Research map / Video / Video generation
DreamSwapV: Mask-guided Subject Swapping for Any Customized Video Editing
Weitao Wang, Zichen Wang, Hongdeng Shen, Yulei Lu, Xirui Fan, Suhui Wu, Jun Zhang, Haoqian Wang, Hao Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use frame-level token concatenation for reference injection instead of channel concatenation (ablation) | DreamSwapV · DreamSwapV-Benchmark (qualitative) | visual detail | Improved |
| use frame-level token concatenation for reference injection instead of cross-attention (ablation) | DreamSwapV · DreamSwapV-Benchmark (qualitative) | visual detail | Improved |
| use adaptive grid sizing for mask augmentation (ablation) | DreamSwapV · DreamSwapV-Benchmark (qualitative) | visual fidelity | Improved |
| add random cropping to reference augmentation (ablation) | DreamSwapV · training reference augmentation | hallucination and detail | Hurt |
| compare DreamSwapV with Kling 1.6 overall | DreamSwapV · DreamSwapV-Benchmark | Total Average: 68.63% -> 71.41% | Improved |
| compare DreamSwapV with Kling 1.6 on VBench indicators | DreamSwapV · DreamSwapV-Benchmark | VBench Average: 79.79% -> 80.44% | Improved |
| compare DreamSwapV with Kling 1.6 on background preservation | DreamSwapV · DreamSwapV-Benchmark | Background Preservation: 39.17% -> 52.49% | Improved |
Closest papers by meaning
- VideoSwap: Customized Video Subject Swapping with Interactive Semantic Point Correspondence (CVPR 2024 · similarity 0.89)
- OmniVCus: Feedforward Subject-driven Video Customization with Multimodal Control Conditions (NeurIPS 2025 · similarity 0.87)
- DreamVideo: Composing Your Dream Videos with Customized Subject and Motion (CVPR 2024 · similarity 0.86)
- PolyVivid: Vivid Multi-Subject Video Generation with Cross-Modal Interaction and Enhancement (NeurIPS 2025 · similarity 0.86)
- HOI-Swap: Swapping Objects in Videos with Hand-Object Interaction Awareness (NeurIPS 2024 · similarity 0.86)
- DreamSalon: A Staged Diffusion Framework for Preserving Identity-Context in Editable Face Generation (CVPR 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
