Efficient Progressive High Dynamic Range Image Restoration via Attention and Alignment Network
Abstract
HDR is an important part of computational photography technology. In this paper, we propose a lightweight neural network called Efficient Attention-and-alignment-guided Progressive Network (EAPNet) for the challenge NTIRE 2022 HDR Track 1 and Track 2. We introduce a multi-scale lightweight encoding module to extract features. Besides, we propose Progressive Dilated U-shape Block (PDUB) which is a progressive plug-and-play module for dynamically tuning MAccs and PSNR. Finally, we use fast and low-power feature-alignment module to deal with misalignment problem in place of the time-consuming Deformable Convolutional Network (DCN). The experiments show that our method achieves about 20× compression on MAccs with better PSNR-µ and PSNR compared to the state-of-the-art method. We got the 2nd place of both two tracks during the testing phase. Fig. 1 shows the visualized result of NTIRE 2022 HDR challenge.
Cite
Text
Yu et al. "Efficient Progressive High Dynamic Range Image Restoration via Attention and Alignment Network." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022. doi:10.1109/CVPRW56347.2022.00121Markdown
[Yu et al. "Efficient Progressive High Dynamic Range Image Restoration via Attention and Alignment Network." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2022.](https://mlanthology.org/cvprw/2022/yu2022cvprw-efficient/) doi:10.1109/CVPRW56347.2022.00121BibTeX
@inproceedings{yu2022cvprw-efficient,
title = {{Efficient Progressive High Dynamic Range Image Restoration via Attention and Alignment Network}},
author = {Yu, Gaocheng and Zhang, Jin and Ma, Zhe and Wang, Hongbin},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
year = {2022},
pages = {1123-1130},
doi = {10.1109/CVPRW56347.2022.00121},
url = {https://mlanthology.org/cvprw/2022/yu2022cvprw-efficient/}
}