaries.
Product experience
Training modes built around feedback loops.
01
Guided individual training
Users choose an exercise, difficulty, and target repetitions. The trainer then guides the movement and keeps the session focused with clear visual progress.
02
Live pose detection
Detected body key points are drawn directly over the camera image. The app can compare posture against the expected position and count correct and incorrect repetitions in the moment.
03
Battle mode and evaluation
A playful competitive mode lets two users challenge each other, while the final evaluation summarizes performance with counts, rating, and success percentage.
Results
The prototype in motion.
The included results video shows the student prototype being used as an interactive training assistant.
Exercise recognition
Movements supported by the trainer.
Data security
No personal data is collected or used.
This project webpage is provided for informational purposes only. It does not collect, store, process, analyze, or transmit personal data. No cookies, tracking technologies, analytics services, user accounts, contact forms, or server-side user profiles are used on this page. Any media shown on the website is static project material and is not connected to live user input.