PoseNet Tetris

Utilized TensorFlow.js version of PoseNet for real-time human pose estimation to develop a functional Tetris game. Once the start button is hit, the entire game is played via the computer camera using PoseNet's machine trained human pose estimation built using Google's TensorFlow.js.

STACK: TensorFlow.js, PoseNet, React, Redux, React-Redux, Semantic UI

• Transformed minimally available Tensorflow PoseNet documentation to adapt with React.js to make manipulating the DOM scalable for a gaming experience.
• Developed pose algorithms to successfully translate human poses to into an interactive movable block experience.
• Developed javascript gaming theory, and front-end dynamic rendering using Redux in an agile team workflow environment.