FormPro AI

Inspiration

I used to be very skinny two years ago, I was in the 9th grade and was 5 foot 8 and weighed 95 lbs. I was so insecure I wouldn't even wear a t-shirt in public, I was scared to go into the gym for all sorts of reasons, but eventually one day started going, learned how to eat properly, and learned the sciences on lifting. I am now in the 11th grade and am 5 foot 10 and weigh 170lbs. I don't want anyone else to go through the same process I want through, I want the gym to be a place where everyone can feel safe and want to go to, and that is what motivated me to start this project. (Told from POV of Aditya Donkada).

What it does

My project utilizes computer vision to analyze and improve the exercise form of gym goers by measuring the angles between joints in real-time. It offers immediate visual feedback on their performance, helping to ensure correct posture and reduce the risk of injuries during workouts.

How we built it

We built this project using a combination of OpenCV and MediaPipe, which are powerful tools for computer vision tasks. OpenCV handles the video capture and image processing, allowing us to manipulate and analyze the video feed in real-time. MediaPipe, on the other hand, provides robust models for pose estimation, enabling us to detect and track body landmarks accurately. We integrated these technologies to analyze the angles between joints, calculate their correctness during exercises, and provide immediate, actionable feedback on users' form through the video feed. This integration not only enhances the usability of the app but also ensures that the feedback is based on precise and reliable pose estimation.

Challenges we ran into

We encountered significant challenges in accurately calculating angles between joints for proper lifting form, which required refining our mathematical models and adjusting pose estimation techniques to handle different body types and lighting conditions. Additionally, ensuring real-time feedback and developing an intuitive user interface demanded extensive code optimization and thoughtful UI design to effectively communicate complex biomechanical data to users.

Accomplishments that we're proud of

We are proud of being able to make an idea that we had in our heads come to life and have a very base prototype working.

What we learned

We have learned how to read technology documentation, implement new technologies we have never seen before, and just overall how to leverage computer vision to help make the world a better place.

What's next for FormPro AI

We are going to have our friends, families, and local gyms use this base prototype and give us feedback on what to improve. We will then take that feedback and improve our product till our clients are satisfied with these base lifts, then we are looking to add in new features such as a Nutrition Chatbot, more exercises to choose from, and many more features to come.