Reproduce “ReMoVe: Refining a Video Model's Own Motion for Coherent Human Video Generation” 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
Jie Li, Muxin Pu, Hongyi Cai, Mingkang Dong, Jianyuan Guo, Tao Huang
ReMoVe refines human-motion videos by using a video generator’s output as a draft. It recovers body motion, anchors reliable poses while a motion model fills unreliable stretches, then renders mesh and scene-depth conditions to guide video generation. On 152 Motion-X++ videos, it scores 92.66 on VBench-2.0 Human Anatomy versus 90.88 for baseline, a benchmark gain, not proof of physical correctness.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.