Deep Contextual Structure and Semantic Feature Enhancement Stereo Network

Depth estimation is one of the fundamental tasks of computer vision. Stereo matching is the most critical step to obtain the accurate depth information through stereo vision. At present, thin structure regions, depth discontinuity regions, and large textureless regions are still the difficult issues...

Full description

Saved in:
Bibliographic Details
Main Authors: Guowei An, Yaonan Wang, Kai Zeng, Qing Zhu, Xiaofang Yuan, Yang Mo
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10556539/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850258298414039040
author Guowei An
Yaonan Wang
Kai Zeng
Qing Zhu
Xiaofang Yuan
Yang Mo
author_facet Guowei An
Yaonan Wang
Kai Zeng
Qing Zhu
Xiaofang Yuan
Yang Mo
author_sort Guowei An
collection DOAJ
description Depth estimation is one of the fundamental tasks of computer vision. Stereo matching is the most critical step to obtain the accurate depth information through stereo vision. At present, thin structure regions, depth discontinuity regions, and large textureless regions are still the difficult issues for stereo matching. To address the blur in thin structure regions and the dilation in depth discontinuity regions, the contextual structure enhancing module is proposed to enhance the extraction ability for local contextual features of the feature extraction network. To reduce the matching ambiguity in large textureless regions, the semantic feature enhancing module is proposed to enhance the aggregation ability for semantic features of the cost aggregation network. Extensive experiment results show that the proposed stereo network perform well in thin structure regions, depth discontinuity regions and large textureless regions and has achieved excellent performance on Scene Flow datasets, KITTI 2012 datasets, KITTI 2015 datasets and Middlebury datasets.
format Article
id doaj-art-3a6d81f314444ea68d76d2b41a28dcc2
institution OA Journals
issn 2169-3536
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-3a6d81f314444ea68d76d2b41a28dcc22025-08-20T01:56:13ZengIEEEIEEE Access2169-35362024-01-011218120518121610.1109/ACCESS.2024.341395710556539Deep Contextual Structure and Semantic Feature Enhancement Stereo NetworkGuowei An0https://orcid.org/0000-0001-6490-4277Yaonan Wang1https://orcid.org/0000-0002-0519-6458Kai Zeng2https://orcid.org/0000-0002-2745-1253Qing Zhu3https://orcid.org/0000-0001-7785-6374Xiaofang Yuan4https://orcid.org/0000-0001-7280-7207Yang Mo5https://orcid.org/0000-0001-6734-8691College of Electrical and Information Engineering, Hunan University, Changsha, ChinaCollege of Electrical and Information Engineering, Hunan University, Changsha, ChinaCollege of Electrical and Information Engineering, Hunan University, Changsha, ChinaCollege of Electrical and Information Engineering, Hunan University, Changsha, ChinaCollege of Electrical and Information Engineering, Hunan University, Changsha, ChinaCollege of Electrical and Information Engineering, Hunan University, Changsha, ChinaDepth estimation is one of the fundamental tasks of computer vision. Stereo matching is the most critical step to obtain the accurate depth information through stereo vision. At present, thin structure regions, depth discontinuity regions, and large textureless regions are still the difficult issues for stereo matching. To address the blur in thin structure regions and the dilation in depth discontinuity regions, the contextual structure enhancing module is proposed to enhance the extraction ability for local contextual features of the feature extraction network. To reduce the matching ambiguity in large textureless regions, the semantic feature enhancing module is proposed to enhance the aggregation ability for semantic features of the cost aggregation network. Extensive experiment results show that the proposed stereo network perform well in thin structure regions, depth discontinuity regions and large textureless regions and has achieved excellent performance on Scene Flow datasets, KITTI 2012 datasets, KITTI 2015 datasets and Middlebury datasets.https://ieeexplore.ieee.org/document/10556539/Stereo matchingdeep learningneural networkfeature extractionaggregation
spellingShingle Guowei An
Yaonan Wang
Kai Zeng
Qing Zhu
Xiaofang Yuan
Yang Mo
Deep Contextual Structure and Semantic Feature Enhancement Stereo Network
IEEE Access
Stereo matching
deep learning
neural network
feature extraction
aggregation
title Deep Contextual Structure and Semantic Feature Enhancement Stereo Network
title_full Deep Contextual Structure and Semantic Feature Enhancement Stereo Network
title_fullStr Deep Contextual Structure and Semantic Feature Enhancement Stereo Network
title_full_unstemmed Deep Contextual Structure and Semantic Feature Enhancement Stereo Network
title_short Deep Contextual Structure and Semantic Feature Enhancement Stereo Network
title_sort deep contextual structure and semantic feature enhancement stereo network
topic Stereo matching
deep learning
neural network
feature extraction
aggregation
url https://ieeexplore.ieee.org/document/10556539/
work_keys_str_mv AT guoweian deepcontextualstructureandsemanticfeatureenhancementstereonetwork
AT yaonanwang deepcontextualstructureandsemanticfeatureenhancementstereonetwork
AT kaizeng deepcontextualstructureandsemanticfeatureenhancementstereonetwork
AT qingzhu deepcontextualstructureandsemanticfeatureenhancementstereonetwork
AT xiaofangyuan deepcontextualstructureandsemanticfeatureenhancementstereonetwork
AT yangmo deepcontextualstructureandsemanticfeatureenhancementstereonetwork