A novel gradient and semantic-aware transformer network for low-light image enhancement: A novel gradient and semantic-aware...: T. Zhan et al.
Saved in:
| Title: | A novel gradient and semantic-aware transformer network for low-light image enhancement: A novel gradient and semantic-aware...: T. Zhan et al. |
|---|---|
| Authors: | Zhan, Tianming1,2 (AUTHOR) ztm@nau.edu.cn, Lu, Chenyang1 (AUTHOR), Wu, Huapeng1 (AUTHOR), Wang, Chenyun1 (AUTHOR) |
| Source: | Multimedia Systems. Jun2025, Vol. 31 Issue 3, p1-14. 14p. |
| Subjects: | Discrete wavelet transforms, Image intensifiers, Deep learning, Artificial intelligence, Image processing |
| Abstract: | The advent of deep learning has significantly propelled the advancement of low-light image enhancement techniques, yielding promising experimental outcomes. However, a series of image degradation problems such as noise and texture details have not been effectively handled, leaving room for further improvement of low-light image enhancement performance. In this work, we introduce a novel framework, the gradient and semantic-aware transformer network (GSTN), specifically tailored for low-light image enhancement. Our model comprises three pivotal components: the pre-lighten network (PLNet), which serves to light up the image to present more details and extract the illumination feature; the prior-guided enhancement module, designed to restore image details and mitigate noise leveraging the original gradient features; and the illuminance adjustment module (IAM), which refines the illumination of the enhanced image. In addition, we introduce discrete wavelet transform to implement cross-domain feature interactions and multi-scale feature fusion. Extensive experiments show that that our methods obtains better results in comparison with some state-of-the-art low-light image enhancement methods on different low-light datasets. [ABSTRACT FROM AUTHOR] |
| Copyright of Multimedia Systems is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | The advent of deep learning has significantly propelled the advancement of low-light image enhancement techniques, yielding promising experimental outcomes. However, a series of image degradation problems such as noise and texture details have not been effectively handled, leaving room for further improvement of low-light image enhancement performance. In this work, we introduce a novel framework, the gradient and semantic-aware transformer network (GSTN), specifically tailored for low-light image enhancement. Our model comprises three pivotal components: the pre-lighten network (PLNet), which serves to light up the image to present more details and extract the illumination feature; the prior-guided enhancement module, designed to restore image details and mitigate noise leveraging the original gradient features; and the illuminance adjustment module (IAM), which refines the illumination of the enhanced image. In addition, we introduce discrete wavelet transform to implement cross-domain feature interactions and multi-scale feature fusion. Extensive experiments show that that our methods obtains better results in comparison with some state-of-the-art low-light image enhancement methods on different low-light datasets. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 09424962 |
| DOI: | 10.1007/s00530-025-01710-y |
