Two-stream deep sparse network for accurate and efficient image restoration

doi:10.1016/j.cviu.2020.103029

	Two-stream deep sparse network for accurate and efficient image restoration
	Wang, Shuhui 1; Hu, Ling 1,2; Li, Liang 1; Zhang, Weigang 3; Huang, Qingming 1,2,4
	2020-11-01
发表期刊	COMPUTER VISION AND IMAGE UNDERSTANDING
ISSN	1077-3142
卷号	200 页码:11
摘要	Deep convolutional neural network (CNN) has achieved great success in image restoration. However, previous methods ignore the complementarity between low-level and high-level features, thereby leading to limited image reconstruction quality. In this paper, we propose a two-stream sparse network (TSSN) to explicitly learn shallow and deep features to enforce their respective contribution to image restoration. The shallow stream learns shallow features (e.g., texture edges), and the deep stream learns deep features (e.g., salient semantics). In each stream, sparse residual block (SRB) is proposed to efficiently aggregate hierarchical features by constructing sparse connections among layers in the local block. Spatial-wise and channel-wise attention are used to fuse the shallow and deep stream which recalibrates features by weight assignment in both spatial and channel dimensions. A novel loss function called Softmax-L-1 loss is proposed to increase penalties of pixels that have large L-1 loss (i.e., hard pixels). TSSN is evaluated with three representative IR applications, i.e., single image super-resolution, image denoising and JPEG deblocking. Extensive experiments demonstrate that TSSN outperforms most of state-of-the-art methods on benchmark datasets on both quantitative metric and visual quality.
关键词	Two-stream sparse network Image restoration Image super-resolution Image denoising
DOI	10.1016/j.cviu.2020.103029
收录类别	SCI
语种	英语
资助项目	National Key R&D Program of China[2018AAA0102003] ; National Natural Science Foundation of China[61672497] ; National Natural Science Foundation of China[61771457] ; National Natural Science Foundation of China[61836002] ; National Natural Science Foundation of China[61620106009] ; National Natural Science Foundation of China[U1636214] ; National Natural Science Foundation of China[61931008] ; Key Research Program of Frontier Sciences of CAS[QYZDJ-SSW-SYS013]
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic
WOS记录号	WOS:000579188100001
出版者	ACADEMIC PRESS INC ELSEVIER SCIENCE
引用统计	被引频次：2[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/15709
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Li, Liang
作者单位	1.Chinese Acad Sci, Key Lab Intelligent Informat Proc, Inst Comp Technol, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Sch Comp Sci & Technol, Beijing 100049, Peoples R China 3.Harbin Inst Technol, Weihai 264200, Peoples R China 4.Univ Chinese Acad Sci, Key Lab Big Data Min & Knowledge Management, Beijing 101408, Peoples R China
推荐引用方式 GB/T 7714	Wang, Shuhui,Hu, Ling,Li, Liang,et al. Two-stream deep sparse network for accurate and efficient image restoration[J]. COMPUTER VISION AND IMAGE UNDERSTANDING,2020,200:11.
APA	Wang, Shuhui,Hu, Ling,Li, Liang,Zhang, Weigang,&Huang, Qingming.(2020).Two-stream deep sparse network for accurate and efficient image restoration.COMPUTER VISION AND IMAGE UNDERSTANDING,200,11.
MLA	Wang, Shuhui,et al."Two-stream deep sparse network for accurate and efficient image restoration".COMPUTER VISION AND IMAGE UNDERSTANDING 200(2020):11.