Inspiration

With an increasingly sedentary life-style due to COVID and an increasingly important need to perform fitness, we understand how difficult it can be to perform exercise when you're stuck at home, chocolate bars are within reach, and you don't know where to start. In order to lower the barrier into fitness, we created Fit-tionary that assists users with fitness searches and recommendations, allowing even novices to spring into action!

What it does

Using Jina, a powerful neural network based search system, Fit-tionary allows a user to find out anything they could possibly want to know about a given exercise. Don't know what an exercise you saw is called, what it's supposed to do, or how it should be properly performed? Show Fit-tionary a series of pictures, a gif, or a video of the exercise and let the power of machine learning help you learn everything you've ever wanted to know about the exercise. Know the name, but not what the exercise should look like? Fit-tionary can help! In fact, the Cockroachdb database Fit-tionary uses is constructed to allow it to search along many different indices, making the process of finding related exercises, form videos, and more painless and easy.

How we built it

We used a Python structure and the Jina AI neural search engine to recognize and classify images of various workouts and trained them to associate their names as well as the other details. Since Jina does not work in Windows based environments, we used an Ubuntu Web-based subsystem/terminal to run the backend of the system. We also used the CockroachDB database to store and contain exercise information, images, as well as algorithm training instructions for the Jina AI to utilize and call upon. For the frontend portion we utilized the Jina Box package.

Challenges we ran into

The two greatest challenges encountered as a team during the hacking phase were issues with engine initialization and difficulties with running the programs. None of us had ever used Linux before, so trying to working with WSL terminals and eventually having to obtain Ubuntu was certainly a challenge.

Accomplishments that we're proud of

We're happy to say that despite the lack of documentation and technical support we were able to get our individual tools initialized and running.

What's next for Fit-tionary

We believe there is a definite need for an application like Fit-tionary. We'd like to have a functioning system in place to provide inexperienced gym goers with custom workouts as well as motivation. Additionally other industries may benefit from using a dedicated database extraction system similar to ours. Going forward we plan to completely automate most functions of Fit-tionary, train the system even more using more images/videos, and expand the abilities/features of the app.

Check out the GITHUB and Youtube demo!

Team members: Adrian Best, Adithya Shankar, Ryan Walsh

Built With

Share this project:

Updates