基于双特征聚合网络的轻量图像超分辨率

Shang Li, Guixuan Zhang, Zhengxiong Luo, Jie Liu, Zhi Zeng, Shuwu Zhang
{"title":"基于双特征聚合网络的轻量图像超分辨率","authors":"Shang Li, Guixuan Zhang, Zhengxiong Luo, Jie Liu, Zhi Zeng, Shuwu Zhang","doi":"10.1109/ICCST53801.2021.00104","DOIUrl":null,"url":null,"abstract":"With the power of deep learning, super-resolution (SR) methods enjoy a dramatic boost of performance. However, they usually have a large model size and high computational complexity, which hinders the application in devices with limited memory and computing power. Some lightweight SR methods solve this issue by directly designing shallower architectures, but it will affect SR performance. In this paper, we propose the dual feature aggregation strategy (DFA). It enhances the feature utilization via feature reuse, which largely improves the representation ability while only introducing marginal computational cost. Thus, a smaller model could achieve better cost-effectiveness with DFA. Specifically, DFA consists of local and global feature aggregation modules (LAM and GAM). They work together to further fuse hierarchical features adaptively along the channel and spatial dimensions. Extensive experiments suggest that the proposed network performs favorably against the state-of-the-art SR methods in terms of visual quality, memory footprint, and computational complexity.","PeriodicalId":222463,"journal":{"name":"2021 International Conference on Culture-oriented Science & Technology (ICCST)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Lightweight Image Super-Resolution via Dual Feature Aggregation Network\",\"authors\":\"Shang Li, Guixuan Zhang, Zhengxiong Luo, Jie Liu, Zhi Zeng, Shuwu Zhang\",\"doi\":\"10.1109/ICCST53801.2021.00104\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"With the power of deep learning, super-resolution (SR) methods enjoy a dramatic boost of performance. However, they usually have a large model size and high computational complexity, which hinders the application in devices with limited memory and computing power. Some lightweight SR methods solve this issue by directly designing shallower architectures, but it will affect SR performance. In this paper, we propose the dual feature aggregation strategy (DFA). It enhances the feature utilization via feature reuse, which largely improves the representation ability while only introducing marginal computational cost. Thus, a smaller model could achieve better cost-effectiveness with DFA. Specifically, DFA consists of local and global feature aggregation modules (LAM and GAM). They work together to further fuse hierarchical features adaptively along the channel and spatial dimensions. Extensive experiments suggest that the proposed network performs favorably against the state-of-the-art SR methods in terms of visual quality, memory footprint, and computational complexity.\",\"PeriodicalId\":222463,\"journal\":{\"name\":\"2021 International Conference on Culture-oriented Science & Technology (ICCST)\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-11-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2021 International Conference on Culture-oriented Science & Technology (ICCST)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICCST53801.2021.00104\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 International Conference on Culture-oriented Science & Technology (ICCST)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICCST53801.2021.00104","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

借助深度学习的力量,超分辨率(SR)方法的性能得到了显著提升。然而,它们通常具有较大的模型尺寸和较高的计算复杂度,这阻碍了在内存和计算能力有限的设备中的应用。一些轻量级的SR方法通过直接设计较浅的架构来解决这个问题,但是这会影响SR的性能。本文提出了双特征聚合策略(DFA)。它通过特征重用来提高特征利用率,在只引入边际计算成本的情况下,极大地提高了表征能力。因此,使用DFA,较小的模型可以获得更好的成本效益。具体来说,DFA由局部和全局特征聚合模块(LAM和GAM)组成。它们一起工作,进一步自适应地融合沿着通道和空间维度的分层特征。大量的实验表明,所提出的网络在视觉质量、内存占用和计算复杂性方面优于最先进的SR方法。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Lightweight Image Super-Resolution via Dual Feature Aggregation Network
With the power of deep learning, super-resolution (SR) methods enjoy a dramatic boost of performance. However, they usually have a large model size and high computational complexity, which hinders the application in devices with limited memory and computing power. Some lightweight SR methods solve this issue by directly designing shallower architectures, but it will affect SR performance. In this paper, we propose the dual feature aggregation strategy (DFA). It enhances the feature utilization via feature reuse, which largely improves the representation ability while only introducing marginal computational cost. Thus, a smaller model could achieve better cost-effectiveness with DFA. Specifically, DFA consists of local and global feature aggregation modules (LAM and GAM). They work together to further fuse hierarchical features adaptively along the channel and spatial dimensions. Extensive experiments suggest that the proposed network performs favorably against the state-of-the-art SR methods in terms of visual quality, memory footprint, and computational complexity.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信