RelightFormer: Feed-forward Generative Transformer for Multiview Object Relighting
Summary
RelightFormer introduces a feed-forward generative Transformer for direct single- and multi-view image relighting, using cross-attention for illumination injection and permutation-invariant encodings for unordered views, trained on a massive synthetic dataset to achieve state-of-the-art visual quality.
View Cached Full Text
Cached at: 09/09/26, 08:29 AM
Paper page - RelightFormer: Feed-forward Generative Transformer for Multiview Object Relighting
Source: https://huggingface.co/papers/2609.07414
Abstract
A feed-forward generative Transformer enables direct single- and multi-view image relighting by injecting target illumination via cross-attention and processing unordered views with permutation-invariant encodings, trained on a large synthetic dataset.
Image relighting is traditionally tackled via complex inverse rendering pipelines, which suffer from ill-posed optimization, or single-image generative models that ignore crucial multi-view cues necessary for understanding 3D geometry and material interactions. To address these limitations, we introduce afeed-forward generative Transformerfor direct single- and multi-view image relighting that entirely bypasses explicit intrinsic property estimation. Adapted from a video foundation model, our architecture features alatent illumination modulethat dynamically injects target environment maps into spatial features viacross-attention. Furthermore, we employpermutation-invariant positional encodingsto symmetrically process unordered multi-view inputs without sequential bias. To train this robust data-driven model, we construct the massiveLaval Objaverse Dataset(LOD), comprising 90K objects and 39K unique illuminations. Extensive experiments demonstrate state-of-the-art visual quality, photorealistic relighting quality, and strong zero-shot generalization across single-view, multi-view, and novel-view relighting tasks.
View arXiv pageView PDFProject pageGitHubAdd to collection
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2609.07414 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2609.07414 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2609.07414 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
GenRecon: Bridging Generative Priors for Multi-View 3D Scene Reconstruction
GenRecon introduces a method for 3D scene reconstruction that integrates generative 3D priors with multi-view image conditioning, achieving high-fidelity, editable mesh reconstructions of indoor environments and outperforming existing methods by 16%.
WildRelight: A Real-World Benchmark and Physics-Guided Adaptation for Single-Image Relighting
This paper introduces WildRelight, a new real-world benchmark dataset for single-image relighting that addresses the gap between synthetic and natural scenes. It proposes a physics-guided adaptation framework using diffusion posterior sampling and test-time adaptation to improve model performance on real-world data.
RayDer: Scalable Self-Supervised Novel View Synthesis from Real-World Video
RayDer is a unified feed-forward transformer that consolidates camera estimation, scene reconstruction, and rendering for self-supervised novel view synthesis from real-world video, achieving clean power-law scaling and strong zero-shot performance.
RenderFormer-V2: Neural Rendering with Heterogeneous Scene Primitives
RenderFormer-V2 is a transformer-based neural rendering model that handles diverse light-transport effects using a two-stage sequence-to-sequence architecture with improved attention mechanisms and support for heterogeneous scene primitives.
Relit-LiVE: Relight Video by Jointly Learning Environment Video
This paper introduces Relit-LiVE, a novel video relighting framework that produces physically consistent results without requiring camera pose information by using raw reference images and joint environment video prediction.