HVI-CoLIE: Color-Aware Context-Based Low-Light Image Enhancement via Neural Implicit Representations and Illumination-Guided Restoration
Ingegneria Sismica · 10.06.2026 · 42 (4) · 1-14
DOI: 10.65102/is20261288 ↗Kurz gefasst
Fachartikel in der Zeitschrift Ingegneria Sismica, online erschienen am 10.06.2026. Verfasst von Chaoqun Wang.
Abstract (Originalsprache)
Current low-light image enhancement methods still suffer from poor visual perception, color bias, and over-exposure caused by aggressive brightening when dealing with unknown illumination conditions. In this paper, we propose a zero-shot method termed HVI-CoLIE, which redefines the enhancement process through mapping the 2D coordinates of an underexposed image to its HVI intensity component, while using the restored intensity to guide chromatic restoration and adaptive gamma correction to control over-exposure. We transform the image to HVI space and propose an implicit neural function combined with an illumination-guided Chromatic Restoration Module. Moveover, We introduce an adaptive decision criterion to effectively alleviate the over-exposure problem commonly observed in current zero-shot low-light enhancement methods. Through comprehensive evaluations, we investigate the characteristics of our proposed framework and demonstrate its advantages in image quality and scene adaptability. Moreover, we further evaluate its performance on downstream tasks under low-light scenarios, highlighting the practical utility of HVI-CoLIE.
Zitieren
APA
Wang, C. (2026). HVI-CoLIE: Color-Aware Context-Based Low-Light Image Enhancement via Neural Implicit Representations and Illumination-Guided Restoration. Ingegneria Sismica, 42(4), 1-14. https://doi.org/10.65102/is20261288
Vancouver
Wang C. HVI-CoLIE: Color-Aware Context-Based Low-Light Image Enhancement via Neural Implicit Representations and Illumination-Guided Restoration. Ingegneria Sismica. 2026;42(4):1-14. doi:10.65102/is20261288
BibTeX
@article{Wang2026,
author = {Wang, Chaoqun},
title = {HVI-CoLIE: Color-Aware Context-Based Low-Light Image Enhancement via Neural Implicit Representations and Illumination-Guided Restoration},
journal = {Ingegneria Sismica},
year = {2026},
volume = {42},
number = {4},
pages = {1--14},
doi = {10.65102/is20261288}
}
EUVIMED