Exercise Evaluator

• 

  The ending pose of the pushup evaluated by our model.

• 

  The starting pose of the pushup evaluated by our model.

Inspiration

We were inspired to create an exercise evaluator because we are interested in computer vision and machine learning and wanted to create our own model that combines the two together.

What it does

Our model evaluates a user's exercise video and determines how well their form is.

How we built it

We used Python, PyTorch, MediaPipe and more Python libraries in Google Colab to create our model.

Challenges we ran into

We had issues with combining the computer vision model and machine learning model together after building them separately. We had to adapt data from the computer vision model to input it into the classification model.

Accomplishments that we're proud of

We are proud of learning PyTorch and other new technologies that helped us build this model from scratch.

What we learned

We learned how to make machine learning and computer vision models using Python from scratch particularily using PyTorch.

What's next for Exercise Evaluator

We want to improve the accuracy of our model, make it more user friendly with an app, and expand our model to include more exercises.

Built With

• mediapipe
• python
• pytorch