Learning Semantic Correspondence with Sparse Annotations

doi:10.1007/978-3-031-19781-9_16

	Learning Semantic Correspondence with Sparse Annotations
	Huang, Shuaiyi 1; Yang, Luyu 1; He, Bo 1; Zhang, Songyang 2; He, Xuming3,4 ; Shrivastava, Abhinav 1
	2022
会议录名称	LECTURE NOTES IN COMPUTER SCIENCE (INCLUDING SUBSERIES LECTURE NOTES IN ARTIFICIAL INTELLIGENCE AND LECTURE NOTES IN BIOINFORMATICS)
ISSN	0302-9743
卷号	13674 LNCS
页码	267-284
发表状态	已发表
DOI	10.1007/978-3-031-19781-9_16
摘要	Finding dense semantic correspondence is a fundamental problem in computer vision, which remains challenging in complex scenes due to background clutter, extreme intra-class variation, and a severe lack of ground truth. In this paper, we aim to address the challenge of label sparsity in semantic correspondence by enriching supervision signals from sparse keypoint annotations. To this end, we first propose a teacher-student learning paradigm for generating dense pseudo-labels and then develop two novel strategies for denoising pseudo-labels. In particular, we use spatial priors around the sparse annotations to suppress the noisy pseudo-labels. In addition, we introduce a loss-driven dynamic label selection strategy for label denoising. We instantiate our paradigm with two variants of learning strategies: a single offline teacher setting, and mutual online teachers setting. Our approach achieves notable improvements on three challenging benchmarks for semantic correspondence and establishes the new state-of-the-art. Project page: https://shuaiyihuang.github.io/publications/SCorrSAN. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
关键词	Computer vision Learning systems Background clutter Complex scenes De-noising Ground truth Intra-class variation Learning semantics Pseudo-label Semantic correspondence Sparse annotation Teachers'
会议名称	17th European Conference on Computer Vision, ECCV 2022
出版地	GEWERBESTRASSE 11, CHAM, CH-6330, SWITZERLAND
会议地点	Tel Aviv, Israel
会议日期	October 23, 2022 - October 27, 2022
URL	查看原文
收录类别	EI ; CPCI-S
语种	英语
WOS研究方向	Computer Science ; Imaging Science & Photographic Technology
WOS类目	Computer Science, Artificial Intelligence ; Imaging Science & Photographic Technology
WOS记录号	WOS:000904096200016
出版者	Springer Science and Business Media Deutschland GmbH
EI入藏号	20224813184283
EI主题词	Semantics
EISSN	1611-3349
EI分类号	723.5 Computer Applications ; 741.2 Vision
原始文献类型	Conference article (CA)
引用统计
文献类型	会议论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/272836
专题	信息科学与技术学院_PI研究组_何旭明组
通讯作者	Huang, Shuaiyi
作者单位	1.Univ Maryland, College Pk, MD 20742 USA 2.Shanghai AI Lab, Shanghai, Peoples R China 3.ShanghaiTech Univ, Shanghai, Peoples R China 4.Shanghai Engn Res Ctr Intelligent Vision & Imagin, Shanghai, Peoples R China
推荐引用方式 GB/T 7714	Huang, Shuaiyi,Yang, Luyu,He, Bo,et al. Learning Semantic Correspondence with Sparse Annotations[C]. GEWERBESTRASSE 11, CHAM, CH-6330, SWITZERLAND:Springer Science and Business Media Deutschland GmbH,2022:267-284.