Time for squats with Shia LaBeouf and AI!! ML Image Recognition using Google's Teachable Machine

A HACKUCI 2020 submission.

In the modern era of extreme developer laziness and muscular atrophy, we were inspired to move our butts.

git off your ass forces lazy techies to bounce their stagnant tushies at set intervals of time. Users must complete squats in order to regain access to their web browser. Big brother Tensorflow ML models, and the omnipotent, compelling power of Shia LaBeouf force users into non-consensual activity.

git off your ass is built with substantial dependence on the OTC drug known as "Caffeine". It is trained with Google's Tensorflow-based Teachable Machine and the skinny (and now sore) legs of multiple UCI Computer Science majors.

Training Image models to recognize movement is highly dependent on variations in light and color. Many thousands of pictures were taken to train our rudimentary Squat-model. Getting this right took many iterations and struggles.

We're most proud of the practical applicability, the technical complexity, and comedic value that the app provide the UCI community.

We learned how to train complex computer vision ML models. We used new ML5 and P5js tools to render real-time video Machine Learning analysis. Squatting in-between workflow improves mental efficiency and hacker happiness.

More frequent exercise between hacking sessions! Even greater love for Shia LaBeouf (if that's humanly possible)!