[ AI ] dancer
Machine learning analysis of dance improvisation.
Performer: Evangelos Biskas
The objective of this project is to create a Machine Learning algorithm capable of generating a video composition in real time, in response to music and based on recorded material of Biskas's improvisation techniques.
Neural Network interactions to music.
Neural Network interactions to music.
This NN was created using audio MEL coefficients as an input signal, and the skeleton keypoints coordinates taken from 20 hours of Evangelos's improvisation videos.
The Neural Net is now trying to create a new "skeleton" as a reaction to a real-time music input.