Fit Trainer is an AI Trainer were it will take your live video or recorded video and count your reps and give you live feedback on your execersize form and overall.
Many people likes to go to the gym and get fit or get muscles, but does not know how to perform each exercise.
It will take either your live video or a recorded video and it will give you live feedback on the screen like your shoulder are too low keep them up. It will also count your reps and record your activity like how you did in each reps.
We are using python libraries to open the webcam of the user and then take live video and openCV reads each frame from the webcam and it feeds it into the MediaPipe Ai, Then the MediaPipe pinpoints the joints and the body's critical parts and save all the information. After that our code which is specifically designed to compute angels and motions for shoulder press comes into play. We have a complete frontend where we would upload a recorded gym session and then it will show full range of motions what we did right and what we did wrong and a great feedback with the use of openAI API. The interactive website is made by using react and the styling is done in CSS and little bit of HTML. We used git to intstall dependencies and many libraries which we needed for our project.
Many of the dependencies would not match the version of the libraries we had. This issue would also cause us to not be able to upload anything to github.
Our project as a whole has a great potential and therefore we are proud of what we made.
We learned how to communicate and how to work as a team to solve a problem. for the technical side, we learned how the the openCv reads each frame and how to send that data to media pipe and additionally we learned how does the RestAPI's work in depth.
As of now it will be able to do only shoulder press and on the side we also made the squat version of it and we tested it and it works fine. we want to
Log in or sign up for Devpost to join the conversation.