Inspiration
As computer science students we are often busy hunched over our computers making our ideas into projects. Its really helpful to take a break and stretch every once in a while, so we create this application to promote yoga and stretching.
What it does
Users can select which pose they want to perform on our app, and the camera will open up and our awesome and cool mascot will indicate how accurate the user's pose is. The user will also receive helpful advice on how to perform the pose they chose.
How we built it
We used OpenCV and Mediapipe to capture the camera input and obtain the coordinates of a person in front of the camera. We then created helper functions to help us calcuate the angles between 8 major joints in the body. We used images of people doing the pose to help us find the ideal angles, and we created more helper functions to calculate the accuracy of the pose given the coordinates of a person.
Challenges we ran into
We had some issues with calling the LLM we chose for our advice. We also had some issues with merge conflicts and communication between frontend and backend, but we're glad we were able to resolve those in the end!
Accomplishments that we're proud of
We are proud of how we were able to connect the dataflow from the backend to an image in the frontend as our first project with computer vision. We were also happy to create actionable health feedback from camera input.
What we learned
We learned a lot more about how OpenCV and Mediapipe worked, and also more about how to integrate frontend and backend together.
What's next for FlowState
We would like to be able to add more yoga poses to our app, and we would also like to be able for users to create a guided yoga routine, and to be able to save the routines.
Log in or sign up for Devpost to join the conversation.