
Motivation & Challenge: Reshaping objects in video editing conflicts with the source motion. StableV2V edits the first frame, aligns it with the motion, then propagates the edit across frames.
Results & Venues: Outperformed state-of-the-art methods in visual consistency and inference efficiency. Introduced DAVIS-Edit, a 160-case benchmark with similar/changing-shape splits for text- and image-based prompts.
Published in IEEE Transactions on Circuits and Systems for Video Technology, 2026.
Link: https://ieeexplore.ieee.org/document/11272911
GitHub: https://github.com/AlonzoLeeeooo/StableV2V
1 like · 0 views · Sep 15, 2026
Motivation & Challenge: Reshaping objects in video editing conflicts with the source motion.