Light-A-Video Progressive Light Fusion

A training-free video relighting pipeline that repurposes an existing image relighting model with a Consistent Light Attention module across frames, then linearly blends source and relit appearance over time (Progressive Light Fusion) to remove flicker without any video-specific training.

Properties

family: image-to-video
modality: video -> per-frame relit candidates -> temporally fused relit video
whyMultiModel: A still-image relighting model alone produces per-frame lighting-source drift when run independently on each frame; a second temporal-fusion mechanism operating over the video diffusion's attention and output space is required to enforce cross-frame consistency.
steps: [object Object], [object Object], [object Object]
controls: light source/text prompt for target illumination, blending ratio in PLF, base video diffusion backbone choice (e.g. VideoCrafter2, AnimateDiff)
exampleStack: IC-Light (per-frame) -> Light-A-Video CLA (cross-frame attention) -> Light-A-Video PLF (temporal blend)
useCases: [relighting UGC or archival video without retraining][time-of-day changes for existing video clips][consistent studio-light simulation on handheld footage]
pitfalls: training-free temporal fusion can under-correct for large lighting-direction changes, causing residual low-frequency flicker on fast camera motion
source: Light-A-Video: Training-free Video Relighting via Progressive Light Fusion, arXiv:2502.08590 (ICCV 2025)

Provenance

Source: https://arxiv.org/abs/2502.08590

Light-A-Video Progressive Light Fusion

Properties

family: image-to-video
modality: video -> per-frame relit candidates -> temporally fused relit video
whyMultiModel: A still-image relighting model alone produces per-frame lighting-source drift when run independently on each frame; a second temporal-fusion mechanism operating over the video diffusion's attention and output space is required to enforce cross-frame consistency.
steps: [object Object], [object Object], [object Object]
controls: light source/text prompt for target illumination, blending ratio in PLF, base video diffusion backbone choice (e.g. VideoCrafter2, AnimateDiff)
exampleStack: IC-Light (per-frame) -> Light-A-Video CLA (cross-frame attention) -> Light-A-Video PLF (temporal blend)
useCases: [relighting UGC or archival video without retraining][time-of-day changes for existing video clips][consistent studio-light simulation on handheld footage]
pitfalls: training-free temporal fusion can under-correct for large lighting-direction changes, causing residual low-frequency flicker on fast camera motion
source: Light-A-Video: Training-free Video Relighting via Progressive Light Fusion, arXiv:2502.08590 (ICCV 2025)

Provenance

Source: https://arxiv.org/abs/2502.08590