Steven Anderson

Evaluating how effective machine learning can be at aiding video game accessibility by using gesture recognition as the input method to control a game application. Using Tensorflow, I created a classification algorithm capable of classifying multiple hand gestures (E.g., open palm, fist, or thumbs up). I made a game to test the algorithm, where the gestures are used for input. The user must copy the gesture shown above an enemy's head and press the action button to defeat them, allowing progression through the game. Classic arcade shooter games inspired the game application; this allowed for a game that was not too fast-paced for gesture recognition to be counterproductive.