๐‘๐ž๐ฌ๐ž๐š๐ซ๐œ๐กไธจShape-Consistent Video-to-Video Editing by Liu Chang๐‘๐ž๐ฌ๐ž๐š๐ซ๐œ๐กไธจShape-Consistent Video-to-Video Editing by Liu Chang

๐‘๐ž๐ฌ๐ž๐š๐ซ๐œ๐กไธจShape-Consistent Video-to-Video Editing

Liu Chang

Liu Chang

AI Researcher & First Author
๐Œ๐จ๐ญ๐ข๐ฏ๐š๐ญ๐ข๐จ๐ง & ๐‚๐ก๐š๐ฅ๐ฅ๐ž๐ง๐ ๐ž: Video editing with significant changing object shape normally conflict with source motion in existing studies. To address this critical issue, our paper StableV2V edits the first frame, aligns the new shape with motion and propagates the edited content across frames.
๐‘๐ž๐ฌ๐ฎ๐ฅ๐ญ๐ฌ & ๐•๐ž๐ง๐ฎ๐ž๐ฌ: Improved shape consistency and editing quality; introduced DAVIS-Edit for systematic evaluation. Published in IEEE Transactions on Circuits and Systems for Video Technology, 2026.
๐‹๐ข๐ง๐ค๏ผšhttps://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=11272911
Like this project

Posted Sep 15, 2026

StableV2V enables shape-consistent video editing and introduces DAVIS-Edit. Published in TCSVT, 2026.