Hand Rhythm
Team IMAGE
This demonstrator presents a use of deep learning in computer vision.
From a video stream of somebody, the dominant hand (right or left) is detected using pre-trained neural networks.
The hand orientation recognition is performed online in order to detect the user interaction with the game.
To do this, the player makes gestures (up👆, down👇, left👈, right👉) with its hand (left or right) to activate a column of musical notes.
During the game, the goal is to activate the right column when a note passes through the hit zone in order to score points.
This demonstrator uses the TensorFlowJS technology.
Contributor
Brice Ledanois