Inspiration
Our biggest inspiration was our individual shortcomings when it came to starting the gym. Amaya had severe anxiety that prevented her from stepping in a gym alone and since she had no one to show her what to do she was therefore completely stuck. Yahia was morbidly obese for a very long time and felt completely helpless which put him in a state of depression for a couple of years. Brendan, a Team USA figure skater was injured for an entire year due to improper lifting form. We know that people do not like to step out of there comfort zone, but we also know that there is a rising obesity epidemic in the United States. This app addresses both problems from the comfort of the user's own home.
What it does
Reps is an all around wellness app that uses AI to help you improve form in the gym, keep a record of past workouts, and track nutrition outside of the gym. The nutrition log tracks important macronutrients and keeps a count of our calorie intake throughout the day, and the more unique function of our app tracks your range of motion while performing exercises. Specifically, after you finish a set our app reports back how many full, and partial reps you performed. Full reps are counted when you hit 100% range of motion, and partial reps are recorded when you hit at least 80%, but not quite 100%.
How we built it
We used OpenCV and mediapipe to track different parts on a human body and using an algorithm devised by us to track when certain predetermined exercises were being done.
Challenges we ran into
Our two biggest challenges were getting mediapipe and openCV to fully work together how we wanted it to and getting the mongoDB server to work even though we were connecting on our phyiscal mobile devices using ngrok. The first problem was solved by converting all of our data into numpy arrays that could then be mapped to the values 0 to 100 thus giving us a percentage of how complete the reps were.
Accomplishments that we're proud of
In the 48 hours that we were given, we were able to create a full stack mobile application that is almost ready for deployment. To get the app ready for deployment, the main next steps would be to make it look much more aesthetic as well as get the backend to run faster. Most likely, we will be converting the backend fully to node.js/typescript because it is significantly faster. We would also be moving the database to mySQL.
What we learned
We learned how to utilize AI to analyze different human movements in a quantitative way, thus allowing us to scan for any individual movement we wanted. This was also all of our first time utilizing ngrok to connect to the computer's local host for testing but it proved to be extremely valuable and thus will be used very frequently in our future developement careers.
What's next for Reps
We will be refining the app and then applying to the next YCombinator cycle to see if we can get funding as an official startup. We have also already applied to Telora, a company founded at HackMIT that aims to fund upcoming startups by inviting them to work in florida with them over the summer. Within the next 3 weeks, the app will also be on the app store.
Log in or sign up for Devpost to join the conversation.