Using pretrained StableDiffusion for investigating the generation of the same object from a single image to a new view.

To detect action segments in videos using both temporal information, optical flow, and spatial information, RGB data.