Multi-Dimensional Residual Dense Attention Network for Stereo Matching

doi:10.1109/ACCESS.2019.2911618

	Multi-Dimensional Residual Dense Attention Network for Stereo Matching
	Zhang, Guanghui 1,2; Zhu, Dongchen 1; Shi, Wenjun 1,2; Ye, Xiaoqing 1,2; Li, Jiamao 1,2; Zhang, Xiaolin1,2,3
	2019
发表期刊	IEEE ACCESS
ISSN	2169-3536
卷号	7 期号:99 页码:51681-51690
发表状态	已发表
DOI	10.1109/ACCESS.2019.2911618
摘要	Very deep convolutional neural networks (CNNs) have recently achieved great success in stereo matching. It is still highly desirable to learn a robust feature map to improve ill-posed regions, such as weakly textured regions, reflective surfaces, and repetitive patterns. Therefore, we propose an endto-end multi-dimensional residual dense attention network (MRDA-Net) in this paper, focusing on more comprehensive pixel-wise feature extraction. Our proposed network consists of two parts: the 2D residual dense attention net for feature extraction and the 3D convolutional attention net for matching. Our designed 2D residual dense attention net uses a dense network structure to fully exploit the hierarchical features from preceding convolutional layers and uses residual network structure to fuse low-level structure information and high-level semantic information. The 2D attention module of the net aims to adaptively recalibrate channel-wise features to be more concerned about informative features. Our proposed 3D convolutional attention net further expands attention mechanism for matching. The stacked hourglass module of the net is exploited to extract multi-scale context information as well as geometry information. The novel 3D attention module of the net aggregates hierarchical sub-cost volumes adaptively instead of manually and then achieves a comprehensive recalibrated cost volume for more correct disparity computation. The experiments demonstrate that our approach achieves the state-of-the-art accuracy on Scene Flow dataset and KITTI 2012 and KITTI 2015 Stereo datasets.
关键词	Stereo matching multi-dimensional residual dense attention hierarchical 3D attention mechanism
URL	查看原文
收录类别	SCI ; SCIE ; EI
语种	英语
资助项目	Shanghai Science and Technology Committee, China[17511108202]
WOS研究方向	Computer Science ; Engineering ; Telecommunications
WOS类目	Computer Science, Information Systems ; Engineering, Electrical & Electronic ; Telecommunications
WOS记录号	WOS:000466706300001
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
EI入藏号	20192407025701
EI主题词	Convolution ; Deep neural networks ; Extraction ; Feature extraction ; Neural networks ; Semantics ; Textures
EI分类号	Information Theory and Signal Processing:716.1 ; Data Processing and Image Processing:723.2 ; Chemical Operations:802.3
原始文献类型	Article
引用统计
文献类型	期刊论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/31163
专题	信息科学与技术学院_特聘教授组_张晓林组
通讯作者	Li, Jiamao
作者单位	1.Chinese Acad Sci, Shanghai Inst Microsyst & Informat Technol, Bion Vis Syst Lab, State Key Lab Transducer Technol, Shanghai 200050, Peoples R China 2.Univ Chinese Acad Sci, Sch Elect Elect & Commun Engn, Beijing 100049, Peoples R China 3.ShanghaiTech Univ, Sch Informat & Technol, Shanghai 200050, Peoples R China
推荐引用方式 GB/T 7714	Zhang, Guanghui,Zhu, Dongchen,Shi, Wenjun,et al. Multi-Dimensional Residual Dense Attention Network for Stereo Matching[J]. IEEE ACCESS,2019,7(99):51681-51690.
APA	Zhang, Guanghui,Zhu, Dongchen,Shi, Wenjun,Ye, Xiaoqing,Li, Jiamao,&Zhang, Xiaolin.(2019).Multi-Dimensional Residual Dense Attention Network for Stereo Matching.IEEE ACCESS,7(99),51681-51690.
MLA	Zhang, Guanghui,et al."Multi-Dimensional Residual Dense Attention Network for Stereo Matching".IEEE ACCESS 7.99(2019):51681-51690.