Shavasna

• 

  Dropdown menu to select pose

Inspiration

75% of people in the Bay Area work in the tech industry. This means facing back pain from sitting at their desk all day, and large amounts of stress working in a highly competitive environment.

What it does

Shavasna is an AI-powered, at home yoga teacher. Users are shown different yoga poses and then are monitored through their camera to provide feedback on whether they are performing the poses correctly.

How we built it

We trained and tested our model on live images that we found online, and then used OpenCV to detect our camera movement and the Mediapipe library for keypoint detection.

Challenges we ran into

This was our first time working with OpenCV, a popular computer vision library, and we had trouble training our data to match realtime camera movement.

Accomplishments that we're proud of

Just being able to make this project as a whole! There were a lot of moments when we encountered new bugs that we hadn't seen before, but we pushed through and are very happy with our final product.

What we learned

Aside from learning how to use new libraries such as OpenCV, mediapipe, and PyQt5, we also learned some soft skills such as how to effectively delegate tasks, manage multiple people working in the same repository, and create a good presentation.

What's next for Shavasna

We'd like to extend to adding aquatic physical therapy features, and possibly add personalized feedback based on the type of pain a person is experiencing.

Built With

• mediapipe
• opencv
• python