-
Bicep curl Midway
-
Bicep Curl Fully up
Inspiration
In our modern information and technological world, physical health has become an essential focus for our society and an indispensable focus for our daily lives. Nearly every single one of us have seen sufferring that occurs from health issues and injuries, all of which could have been prevented with adequate exercise. However, when people often first begin to exercise, it is easy to injure yourself with incorrect form. If only we all had our own personal trainers. With our app FormFitness, we can transform the way exercise is done and increases the acessibility of health to all! Lets get moving!
What it does
FormFitness is an exercise assistant and tracker that quickly corrects form for various exercises. It'll track your various joints for the respective exercise, and make sure that your form is as perfect as can be. If not, it'll show you which direction you need to move before you continue, or show you that your form needs to be fixed. It also shows a percentage of the range of motion that you do for each exercise, and only counts your repititions if you correctly reach the full range of motion.
How we built it
We used openCV2 and mediaPipe with revolutionary machine learning technology. We then tracked the user's body and joints to determine their motion and the accuracy of their form. By customizing the tracking and form for different exercises, this app can track multiple different exercises and ample room to add more exercises. Our frontend allows the user to select any exercise, which then pops open a camera window that tracks and determines the form.
Challenges we ran into
Our first challenge was just simply the idea itself. We wanted to do something health with new cutting-edge technologies, but we weren't sure how to connect the two. After a half hour of discussion and brainstorming of several ideas, including an app that helps finding lost dogs, we settled on using machine learning to track one's exercise form and automate the role of physical trainers.
Second obstacle we faced was bridging the frontend with the backend, which used some unique packages. This took up most of our frontend developers time, forcing us to have a relatively minimalistic frontend.
Accomplishments that we're proud of
Bridging the front-end to the very different back-end Leveraging new technologies and applying them to old problems
What we learned
How to effectively use ML frameworks, and libraries to make a full stack web app How to effectively communicate through GitHub for easy version control and collaboration within our team
What's next
We plan to expand our frontend massively, by not only increasing its aestheticness but also by making it accessible to mobile phone users like ios and android. Furthermore, we plan to have a workout graph of reps done of each exercise that allows an user to easily track progress. In addition to these improvements, we plan to use machine learning to recommend exercises to be done based on the users current form and progress with other exercises. This would allow us to be a one stop shop for all exercise related requirements.
Log in or sign up for Devpost to join the conversation.