Research map / Video / Video generation
AnyPortal: Zero-Shot Consistent Video Background Replacement
Wenshuo Gao, Xicheng Lan, Shuai Yang
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 |
|---|---|---|---|
| add text-guided IC-Light second step (two-step harmonization) (ablation) | AnyPortal · ablation set | Fram-Acc: 0.966 -> 0.973 | Improved |
| add text-guided IC-Light second step (two-step harmonization) (ablation) | AnyPortal · ablation set | Tem-Con: 0.989 -> 0.993 | Improved |
| add text-guided IC-Light second step (two-step harmonization) (ablation) | AnyPortal · ablation set | Mtn-Psrv: 0.987 -> 0.987 | No clear effect |
| add consistency enhancement stage with video diffusion (ablation) | AnyPortal · ablation set | Tem-Con: 0.961 -> 0.993 | Improved |
| add consistency enhancement stage with video diffusion (ablation) | AnyPortal · ablation set | ID-Psrv: 0.353 -> 0.313 | Improved |
| add Refinement Projection Algorithm (ablation) | AnyPortal · ablation set | ID-Psrv: 0.371 -> 0.313 | Improved |
| add Refinement Projection Algorithm (ablation) | AnyPortal · ablation set | Tem-Con: 0.987 -> 0.993 | Improved |
| add cross-frame attention to IC-Light (ablation) | AnyPortal · qualitative | inter-frame appearance consistency | Improved |
| naive decode-refine-encode instead of RPA projection (ablation) | AnyPortal · qualitative | background quality | Hurt |
| AnyPortal vs per-frame IC-Light | video background replacement · video background replacement (30 test samples) | Tem-Con: 0.945 -> 0.993 | Improved |
| AnyPortal vs DAS + IC-Light | video background replacement · video background replacement (30 test samples) | ID-Psrv: 0.364 -> 0.313 | Improved |
| AnyPortal vs per-frame IC-Light | video background replacement · video background replacement (30 test samples) | Fram-Acc: 0.983 -> 0.973 | Hurt |
Later papers that compared with this method
1 later accepted paper reports AnyPortal as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- FlowPortal: Residual-Corrected Flow for Training-Free Video Relighting and Background Replacement (CVPR 2026 · better on 6 of 6 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- IllumiCraft: Unified Geometry and Illumination Diffusion for Controllable Video Generation (NeurIPS 2025 · similarity 0.86)
- AnyDoor: Zero-shot Object-level Image Customization (CVPR 2024 · similarity 0.86)
- Zero-shot Video Restoration and Enhancement Using Pre-Trained Image Diffusion Model (AAAI 2025 · similarity 0.86)
- Ground-A-Video: Zero-shot Grounded Video Editing using Text-to-image Diffusion Models (ICLR 2024 · similarity 0.85)
- MotionCraft: Physics-Based Zero-Shot Video Generation (NeurIPS 2024 · similarity 0.85)
- AnyScene: Customized Image Synthesis with Composited Foreground (CVPR 2024 · similarity 0.85)
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.
