Leveraging local and global descriptors in parallel to search correspondences for visual localization

doi:10.1016/j.patcog.2021.108344

	Leveraging local and global descriptors in parallel to search correspondences for visual localization
	Zhang, Pengju 1,2; Zhang, Chaofan 3; Liu, Bingxi 1,2; Wu, Yihong 1,2
刊名	PATTERN RECOGNITION
	2022-02-01
卷号	122
关键词	Visual localization 6DoF pose Parallel search Learning based descriptor
ISSN号	0031-3203
DOI	10.1016/j.patcog.2021.108344
通讯作者	Zhang, Chaofan(zcfan@aiofm.ac.cn) ; Wu, Yihong(yhwu@nlpr.ia.ac.cn)
英文摘要	Visual localization to compute 6DoF camera pose from a given image has wide applications. Both local and global descriptors are crucial for visual localization. Most of the existing visual localization methods adopt a two-stage strategy: image retrieval first is performed by global descriptors, and then 2D-3D correspondences are made by local descriptors from 2D query image points and its nearest neighbor candidates which are the 3D points visible by these retrieved images. The above two stages are serially performed in these methods. However, due to the fact that 3D points obtained from the retrieval feedback are only rely on global descriptors, these methods cannot fully take the advantages of both local and global descriptors. In this paper, we propose a novel parallel search framework, which fully leverages advantages of both local and global descriptors to get nearest neighbor candidates of a 2D query image point. Specifically, besides using deep learning based global descriptors, we also utilize local descriptors to construct random tree structures for obtaining nearest neighbor candidates of the 2D query image point. We propose a new probability model and a new deep learning based local descriptor when constructing the random trees. In addition, a weighted Hamming regularization term to keep discriminativeness after binarization is given in loss function for the proposed local descriptor. The loss function co-trains both real and binary local descriptors of which the results are integrated into the random trees. Experiments on challenging benchmarks show that the proposed localization method can significantly improve the robustness and accuracy compared with the ones which get nearest neighbor candidates of a query local feature just based on either local or global descriptors. (c) 2021 Elsevier Ltd. All rights reserved.
资助项目	National Natural Science Foundation of China[61836015] ; National Natural Science Foundation of China[62002359] ; Natural Science Foundation of Anhui Province of China[2108085QF277] ; Beijing Advanced Discipline Fund[115200S001]
WOS关键词	FEATURES
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	ELSEVIER SCI LTD
WOS记录号	WOS:000704891800013
资助机构	National Natural Science Foundation of China ; Natural Science Foundation of Anhui Province of China ; Beijing Advanced Discipline Fund
内容类型	期刊论文
源URL	[http://ir.hfcas.ac.cn:8080/handle/334002/125547]
专题	中国科学院合肥物质科学研究院
通讯作者	Zhang, Chaofan; Wu, Yihong
作者单位	1.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 2.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China 3.Chinese Acad Sci, Hefei Inst Phys Sci, Anhui Inst Opt & Fine Mech, Beijing 230031, Anhui, Peoples R China
推荐引用方式 GB/T 7714	Zhang, Pengju,Zhang, Chaofan,Liu, Bingxi,et al. Leveraging local and global descriptors in parallel to search correspondences for visual localization[J]. PATTERN RECOGNITION,2022,122.
APA	Zhang, Pengju,Zhang, Chaofan,Liu, Bingxi,&Wu, Yihong.(2022).Leveraging local and global descriptors in parallel to search correspondences for visual localization.PATTERN RECOGNITION,122.
MLA	Zhang, Pengju,et al."Leveraging local and global descriptors in parallel to search correspondences for visual localization".PATTERN RECOGNITION 122(2022).