Deformable scene text detection using harmonic features and modified pixel aggregation network

doi:10.1016/j.patrec.2021.10.006

CORC > 自动化研究所 > 中国科学院自动化研究所 > 模式识别国家重点实验室 > 模式分析与学习团队

	Deformable scene text detection using harmonic features and modified pixel aggregation network
	Jain, Tanmay 2; Palaiahnakote, Shivakumara 1; Pal, Umapada 2; Liu, Cheng-Lin 3,4
刊名	PATTERN RECOGNITION LETTERS
	2021-12-01
卷号	152 页码:135-142
关键词	Natural scene text detection Maximally stable extremal region Fourier harmonic features Pixel aggregation network Deep learning models Deformable text detection
ISSN号	0167-8655
DOI	10.1016/j.patrec.2021.10.006
通讯作者	Palaiahnakote, Shivakumara(shiva@um.edu.my)
英文摘要	Although text detection methods have addressed several challenges in the past, there is a dearth of effective methods for text detection in deformable images, such as images containing text embedded on cloth, banners, rubber, sports jerseys, uniforms, etc. This is because deformable regions contain surfaces of arbitrarily shapes, which lead to poor text quality. This paper presents a new method for deformable text detection in natural scene images. It is observed that although the shapes of characters change in a deformable region, the pixel values and spatial relationship between the pixels do not change. This motivated us to explore extraction of Maximally Stable Extremal Regions (MSER) in an image in which pixels that share common features are grouped into components. The unique character shape variations led us to explore harmonic features to represent the component shape variations, using which a classifier classifies text and non-text components from the output of the MSER step. Additionally, the objective of developing a lightweight method with low computational cost motivated us to introduce a modified Pixel Aggression Network (PAN) for text deformable text detection at a component level. Comprehensive experiments which include experiments on our Deformable Text Dataset (DTD) and standard natural scene text datasets, namely, MSRATD-500, ICDAR 2019 MLT, Total-Text, CTW1500, ICDAR 2019 ArT and DSTA1500 datasets show that the proposed model outperforms the existing methods for our dataset as well as the standard datasets. (c) 2021 Elsevier B.V. All rights reserved.
资助项目	FRGS grant, Ministry of Higher Education, Malaysia[FP1042020] ; TIH, ISI, Kolkata
WOS研究方向	Computer Science
语种	英语
出版者	ELSEVIER
WOS记录号	WOS:000711455600007
资助机构	FRGS grant, Ministry of Higher Education, Malaysia ; TIH, ISI, Kolkata
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/46273]
专题	自动化研究所_模式识别国家重点实验室_模式分析与学习团队
通讯作者	Palaiahnakote, Shivakumara
作者单位	1.Univ Malaya, Fac Comp Sci & Informat Technol, Dept Comp Syst & Technol, Kuala Lumpur, Malaysia 2.Indian Stat Inst, Comp Vis & Pattern Recognit Unit, Kolkata, India 3.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 4.Chinese Acad Sci, Inst Automat, Cheng Lin Liu Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Jain, Tanmay,Palaiahnakote, Shivakumara,Pal, Umapada,et al. Deformable scene text detection using harmonic features and modified pixel aggregation network[J]. PATTERN RECOGNITION LETTERS,2021,152:135-142.
APA	Jain, Tanmay,Palaiahnakote, Shivakumara,Pal, Umapada,&Liu, Cheng-Lin.(2021).Deformable scene text detection using harmonic features and modified pixel aggregation network.PATTERN RECOGNITION LETTERS,152,135-142.
MLA	Jain, Tanmay,et al."Deformable scene text detection using harmonic features and modified pixel aggregation network".PATTERN RECOGNITION LETTERS 152(2021):135-142.