Video and Temporal Diffusion

3D U-Net Architecture

Convolutional neural network structure adapted for video data, combining encoder-decoder paths with 3D residual connections to effectively capture multi-scale spatial and temporal contexts during denoising.

← Zurück