The goal of this project is to translate human choreography/movement to virtual avatars, with AI-generated backgrounds using StreamDiffusion (which ultimately may lead to AI-generated music videos in the future).

Progress Update 1:

Currently have media pipe working with joints being displayed on a black background with white dots. Testing to input this raw joints to StreamDiffusion.

Future paths:

Somehow create an input format using this so that StreamDiffusion can convert this to a pose on an avatar.