Reproduce “ComplexSync: High-Fidelity and Real-Time Lip Sync in Complex Scenarios” 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
Jiaran Cai, Xingpei Ma, Shenneng Huang
ComplexSync edits mouth movements to match speech while preserving identity and scene appearance, including under occlusion and changing lighting. It separates synchronization and fidelity learning, fuses their weights, aligns scene features, and distills generation to one denoising step. On ComplexSync-Val, it reports Occ-RC 95.20 versus DINet’s 57.74 and 72.73 FPS; these benchmark and throughput results do not guarantee universal performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.