mlsanityData to 14 Oct 2026

SAFE: Scene-Aware Feature Modulation for Color Constancy with Learned Color Space in Pure-Color Scenes

arXiv 2608.13967 ↗First posted 14 Aug 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Color constancy on pure-color scenes is challenging: when most pixels share a narrow band of hues, every chromaticity-based cue collapses to a single point and standard estimators become ambiguous. We propose a compact framework that couples two innovations: (i) SAFE, a Scene-Aware FeaturE modulation network that organizes illumination cues into a structured four-token representation, which is then selectively reweighted based on scene complexity features; (ii) the Learned Color Space (LCS), a scene-dependent chromaticity normalization that directly addresses the chromaticity collapse problem for pure-color scenes. Experiment results show that SAFE consistently improves performance in pure-color scenes. Compared to the best-performing baseline in each metric, it reduces the mean angular error by 10%, the best-25% error by 20%, and the worst-25% error by 5.8%.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 31 of its 65 references that are in the map. First posted on arXiv 14 Aug 2026.

Builds on

31 of 65 references.

References · 31

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2016Gaussian Error Linear Units (GELUs)cited by 3,502
  3. 2017FiLM: Visual Reasoning with a General Conditioning Layercited by 2,333
  4. 2018Recovering Realistic Texture in Image Super-resolution by Deep Spatial Feature Transformcited by 371
  5. 2023Spatially-Adaptive Feature Modulation for Efficient Image Super-Resolutioncited by 57
  6. 2020Deep White-Balance Editingcited by 54
  7. 2022Realistic Blur Synthesis for Learning Image Deblurringcited by 53
  8. 2016Fast Fourier Color Constancycited by 37
  9. 2022Multi-stage image denoising with the wavelet transformcited by 37
  10. 2015Convolutional Color Constancycited by 36
  11. 2021IDR: Self-Supervised Image Denoising via Iterative Data Refinementcited by 36
  12. 2021Semantically Contrastive Learning for Low-light Image Enhancementcited by 31
19 more
  1. 2020Cross-Camera Convolutional Color Constancycited by 30
  2. 2023MM-BSN: Self-Supervised Image Denoising for Real-World with Multi-Mask based on Blind-Spot Networkcited by 30
  3. 2021CLCC: Contrastive Learning for Color Constancycited by 29
  4. 2019On Finding Gray Pixelscited by 26
  5. 2023ParamISP: Learned Forward and Inverse ISPs using Camera Parameterscited by 23
  6. 2019Cascading Convolutional Color Constancycited by 22
  7. 2021Auto White-Balance Correction for Mixed-Illuminant Scenescited by 22
  8. 2024Attentive Illumination Decomposition Model for Multi-Illuminant White Balancingcited by 8
  9. 2025GCC: Generative Color Constancy via Diffusing a Color Checkercited by 8
  10. 2025CCMNet: Leveraging Calibrated Color Correction Matrices for Cross-Camera Color Constancycited by 8
  11. 2025Revisiting Image Fusion for Multi-Illuminant White-Balance Correctioncited by 5
  12. 2024Pixel-Wise Color Constancy via Smoothness Techniques in Multi-Illuminant Scenescited by 3
  13. 2025Integral Fast Fourier Color Constancycited by 3
  14. 2025Efficient RAW Image Deblurring with Adaptive Frequency Modulationcited by 2
  15. 2025Perception-Inspired Color Space Design for Photo White Balance Editingcited by 2
  16. 2026RL-AWB: Deep Reinforcement Learning for Auto White Balance Correction in Low-Light Night-time Scenescited by 2
  17. 2018A neural field model for color perception unifying assimilation and contrastcited by 1
  18. 2025Multi-illuminant Color Constancy via Multi-scale Illuminant Estimation and Fusioncited by 1
  19. 2026Why Low-Light Cameras Go Color Blind: Removing Color Bias in Raw Denoisingcited by 1

Built on by

Not cited yet.