2025/05/03 by Anushka Agarwal, Agarwal, Anushka, Hassan, Muhammad Yusuf +2 · 1 citation
Computer Science · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Music Technology and Sound Studies #Music and Audio Processing #Speech and Audio Processing
paper · pdf · doi:10.48550/arxiv.2505.01928
openalex publication_date 2025/05/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We introduce GenSync, a novel framework for multi-identity lip-synced video synthesis using 3D Gaussian Splatting. Unlike most existing 3D methods that require training a new model for each identity , GenSync learns a unified network that synthesizes lip-synced videos for multiple speakers. By incorporating a Disentanglement Module, our approach separates identity-specific features from audio representations, enabling efficient multi-identity video synthesis. This design reduces computational overhead and achieves 6.8x faster training compared to state-of-the-art models, while maintaining high lip-sync accuracy and visual quality.