Panoptic Segmentation with Convex Object Representation

doi:10.1093/comjnl/bxad119

	Panoptic Segmentation with Convex Object Representation
	Yao, Zhicheng 1,2; Wang, Sa 1,2; Zhu, Jinbin 1; Bao, Yungang 1,2
	2023-12-20
发表期刊	COMPUTER JOURNAL
ISSN	0010-4620
页码	11
摘要	The accurate representation of objects holds pivotal significance in the realm of panoptic segmentation. Presently, prevalent object representation methodologies, including box-based, keypoint-based and query-based techniques, encounter a challenge known as the 'representation confusion' issue in specific scenarios, often resulting in the mislabeling of instances. In response, this paper introduces Convex Object Representation (COR), a straightforward yet highly effective approach to address this problem. COR leverages a CNN-based Euclidean Distance Transform to convert the target instance into a convex heatmap. Simultaneously, it offers a parallel embedding method for encoding the object. Subsequently, COR characterizes objects based on the distinctive embedding vectors of their convex vertices. This paper seamlessly integrates COR into a state-of-the-art query-based panoptic segmentation framework. Experimental findings validate that COR successfully mitigates the representation confusion predicament, enhancing segmentation accuracy. The COR-augmented methods exhibit notable improvements of +1.3 and +0.7 points in PQ on the Cityscapes validation and MS COCO panoptic 2017 validation datasets, respectively.
关键词	deep learning computer vision image segmentation panoptic segmentation instance representation
DOI	10.1093/comjnl/bxad119
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[62090020] ; National Natural Science Foundation of China[61672499] ; Youth Innovation Promotion Association of Chinese Academy of Sciences[2013073] ; Strategic Priority Research Program of Chinese Academy of Sciences[XDC05030200]
WOS研究方向	Computer Science
WOS类目	Computer Science, Hardware & Architecture ; Computer Science, Information Systems ; Computer Science, Software Engineering ; Computer Science, Theory & Methods
WOS记录号	WOS:001128325000001
出版者	OXFORD UNIV PRESS
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/38462
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Wang, Sa
作者单位	1.Chinese Acad Sci, Inst Comp Technol, 6 Kexueyuan South Rd Zhongguancun, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, 1 Yanqihu East Rd, Beijing 101408, Peoples R China
推荐引用方式 GB/T 7714	Yao, Zhicheng,Wang, Sa,Zhu, Jinbin,et al. Panoptic Segmentation with Convex Object Representation[J]. COMPUTER JOURNAL,2023:11.
APA	Yao, Zhicheng,Wang, Sa,Zhu, Jinbin,&Bao, Yungang.(2023).Panoptic Segmentation with Convex Object Representation.COMPUTER JOURNAL,11.
MLA	Yao, Zhicheng,et al."Panoptic Segmentation with Convex Object Representation".COMPUTER JOURNAL (2023):11.