FlowPortal: Residual-Corrected Flow for Training-Free Video Relighting and Background Replacement
Authors:
Preprint
Abstract
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. Project Page: https://gaowenshuo.github.io/FlowPortalProject/.Keywords
All Related Versions
This publication has 10 references indexed in Scilit:
- LumiNet: Latent Intrinsics Meets Diffusion Models for Indoor Scene Relighting2025
- MatAnyone: Stable Video Matting with Consistent Memory Propagation2025
- Bilateral Reference for High-Resolution Dichotomous Image SegmentationCAAI Artificial Intelligence Research, 2024
- Colorful Diffuse Intrinsic Image Decomposition in the WildACM Transactions on Graphics, 2024
- SwitchLight: Co-Design of Physics-Driven Architecture and Pre-training Framework for Human Portrait Relighting2024
- Adding Conditional Control to Text-to-Image Diffusion Models2023
- FateZero: Fusing Attentions for Zero-shot Text-based Video Editing2023
- Null-text Inversion for Editing Real Images using Guided Diffusion Models2023
- Neural Video Portrait Relighting in Real-time via Consistency Modeling2021
- RAFT: Recurrent All-Pairs Field Transforms for Optical Flow2020