Unconstrained end-to-end text reading with feature rectification

doi:10.1016/j.patrec.2021.05.017

CORC > 自动化研究所 > 中国科学院自动化研究所 > 复杂系统管理与控制国家重点实验室 > 影像分析与机器视觉团队

	Unconstrained end-to-end text reading with feature rectification
	Du, Chen 1,2; Wang, Yanna 1; Wang, Chunheng 1; Xiao, Baihua 1; Shi, Cunzhao 1
刊名	PATTERN RECOGNITION LETTERS
	2021-09-01
卷号	149 页码:1-8
关键词	Text recognition Text detection Position-sensitive network Features incompatibility End-to-end
ISSN号	0167-8655
DOI	10.1016/j.patrec.2021.05.017
通讯作者	Wang, Chunheng(chunheng.wang@ia.ac.cn)
英文摘要	We propose an end-to-end trainable network that can simultaneously localize and recognize irregular text from images. Specifically, we find the feature incompatibility problem, which arises from the contradiction between detection and recognition tasks for feature extraction of the convolutional neural network, and propose to introduce the larger-scale features for the recognition part to improve the accuracy of recognition instead of using the same feature with the detection. To extract effective text features for perspective and curved text recognition, we propose a position-sensitive network to rectify the text proposal features in the recognition branch. The position-sensitive network, which is trained in a weak supervision way, takes the proposal detection feature as input and outputs the feature rectification information. Experiments demonstrate that the proposed method can achieve state-of-the-art or highly competitive performance compared with baselines on a number of benchmarks. (c) 2021 Elsevier B.V. All rights reserved.
资助项目	Key Programs of the Chinese Academy of Sciences[ZDBSSSWJSC004] ; Key Programs of the Chinese Academy of Sciences[ZDBSSSWJSC005] ; National Natural Science Foundation of China (NSFC)[62073326] ; National Natural Science Foundation of China (NSFC)[62071469]
WOS关键词	NETWORK
WOS研究方向	Computer Science
语种	英语
出版者	ELSEVIER
WOS记录号	WOS:000680052800001
资助机构	Key Programs of the Chinese Academy of Sciences ; National Natural Science Foundation of China (NSFC)
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/45649]
专题	自动化研究所_复杂系统管理与控制国家重点实验室_影像分析与机器视觉团队
通讯作者	Wang, Chunheng
作者单位	1.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing, Peoples R China
推荐引用方式 GB/T 7714	Du, Chen,Wang, Yanna,Wang, Chunheng,et al. Unconstrained end-to-end text reading with feature rectification[J]. PATTERN RECOGNITION LETTERS,2021,149:1-8.
APA	Du, Chen,Wang, Yanna,Wang, Chunheng,Xiao, Baihua,&Shi, Cunzhao.(2021).Unconstrained end-to-end text reading with feature rectification.PATTERN RECOGNITION LETTERS,149,1-8.
MLA	Du, Chen,et al."Unconstrained end-to-end text reading with feature rectification".PATTERN RECOGNITION LETTERS 149(2021):1-8.