中国邮电高校学报(英文) ›› 2022, Vol. 29 ›› Issue (2): 52-62.doi: 10.19682/j.cnki.1005-8885.2022.0001

• Image Processing • 上一篇    下一篇

Low-light image enhancement algorithm using a residual network with semantic information

段炼,唐贵进   

  1. 南京邮电大学
  • 收稿日期:2021-01-19 修回日期:2021-07-10 出版日期:2022-04-26 发布日期:2022-04-26
  • 通讯作者: 唐贵进 E-mail:tanggj@njupt.edu.cn
  • 基金资助:
    南京邮电大学科研基金项目;江苏省科协提升计划项目

Low-light image enhancement algorithm using a residual network with semantic information

DUAN Lian1,   

  • Received:2021-01-19 Revised:2021-07-10 Online:2022-04-26 Published:2022-04-26
  • Supported by:
    This work was supported by the Promotion Plan of Jiangsu Association for Science and Technology (TJ215039), and the Research Foundation of Nanjing University of Posts and Telecommunications (NY219076).

摘要:

Aiming to solve the poor performance of low illumination enhancement algorithms on uneven illumination images, a low-light image enhancement (LIME) algorithm based on a residual network was proposed. The algorithm constructs a deep network that uses residual modules to extract image feature information and semantic modules to extract image semantic information from different levels. Moreover, a composite loss function was also designed for the process of low illumination image enhancement, which dynamically evaluated the loss of an enhanced image from three factors of color, structure, and gradient. It ensures that the model can correctly enhance the image features according to the image semantics, so that the enhancement results are more in line with the human visual experience. Experimental results show that compared with the state-of-the-art algorithms, the semantic-driven residual low-light  network (SRLLN) can effectively improve the quality of low illumination images, and achieve better subjective and objective evaluation indexes on different types of images.

关键词: image enhancement, convolutional neural network (CNN), residual learning, image semantic

Abstract:

Aiming to solve the poor performance of low illumination enhancement algorithms on uneven illumination images, a low-light image enhancement (LIME) algorithm based on a residual network was proposed. The algorithm constructs a deep network that uses residual modules to extract image feature information and semantic modules to extract image semantic information from different levels. Moreover, a composite loss function was also designed for the process of low illumination image enhancement, which dynamically evaluated the loss of an enhanced image from three factors of color, structure, and gradient. It ensures that the model can correctly enhance the image features according to the image semantics, so that the enhancement results are more in line with the human visual experience. Experimental results show that compared with the state-of-the-art algorithms, the semantic-driven residual low-light  network (SRLLN) can effectively improve the quality of low illumination images, and achieve better subjective and objective evaluation indexes on different types of images.

Key words: image enhancement, convolutional neural network (CNN), residual learning, image semantic

中图分类号: