- Utilize motion capture data in .npy format with shape (# joints, # timesteps, # dimensions) to visualize dance sequences.
- Construct a 3D plotting function to depict dancer movements over time, with joints represented as points.
- Train a generative model like a Variational Autoencoder (VAE) with LSTM layers to generate short dance phrases.
- Compare input sequences from a test set with their decoded counterparts using the VAE model.
- Generate new dance sequences using the trained model and visualize them alongside real sequences for evaluation.
-
Notifications
You must be signed in to change notification settings - Fork 0
Variational Autoencoder (VAE) with LSTM layers to generate short dance phrases. Compare input sequences from a test set with their decoded counterparts using the VAE model. Generate new dance sequences using the trained model and visualize them alongside real sequences for evaluation.
CoderSahel/AI-Enabled-Choreography
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
About
Variational Autoencoder (VAE) with LSTM layers to generate short dance phrases. Compare input sequences from a test set with their decoded counterparts using the VAE model. Generate new dance sequences using the trained model and visualize them alongside real sequences for evaluation.