Saved in:
Bibliographic Details
Main Authors: Huang, Wei, Huang, Joyce
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.07329
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918489260818432
author Huang, Wei
Huang, Joyce
author_facet Huang, Wei
Huang, Joyce
contents We introduce GC-ART (Global Curve Adaptive Rational Tone-mapping), a lightweight differentiable pre-processing module for robust image classification. GC-ART predicts an endpoint-pinned rational tone curve from per-channel soft histograms using a 643-parameter MLP, then applies the curve pointwise before the classifier. The module is trained end-to-end with cross-entropy and a soft monotonicity penalty. On CIFAR-10 with a CIFAR-style ResNet-18, GC-ART matches clean accuracy with the unenhanced baseline and other learned enhancers, improves over the baseline on multiplicative darkening, and achieves the best learned-method result on contrast corruption (48.45% vs. 46.27% for the baseline and 47.13% for Zero-DCE++). These results suggest that histogram-conditioned rational curves can learn useful global tone corrections, including contrast-expanding behavior, while preserving edge locations by construction through pointwise mapping. GC-ART also uses substantially fewer FLOPs than convolutional learned enhancers at 32 x 32. The current hyperparameters are untuned, leaving room for systematic improvement.
format Preprint
id arxiv_https___arxiv_org_abs_2605_07329
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle GC-ART: Global Learnable Second-Order Rational Tone Curves for Illumination Robustness
Huang, Wei
Huang, Joyce
Computer Vision and Pattern Recognition
We introduce GC-ART (Global Curve Adaptive Rational Tone-mapping), a lightweight differentiable pre-processing module for robust image classification. GC-ART predicts an endpoint-pinned rational tone curve from per-channel soft histograms using a 643-parameter MLP, then applies the curve pointwise before the classifier. The module is trained end-to-end with cross-entropy and a soft monotonicity penalty. On CIFAR-10 with a CIFAR-style ResNet-18, GC-ART matches clean accuracy with the unenhanced baseline and other learned enhancers, improves over the baseline on multiplicative darkening, and achieves the best learned-method result on contrast corruption (48.45% vs. 46.27% for the baseline and 47.13% for Zero-DCE++). These results suggest that histogram-conditioned rational curves can learn useful global tone corrections, including contrast-expanding behavior, while preserving edge locations by construction through pointwise mapping. GC-ART also uses substantially fewer FLOPs than convolutional learned enhancers at 32 x 32. The current hyperparameters are untuned, leaving room for systematic improvement.
title GC-ART: Global Learnable Second-Order Rational Tone Curves for Illumination Robustness
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.07329