摘要
使用图像增强方法和深度学习的方法可以提高低照度图像亮度,改善图像质量。文章首先对传统的低照度图像增强算法分类介绍,总结这些算法近年来的改进过程,然后重点介绍当下适用于低照度图像增强的网络模型,同时对这些网络结构和适用于该网络的部分方法进行梳理,最后介绍实验所需要的数据库与增强后图像的评价准则,提出了目前深度学习在该领域所面临的挑战,旨在为低照度图像增强的发展方向提供思考。
Image enhancement and deep learning can improve the brightness and quality of low illumination image.In this paper,the traditional low illumination image enhancement algorithms are introduced,and the improvement process of these algorithms in recent years is summarized.Then,the network models suitable for low illumination image enhancement are mainly introduced.At the same time,these network structures and some methods suitable for the network are sorted out.Finally,the database needed for the experiment and the evaluation criteria of the enhanced image are introduced,and the objectives are proposed The challenge of pre deep learning in this field aims to provide thinking for the development direction of low illumination image enhancement.
作者
燕雨洁
张煜朋
贾珍珠
苏红旗
Yan Yujie;Zhang Yupeng;Jia Zhenzhu;Su Hongqi(China University of mining and Technology(Beijing),Beijing 100083,China)
出处
《无线互联科技》
2021年第1期77-80,共4页
Wireless Internet Technology
关键词
低照度图像增强
深度神经网络
生成对抗网络
low illumination image enhancement
deep neural network
generating countermeasure network
作者简介
燕雨洁(1996-),女,安徽亳州人,硕士研究生,研究方向:图形图像处理。
