模式识别与人工智能
2025年4月3日 星期四   首 页     期刊简介     编委会     投稿指南     伦理声明     联系我们                                                                English
模式识别与人工智能  2022, Vol. 35 Issue (12): 1101-1121    DOI: 10.16451/j.cnki.issn1003-6059.202212005
“基于深度学习的图像理解及应用”专题 最新目录| 下期目录| 过刊浏览| 高级检索 |
基于l1诱导轻量级深度网络的图像超分辨率重建
张大宝1, 赵建伟1,2, 周正华1
1.中国计量大学 理学院 杭州 310018;
2.中国计量大学 信息工程学院 浙江省电磁波信息技术与计量检测重点实验室 杭州 310018
Image Super-Resolution Reconstruction Based on l1 Induced Lightweight Deep Networks
ZHANG Dabao1, ZHAO Jianwei1,2, ZHOU Zhenghua1
1. College of Sciences, China Jiliang University, Hangzhou 310018;
2. Key Laboratory of Electromagnetic Wave Information Technology and Metrology of Zhejiang Province, College of Information Engineering, China Jiliang University, Hangzhou 310018

全文: PDF (2411 KB)   HTML (1 KB) 
输出: BibTeX | EndNote (RIS)      
摘要 现有的基于深度学习的超分辨率重建方法主要通过加深网络以提高网络的重建性能,但是加深网络会导致网络权值数量急剧增加,给网络训练和存储带来巨大负担.考虑到噪声的稀疏性、网络训练的成本及重建图像边缘的清晰度,文中融合模型驱动与数据驱动的思想,提出基于l1诱导轻量级深度网络的图像超分辨率重建方法.先利用分裂Bregman算法和软阈值算子,构建边缘正则的l1重建模型,并推演有效的迭代算法.再在迭代算法的指导下,设计相应的递归深度网络进行图像重建.因此,文中网络是在优化模型指导下设计的,推导出的递归结构由于其权值共享的特性,可减少网络权值的数量.实验表明,文中方法在网络权值数量较少时,仍能取得较优的图像重建性能.
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
张大宝
赵建伟
周正华
关键词 超分辨率重建深度学习轻量级网络模型驱动数据驱动    
Abstract:Existing deep-learning based super-resolution reconstruction methods improve the reconstruction performance of networks by deepening networks. However, sharp increase of the number of network weights is caused by deepening networks, resulting in a huge burden for the storage and training network. With the consideration of the sparsity of noise, the cost of training network and the sharpness of reconstructed edges, an image super-resolution reconstruction is proposed based on l1 induced lightweight deep networks integrating with the idea of model-driven and data-driven. Firstly, the split Bregman algorithm and soft threshold operator are utilized to deduce an effective iterative algorithm from the l1 reconstruction optimization model with an edge regularization term. Secondly, a corresponding recursive deep network is designed for image reconstruction under the guidance of the iterative algorithm. Therefore, the proposed deep network is designed under the guidance of the reconstruction optimization model, and its derived recursive structure reduces the number of network weights due to its property of weight sharing. Experimental results show that the proposed method achieves good reconstruction performance with less number of network weights.
Key wordsSuper-Resolution Reconstruction    Deep Learning    Lightweight Network    Model Driven    Data Driven   
收稿日期: 2022-10-08     
ZTFLH: TP391  
基金资助:浙江省自然科学基金项目(No.LY22F020002,LSY19F020001)资助
通讯作者: 赵建伟,博士,教授,主要研究方向为深度学习、图像处理等.E-mail:zhaojw@amss.ac.cn.   
作者简介: 张大宝,硕士研究生,主要研究方向为深度学习、图像处理等.E-mail:2569049998@qq.com.周正华,博士,副教授,主要研究方向为深度学习、图像处理等.E-mail:zzhzjw2003@163.com.
引用本文:   
张大宝, 赵建伟, 周正华. 基于l1诱导轻量级深度网络的图像超分辨率重建[J]. 模式识别与人工智能, 2022, 35(12): 1101-1121. ZHANG Dabao, ZHAO Jianwei, ZHOU Zhenghua. Image Super-Resolution Reconstruction Based on l1 Induced Lightweight Deep Networks. Pattern Recognition and Artificial Intelligence, 2022, 35(12): 1101-1121.
链接本文:  
http://manu46.magtech.com.cn/Jweb_prai/CN/10.16451/j.cnki.issn1003-6059.202212005      或     http://manu46.magtech.com.cn/Jweb_prai/CN/Y2022/V35/I12/1101
版权所有 © 《模式识别与人工智能》编辑部
地址:安微省合肥市蜀山湖路350号 电话:0551-65591176 传真:0551-65591176 Email:bjb@iim.ac.cn
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持:support@magtech.com.cn