Revisiting Image Fusion for Multi-Illuminant White-Balance Correction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Serrano-Lozano, David, Arora, Aditya, Herranz, Luis, Derpanis, Konstantinos G., Brown, Michael S., Vazquez-Corral, Javier
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917960978792448
author Serrano-Lozano, David
Arora, Aditya
Herranz, Luis
Derpanis, Konstantinos G.
Brown, Michael S.
Vazquez-Corral, Javier
author_facet Serrano-Lozano, David
Arora, Aditya
Herranz, Luis
Derpanis, Konstantinos G.
Brown, Michael S.
Vazquez-Corral, Javier
contents White balance (WB) correction in scenes with multiple illuminants remains a persistent challenge in computer vision. Recent methods explored fusion-based approaches, where a neural network linearly blends multiple sRGB versions of an input image, each processed with predefined WB presets. However, we demonstrate that these methods are suboptimal for common multi-illuminant scenarios. Additionally, existing fusion-based methods rely on sRGB WB datasets lacking dedicated multi-illuminant images, limiting both training and evaluation. To address these challenges, we introduce two key contributions. First, we propose an efficient transformer-based model that effectively captures spatial dependencies across sRGB WB presets, substantially improving upon linear fusion techniques. Second, we introduce a large-scale multi-illuminant dataset comprising over 16,000 sRGB images rendered with five different WB settings, along with WB-corrected images. Our method achieves up to 100\% improvement over existing techniques on our new multi-illuminant image fusion dataset.
format Preprint
id arxiv_https___arxiv_org_abs_2503_14774
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Revisiting Image Fusion for Multi-Illuminant White-Balance Correction
Serrano-Lozano, David
Arora, Aditya
Herranz, Luis
Derpanis, Konstantinos G.
Brown, Michael S.
Vazquez-Corral, Javier
Computer Vision and Pattern Recognition
White balance (WB) correction in scenes with multiple illuminants remains a persistent challenge in computer vision. Recent methods explored fusion-based approaches, where a neural network linearly blends multiple sRGB versions of an input image, each processed with predefined WB presets. However, we demonstrate that these methods are suboptimal for common multi-illuminant scenarios. Additionally, existing fusion-based methods rely on sRGB WB datasets lacking dedicated multi-illuminant images, limiting both training and evaluation. To address these challenges, we introduce two key contributions. First, we propose an efficient transformer-based model that effectively captures spatial dependencies across sRGB WB presets, substantially improving upon linear fusion techniques. Second, we introduce a large-scale multi-illuminant dataset comprising over 16,000 sRGB images rendered with five different WB settings, along with WB-corrected images. Our method achieves up to 100\% improvement over existing techniques on our new multi-illuminant image fusion dataset.
title Revisiting Image Fusion for Multi-Illuminant White-Balance Correction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.14774