FlowPortal: Residual-Corrected Flow for Training-Free Video Relighting and Background Replacement
Wenshuo Gao, Junyi Fan, Jiangyue Zeng, Shuai Yang
摘要
Video relighting with background replacement is a challenging task critical for applications in film production and creative media. Existing methods struggle to balance temporal consistency, spatial fidelity, and illumination naturalness. To address these issues, we introduce FlowPortal, a novel training-free flow-based video relighting framework. Our core innovation is a Residual-Corrected Flow mechanism that transforms a standard flow-based model into an editing model, guaranteeing perfect reconstruction when input conditions are identical and enabling faithful relighting when they differ, resulting in high structural consistency. This is further enhanced by a Decoupled Condition Design for precise lighting control and a High-Frequency Transfer mechanism for detail preservation. Additionally, a masking strategy isolates foreground relighting from background pure generation process. Experiments demonstrate that FlowPortal achieves superior performance in temporal coherence, structural preservation, and lighting realism, while maintaining high efficiency.
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了每一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
它引用的顶会 Paper22
- Directly Denoising Diffusion ModelsDan Zhang, Jingjing Wang, Feng LuoICML 2024 · 被引用 11,724 次
- Adding Conditional Control to Text-to-Image Diffusion ModelsLvmin Zhang, Anyi Rao, Maneesh AgrawalaICCV 2023 · 被引用 6,759 次
- FateZero: Fusing Attentions for Zero-shot Text-based Video EditingChenyang Qi, Xiaodong Cun, Yong Zhang, Chenyang Lei 等ICCV 2023 · 被引用 510 次
- Prompt-to-Prompt Image Editing with Cross-Attention ControlAmir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman 等ICLR 2023 · 被引用 361 次
- A Latent Space of Stochastic Diffusion Models for Zero-Shot Image Editing and GuidanceChen Henry Wu, Fernando De la TorreICCV 2023 · 被引用 141 次
相关 Paper
- AnyPortal: Zero-Shot Consistent Video Background ReplacementWenshuo Gao, Xicheng Lan, Shuai YangICCV 2025
- Light-a-Video: Training-Free Video Relighting via Progressive Light FusionYujie Zhou, Jiazi Bu, Pengyang Ling, Pan Zhang 等ICCV 2025 · 被引用 4 次
- Neural Video Portrait Relighting in Real-time via Consistency ModelingLongwen Zhang, Qixuan Zhang, Minye Wu, Jingyi Yu 等ICCV 2021 · 被引用 59 次
- Comprehensive Relighting: Generalizable and Consistent Monocular Human Relighting and HarmonizationJunying Wang, Jingyuan Liu, Xin Sun, Krishna Kumar Singh 等CVPR 2025
- Pixel Cube: Diffusion-based Portrait Video Relighting Through Realistic Lighting ReproductionYufan Zhang, Yu Ji, Ayo Ajiboye, Rundi Wu 等SIGGRAPH 2026
