Abstract
Abstract
In digital imaging Low light image improvement is a crucial issue, with applications in medical imaging, surveillance and digital imaging. Images captured under substandard illumination usually appear dark and noisy: contrast is lower, hiding crucial details, while ISO (international Organization for Standardization) settings introduce grainy noise that devalue quality. These issues make images a problem for both human interpretation and automated vision system.
Traditional improvement methods such as histogram equalization and Retinex -based techniques enhance brightness but usually cause artifacts to boost noise. Deep learning approaches achieve strong results but require large datasets, heavy computation, and may fail to generalize well. By contrast, lightweight and interpretable solutions remain unexplored.
This research proposes a simple yet efficient framework that combines Gaussian denoising with Contrast-Limited Adaptive Histogram Equalization (CLAHE). Gaussian filtering restrains high – frequency noise while maintaining edges, and CLAHE improves local contrast without magnifying uniform regions. The combined methods aim to deliver clearer, more natural low light images. Evaluation will use quantitative metrics (PSNR, SSIM, MSE & LOE) and qualitative comparisons to exhibit improvements in visibility, detail preservation, and noise suppression.
Faculty Advisor
Dr. Anas Hourani
Department/Program
Computer Science
Submission Type
in-person poster
Date
3-5-2026
Rights
Copyright the Author(s)
Recommended Citation
Adesoji, Daniel
(2026)
"Enhancing Low-Light and Noisy Images Using Gaussian Denoising and CLAHE (Contrast-Limited Adaptive Histogram Equalization).,"
SACAD: Scholarly Activities: Vol. 2026, Article 19.
Available at:
https://scholars.fhsu.edu/sacad/vol2026/iss2026/19
