基于SPL迭代思想的图像压缩感知重构神经网络

doi:10.12263/DZXB.20200618

电子学报 ›› 2021, Vol. 49 ›› Issue (6): 1195-1203.DOI: 10.12263/DZXB.20200618

所属专题：压缩感知

基于SPL迭代思想的图像压缩感知重构神经网络

裴翰奇¹, 杨春玲¹, 魏志超¹, 曹燕²

1. 华南理工大学电子与信息学院, 广东广州 510640;
2. 华南理工大学国家移动超声探测工程技术研究中心, 广东广州 510640

收稿日期:2020-06-28 修回日期:2020-10-10 出版日期:2021-06-25 发布日期:2022-06-25
通讯作者: 杨春玲(通信作者) 女,1970年生于河南新乡,华南理工大学电子与信息学院博士生导师.研究方向:图像/视频压缩编码,图像质量评价.E-mail:eeclyang@scut.edu.cn
作者简介:裴翰奇 男,1997年生于江西抚州,华南理工大学电子与信息学院研究生.研究方向:图像压缩感知.E-mail:AsukaHellkite@outlook.com;魏志超 男,1996年生于河南禹州,华南理工大学电子与信息学院研究生.研究方向:视频压缩感知.E-mail:zcw781432@outlook.com;曹燕女,1980年生于重庆江津,华南理工大学国家移动超声探测工程技术研究中心副教授.研究方向:声信号处理.E-mail:eeyancao@scut.edu.cn
基金资助:
广东省自然科学基金重点项目（No.2017A030311028）；广东省"短距离无线探测与通信"重点实验室（No.2014B030301010，No.2017B030314003）；国家移动超声探测工程技术研究中心（No.2013FU125X02）

Image Compressive Sensing Reconstruction Network Based on Iterative SPL Theory

PEI Han-qi¹, YANG Chun-ling¹, WEI Zhi-chao¹, CAO Yan²

1. School of Electronic and Information Engineering, South China University of Technology, Guangzhou, Guangdong 510640, China;
2. The National Engineering Technology Research Center for Mobile Ultrasonic Detection, South China University of Technology, Guangzhou, Guangdong 510640, China

Received:2020-06-28 Revised:2020-10-10 Online:2021-06-25 Published:2022-06-25

摘要/Abstract

摘要： 由于神经网络强大的学习能力与快速的运行速度，近年来基于深度学习的图像压缩感知（Image Compressive Sensing，ICS）研究备受关注.然而，大多数现有ICS神经网络的结构设计忽略了传统迭代重构算法中的数学理论基础，无法有效利用信号中的先验结构知识，可解释性较差.为了保留优化算法核心思想并同时利用深度学习的高性能，本文使用可学习的卷积层替代了传统平滑投影Landweber算法（Smooth Projected Landweber，SPL）中的人工设计参数，提出一种新型ICS神经网络SPLNet.在SPLNet中，设计了一个独特的网络结构SPLBlock实现SPL迭代过程中的三个核心步骤：（1）去除块效应的维纳滤波器；（2）在凸投影集合上的近似操作；（3）实现稀疏表示及去噪的变换域双变量收缩.仿真实验结果表明：与现有最优的ICS优化迭代算法GSR相比， SPLNet的重构图像平均PSNR提升了0.78dB；与最优的神经网络框架SCSNet相比，SPLNet的重构图像平均PSNR提升了0.92dB.

关键词: 压缩感知, 图像压缩感知, 深度网络, 卷积神经网络, 图像采样, 图像重构

Abstract: Due to its great learning ability and fast processing speed,deep learning-based image compressive sensing (ICS) methods attract a lot of attention in recent years.However,the design of most existing ICS neural networks architecture ignore the mathematical theory in iterative optimization-based methods and cannot effectively use the prior structure knowledge in the signal,leading to lack of the interpretability.In order to retain the core ideas of the optimization algorithm and utilize the high performance of deep learning,this paper uses learnable convolutional layers to replace the predefined filters and artificial design parameters in the traditional smooth projected Landweber algorithm (SPL),and proposes a ICS neural network named SPLNet.In SPLNet,we design a unique network structure SPLBlock to implement three key steps in SPL iteration:(1) Wiener filter for removal of blocking artifacts; (2) approximation with projection onto the convex set; (3) bivariate shrinkage on transform domain for sparse representation and denoising.Experimental results indicate that,compared with current state-of-the-art ICS optimization iterative algorithm GSR,the average reconstructed image PSNR of SPLNet are improved by 0.78dB,and compared with state-of-the-art neural network framework SCSNet,the average reconstructed image PSNR of SPLNet are improved by 0.92dB.

Key words: compressive sensing, image compressive sensing, deep networks, convolutional neural networks, image sampling, image reconstruction

中图分类号:

TN919.8

裴翰奇, 杨春玲, 魏志超, 曹燕. 基于SPL迭代思想的图像压缩感知重构神经网络[J]. 电子学报, 2021, 49(6): 1195-1203.

PEI Han-qi, YANG Chun-ling, WEI Zhi-chao, CAO Yan. Image Compressive Sensing Reconstruction Network Based on Iterative SPL Theory[J]. Acta Electronica Sinica, 2021, 49(6): 1195-1203.

参考文献

[1] Donoho D L.Compressed sensing[J].IEEE Transactions on Information Theory,2006,52(4):1289-1306.
[2] Gan L.Block compressed sensing of natural images[A].200715th International Conference on Digital Signal Processing[C].Cardiff:IEEE,2007.403-406.
[3] Mun S,Fowler J E.Block compressed sensing of images using directional transforms[A].200916th IEEE International Conference on Image Processing (ICIP)[C].Cairo:IEEE,2009.3021-3024.
[4] Fowler J E,Mun S,Tramel E W.Multiscale block compressed sensing with smoothed projected landweber reconstruction[A].201119th European Signal Processing Conference[C].Barcelona:IEEE,2011.564-568.
[5] Mun S,Fowler J E.Residual reconstruction for block-based compressed sensing of video[A].2011 Data Compression Conference[C].Snowbird:IEEE,2011.183-192.
[6] Chen C,Tramel E W,Fowler J E.Compressed-sensing recovery of images and video using multihypothesis predictions[A].2011 Conference Record of the Forty Fifth Asilomar Conference on Signals,Systems and Computers (ASILOMAR)[C].USA:IEEE,2011.1193-1198.
[7] Zhang J,Zhao D,Gao W.Group-based sparse representation for image restoration[J].IEEE Transactions on Image Processing,2014,23(8):3336-3351.
[8] Fan Q,Zhuo W,Tang C K,et al.Few-shot object detection with attention-RPN and multi-relation detector[A].Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition[C].USA:IEEE,2020.4013-4022.
[9] Chen Z,Zhong B,Li G,et al.Siamesebox adaptive network for visual tracking[A].Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition[C].USA:IEEE,2020.6668-6677.
[10] Xiang X,Tian Y,Zhang Y,et al.Zooming Slow-Mo:Fast and accurate one-stage space-time video super-resolution[A].Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition[C].USA:IEEE,2020.3370-3379.
[11] Kulkarni K,Lohit S,Turaga P,et al.ReconNet:Non-iterative reconstruction of images from compressively sensed measurements[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].Las Vegas:IEEE,2016.449-458.
[12] Yao H,Dai F,Zhang S,et al.DR2-Net:Deep residual reconstruction network for image compressive sensing[J].Neurocomputing,2019,359:483-493.
[13] Zhang J,Ghanem B.ISTA-Net:Interpretable optimization-inspired deep network for image compressive sensing[A].Proceedings of the IEEEConference on Computer Vision and Pattern Recognition[C].Salt Lake City:IEEE,2018.1828-1837.
[14] 练秋生,富利鹏,陈书贞,等.基于多尺度残差网络的压缩感知重构算法[J].自动化学报,2019,45(11):2082-2091. LIAN Qiu-sheng,FU Li-peng,CHEN Shu-zhen,et al.A compressed sensing algorithm based on multi-scale residual reconstruction network[J].Acta Automatica Sinica,2019,45(11):2082-2091.(in Chinese)
[15] Shi W,Jiang F,Zhang S,et al.Deep networks for compressed image sensing[A].2017 IEEE International Conference on Multimedia and Expo (ICME)[C].Hong Kong:IEEE,2017.877-882.
[16] Shi W,Jiang F,Liu S,et al.Multi-scale deep networks for image compressed sensing[A].201825th IEEE International Conference on Image Processing (ICIP)[C].Athens:IEEE,2018.46-50.
[17] Shi W,Jiang F,Liu S,et al.Image Compressed sensing using convolutional neural network[J].IEEE Transactions on Image Processing,2019,29:375-388.
[18] Shi W,Jiang F,Liu S,et al.Scalable convolutional neural network for image compressed sensing[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].Long Beach:IEEE,2019.12290-12299.
[19] Cui W,Jiang F,Gao X,et al.Deep neural network based sparse measurement matrix for image compressed sensing[A].201825th IEEE International Conference on Image Processing (ICIP)[C].Athens:IEEE,2018.3883-3887.
[20] He K,Zhang X,Ren S,et al.Delving deep into rectifiers:Surpassing human-level performance onimagenet classification[A].Proceedings of the IEEE International Conference on Computer Vision[C].Seoul:IEEE,2015.1026-1034.
[21] Arbelaez P,Maire M,Fowlkes C,et al.Contour detection and hierarchical image segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2010,33(5):898-916.
[22] Kingma D P,Ba J.Adam:A Method for Stochastic Optimization[EB/OL].https://arxiv.org/abs/1412.6980,2014.
[23] Abadi M,Barham P,Chen J,et al.Tensorflow:A system for large-scale machine learning[A].12th {USENIX} Symposium on Operating Systems Design and Implementation ({OSDI} 16)[C].Savannah:USENIX,2016.265-283.
[24] Dong H,Supratak A,Mai L,et al.Tensorlayer:A versatile library for efficient deep learning development[A].Proceedings of the 25th ACM International Conference on Multimedia[C].USA:ACM,2017.1201-1204.

基于SPL迭代思想的图像压缩感知重构神经网络

Image Compressive Sensing Reconstruction Network Based on Iterative SPL Theory

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	廖勇, 李玉杰. 一种轻量化低复杂度的FDD大规模MIMO系统CSI反馈方法[J]. 电子学报, 2022, 50(5): 1211-1217.
[2]	吕明久, 马建朝, 韦旭, 陈文峰, 杨军, 马晓岩. 基于矩阵化原子范数的高频雷达距离-多普勒二维估计方法[J]. 电子学报, 2022, 50(5): 1150-1158.
[3]	张云, 化青龙, 姜义成, 徐丹. 基于混合型复数域卷积神经网络的三维转动舰船目标识别[J]. 电子学报, 2022, 50(5): 1042-1049.
[4]	苏林林, 陈亮, 陈菲菲, 周鑫, 焦振航, 刘钊良. 面向B5G/6G的GFDM信号高精度测距与定位研究[J]. 电子学报, 2022, 50(4): 849-859.
[5]	崔亚奇, 何友, 唐田田, 熊伟. 一种深度学习航迹关联方法[J]. 电子学报, 2022, 50(3): 759-763.
[6]	伍邦谷, 张苏林, 石红, 朱鹏飞, 王旗龙, 胡清华. 基于多分支结构的不确定性局部通道注意力机制[J]. 电子学报, 2022, 50(2): 374-382.
[7]	李居朋, 王颖慧, 李刚. 医学图像关键点检测深度学习方法研究与挑战[J]. 电子学报, 2022, 50(1): 226-237.
[8]	左婷, 王法松, 张建康, 李睿. 室内可见光通信系统中基于压缩感知的空移键控信号检测方法[J]. 电子学报, 2022, 50(1): 36-44.
[9]	赵琰, 赵凌君, 匡纲要. 基于注意力机制特征融合网络的SAR图像飞机目标快速检测[J]. 电子学报, 2021, 49(9): 1665-1674.
[10]	冯西安, 寇思玮, 谭伟杰, 毕杨. 水声信号处理中的稀疏表示理论及应用[J]. 电子学报, 2021, 49(9): 1840-1851.
[11]	管永明, 王刚, 骆凯波, 吕梁, 吕晓雯, 史玉良. 基于极端不平衡学习的泛化低压异常箱表关系识别研究与应用[J]. 电子学报, 2021, 49(8): 1507-1514.
[12]	徐频捷, 陈逸杰, 李之南, 赵地. 基于事件驱动的车道线识别算法研究[J]. 电子学报, 2021, 49(7): 1379-1385.
[13]	罗会兰, 袁璞, 童康. 基于深度学习的显著性目标检测方法综述[J]. 电子学报, 2021, 49(7): 1417-1427.
[14]	程旭, 宋晨, 史金钢, 周琳, 张毅锋, 郑钰辉. 基于深度学习的通用目标检测研究综述[J]. 电子学报, 2021, 49(7): 1428-1438.
[15]	张昱, 刘开峰, 张全新, 王艳歌, 高凯龙. 基于组合-卷积神经网络的中文新闻文本分类[J]. 电子学报, 2021, 49(6): 1059-1067.