You can view the videos using full screen to see more details.
If the videos cannot be correctly loaded, please look into the folders we have attached.

Video Visualization Format

Dataset: GIF-Faces (dithered)      Model: CCDNet      Loss: all losses enabled

Dataset: GIF-Faces (not dithered)      Model: CCDNet      Loss: all losses enabled

Dataset: GIF-Moments (not dithered)      Model: CCDNet      Loss: all losses enabled

Dataset: GIF-Faces (not dithered)      Model: CCDNet      Loss: all losses enabled

Temporal Downsampling Factor: 2

Temporal Downsampling Factor: 4

Temporal Downsampling Factor: 8