DEEP LEARNING-BASED LOW LIGHT IMAGE ENHANCEMENT FOR IMPROVED VISIBILITY

Abstract

Low-light images often suffer from poor visibility, low contrast, and high noise, which significantly affect the performance of computer vision systems and visual perception. This paper, Deep Learning-Based Low Light Image Enhancement for Improved Visibility, presents an intelligent framework that employs deep learning techniques to enhance image quality in challenging illumination conditions. The proposed model utilizes a convolutional neural network (CNN) architecture trained on paired low-light and normal-light datasets to learn complex illumination mappings and restore brightness, contrast, and detail. In addition, the system integrates noise reduction and color correction modules to produce visually pleasing and realistic outputs while preserving natural textures. Experimental results demonstrate that the deep learning-based enhancement model achieves superior performance compared to traditional histogram equalization and Retinex-based methods, in terms of both quantitative metrics (PSNR, SSIM) and subjective visual quality. This approach enhances image visibility, making it highly beneficial for applications in surveillance, autonomous driving, medical imaging, and low-light photography.