FreeLit: Paired-Free Indoor Relighting via Physics-Guided Diffusion

2026-07-15

AI Summary

This work addresses the challenging problem of indoor scene relighting, which is complicated by geometric complexity and the coupling of local illumination. Existing approaches rely on paired multi-illumination data and exhibit limited robustness under low-light conditions. To overcome these limitations, we propose a controllable relighting framework that operates without paired supervision, leveraging a physics-guided illumination prior to generate structured lighting maps and pseudo-labels that drive a diffusion model for explicit control over light source position, color, and intensity. Key innovations include a physics-informed intrinsic decomposition strategy integrated with diffusion-based generation, structure-aware distillation, an illumination-invariant reflectance constraint, and consistency regularization. We also introduce a novel evaluation metric tailored to controllability. Our method achieves physically consistent, stable, and accurate relighting without paired training data, significantly improving robustness in low-light scenarios and alignment with user-specified lighting parameters.

Abstract

Image-based indoor scene relighting remains challenging due to the complex interplay between cluttered geometry and local illumination, requiring precise modeling of light position, color, and intensity. Existing data-driven methods implicitly learn this relationship via paired multi-illumination datasets. Nevertheless, this data is costly and fails to scale, which is essential for accurate light-source-level control. Conversely, inverse-rendering methods reduce the data dependency by incorporating physical priors; however, they lack the robustness of intrinsic estimation in challenging conditions. In this paper, we present FreeLit, a paired-free framework for controllable indoor relighting that explicitly manipulates light-source location, color, and intensity. Instead of relying on paired supervision, we construct a physics-guided illumination prior from intrinsic scene properties, generating a structured lightmap along with a pseudo-relit image to guide diffusion-based synthesis. To address instability in intrinsic estimation, especially in low-light scenes, we introduce a relighting-guided intrinsic stabilization strategy that enforces illumination-invariant reflectance through structure-aware distillation and consistency constraints. Furthermore, we propose controllability-oriented evaluation metrics to quantify alignment with user-specified illumination color and intensity. Experimental results demonstrate that FreeLit achieves stable, physically consistent, and controllable relighting, with improved robustness in low-light indoor scenes, without requiring paired supervision.

Problem

• indoor relighting
• light-source control
• paired-free learning
• intrinsic estimation
• illumination modeling

Innovation

• paired-free relighting
• physics-guided diffusion
• intrinsic decomposition
• controllable illumination
• relighting-guided stabilization