arxiv preprint – Explorative Inbetweening of Time and Space

In this episode, we discuss Explorative Inbetweening of Time and Space by Haiwen Feng, Zheng Ding, Zhihao Xia, Simon Niklaus, Victoria Abrevaya, Michael J. Black, Xuaner Zhang. The paper presents a method for generating video sequences from just a starting and ending frame, called bounded generation, by utilizing a new sampling strategy named Time Reversal Fusion. This strategy merges the forward and backward denoising processes guided by the start and end frames to create videos that naturally transition between the two given frames, enable smooth inbetweening of motion, and create looping videos when the frames are the same. Time Reversal Fusion is shown to outperform previous methods in terms of generating complex movements and 3D-consistent visuals without additional training on a model.