OpenCV Bicep Curl Implementation
OpenCV Jumping Jacks Implementation
OpenCV Plank Implementation

FORM FOCUS

Inspiration

Our priority as a group was coming up with an idea that implements AI in a way that could be useful to us and our community. As college students, fitness is something that is always prevalent in our day to day life. When we realized that somtimes we struggle to work out because we don't know the correct form, leading to soreness and general embarassment, we decided to to create an app that corrects its users on common mistakes and gives them helpful tips!

What it does

We are creating an exercise app that will record and store the user's name, gender, age, weight, height, and exercise goals. The app will then recommend an exercise and provide an implementation of motion tracking code in order to assist the user. This currently comes in the form of a counter and form checker depending on the recommended exercise.

How we built it

We first split the work between us four after deciding on a team leader. Our team leader gave us all tasks to complete including himself and we tried our best to accomplish this while working together and helping each other.

Challenges we ran into

One challenge we faced was understanding the use and importance of each other's code. When faced with code that you didn't write yourself, it can sometimes be confusing what each segment is meant to do. We fixed this issue by collectively agreeing to leave lots of comments which allowed for better readability and helped us assist each other easier. Another challenge we faced was implementing the motion tracking AI in a way that could actually help the user. Small mistakes would completely change the applicability of the software and figuring it all out was definitely tough. Sometimes simply changing our tester would completely change our perception on whether or not the app was working. We were able to conquer this by being thorough and dectuple checking our code to make sure we used the AI in the best way possible. We also universalized some of our code to make sure different users could get the motion tracking AI to work for them. Time was also a challenge and even though we feel like this app has a ton of potential, we weren't able to fully realize our dream under the strict deadline.

Accomplishments that we're proud of

Implementing the OpenCV AI in a way that benefits the user was the most satisfying accomplishment of the whole project. Calculating the different angles and getting the camera to work for the user in the first place was incredibly tough and took a lot of time and effort. Getting the buttons to work and styling our application was also incredibly rewarding, since learning the intricacies of Streamlit was no easy feat.

What we learned

In a general sense, we learned to work together as a team to accomplish a difficult task in a rather short amount of time. In a more specific sense, we learned how to use Streamlit to fulfill our vision and we became familiar with OpenCV's incredibly complex AI vision technology.

What's next for Form Focus

Improving the legitimacy of our form checking by optimizing our angle formulas and other functions is a big priority. Maybe inviting trainers to test the strictness of our form checking and give us advice on what metrics to improve. We'd like to expand the amount of exercises the app offers and build on what we can do with the data that the user gives us. Implementing another AI to help recommend the user exercises based on their data is something that we plan to do. Making the app look sleeker and more universally appealing is also a next step.