Light4D: Training-Free Extreme Viewpoint 4D Video Relighting

TL;DR

Light4D is a training-free framework that synthesizes temporally consistent 4D videos under new lighting and extreme viewpoints, overcoming data scarcity and maintaining geometric and appearance fidelity.

Contribution

It introduces Disentangled Flow Guidance and Temporal Consistent Attention to achieve high-quality, consistent 4D relighting without training data.

Findings

Achieves competitive temporal consistency and lighting fidelity.

Handles camera rotations from -90 to 90 degrees.

Eliminates appearance flickering through regularization.

Abstract

Recent advances in diffusion-based generative models have established a new paradigm for image and video relighting. However, extending these capabilities to 4D relighting remains challenging, due primarily to the scarcity of paired 4D relighting training data and the difficulty of maintaining temporal consistency across extreme viewpoints. In this work, we propose Light4D, a novel training-free framework designed to synthesize consistent 4D videos under target illumination, even under extreme viewpoint changes. First, we introduce Disentangled Flow Guidance, a time-aware strategy that effectively injects lighting control into the latent space while preserving geometric integrity. Second, to reinforce temporal consistency, we develop Temporal Consistent Attention within the IC-Light architecture and further incorporate deterministic regularization to eliminate appearance flickering. Extensive experiments demonstrate that our method achieves competitive performance in temporal consistency and lighting fidelity, robustly handling camera rotations from -90 to 90. Code: https://github.com/AIGeeksGroup/Light4D. Website: https://aigeeksgroup.github.io/Light4D.