Diversity-Preserved Distribution Matching Distillation for Fast Visual Synthesis
Paper • 2602.03139 • Published • 45
Image and Video Synthesis and Generation
This collection features VCLab's significant efforts in accelerating, distilling, and improving the image/video synthesis and generation models.
Note [arXiv 2026] Diversity-preserved DMD for fast synthesis. | Code: https://github.com/Multimedia-Analytics-Laboratory/dpdmd
CoCoEdit: Content-Consistent Image Editing via Region Regularized Reinforcement Learning
Paper • 2602.14068 • PublishedNote [arXiv 2026] Content-consistent editing via region-regularized RL. | Code: https://github.com/langmanbusi/CoCoEdit
DNAEdit: Direct Noise Alignment for Text-Guided Rectified Flow Editing
Paper • 2506.01430 • PublishedNote [NeurIPS 2025 Spotlight] Direct noise alignment for rectified flow editing. | Code: https://xiechenxi99.github.io/DNAEdit/
GPSToken: Gaussian Parameterized Spatially-adaptive Tokenization for Image Representation and Generation
Paper • 2509.01109 • Published • 1Note [NeurIPS 2025] Gaussian-parameterized spatial tokens. | Code: https://github.com/xtudbxk/GPSToken
InsViE-1M: Effective Instruction-based Video Editing with Elaborate Dataset Construction
Paper • 2503.20287 • PublishedNote [ICCV 2025] 1M-scale instruction-based video editing. | Code: https://github.com/langmanbusi/InsViE
Exact Feature Distribution Matching for Arbitrary Style Transfer and Domain Generalization
Paper • 2203.07740 • PublishedNote [ECCV 2022 Oral] Exact feature distribution matching. | Code: https://github.com/YBZh/EFDM