Home

This work describes a system, USquat, that utilizes computer vision and machine learning for understanding and analysis of a particular exercise, squatting, as proof of concept for the creation of a detective and corrective exercise assistant. Squatting was chosen as it is a complicated form of exercise and is often mis performed.

* See the USquatTrain repository for code and information about the creation of the core USquat Tensorflow lite LSTM model used in the USquat application.