Reproduce “MoviEdit: Long-form Multi-Shot Video Editing Made Easy” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Suhyeon Lee, Zhengang Li, Soo Ye Kim, Peihao Zhu, Jong Chul Ye, Chun-Hao Paul Huang, Duygu Ceylan, Yan Kang and 1 more
MoviEdit edits long, multi-shot videos by first creating a low-resolution whole-video draft with bidirectional context, then rendering detailed chunks guided by that draft, source chunks, and recent outputs. Its training data includes 86,619 multi-shot editing pairs. On MoviEdit-Bench, text-guided video-to-video edit success is 0.790 versus 0.764 for Runway Aleph 2.0. The benchmark uses generated videos and VLM-based frame evaluation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.