MHDT: A Deep-Learning-Based Text Detection Algorithm for Unstructured Data in Banking

Shenglan Ma; Lingling Yang; Hao Wang; Hong Xiao; Hong Ning Dai; Shuhan Cheng; Tongsen Wang

doi:10.1145/3318299.3318327

MHDT: A Deep-Learning-Based Text Detection Algorithm for Unstructured Data in Banking

Shenglan Ma, Lingling Yang, Hao Wang, Hong Xiao, Hong Ning Dai, Shuhan Cheng, Tongsen Wang

Department of Computer Science

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

Abstract

Text detection in natural scene images becomes highly demanded for unstructured data in banking. In this paper, we propose a new deep learning algorithm called MSER, Hu-moment and Deep learning for Text detection (MHDT) based on Maximum Stable Extremal Regions (MSER) and Hu-moment features. Firstly, we extract MSERs as candidate characters. Secondly, a character classifier is introduced with Hu-moment features to reduce the number of input for clustering. After single linkage clustering, a text classifier trained from a Deep Brief Network is used to delete non-text. The proposed algorithm is evaluated on the ICDAR database, and the experimental results show that the proposed algorithm yields high precision and recall rate.

Original language	English
Title of host publication	ICMLC '19: Proceedings of the 2019 11th International Conference on Machine Learning and Computing
Publisher	Association for Computing Machinery (ACM)
Pages	295-300
Number of pages	6
ISBN (Print)	9781450366007
DOIs	https://doi.org/10.1145/3318299.3318327
Publication status	Published - 22 Feb 2019
Event	11th International Conference on Machine Learning and Computing, ICMLC 2019 - Zhuhai, China Duration: 22 Feb 2019 → 24 Feb 2019 https://dl.acm.org/doi/proceedings/10.1145/3318299

Conference

Conference	11th International Conference on Machine Learning and Computing, ICMLC 2019
Country/Territory	China
City	Zhuhai
Period	22/02/19 → 24/02/19
Internet address	https://dl.acm.org/doi/proceedings/10.1145/3318299

Scopus Subject Areas

Software
Human-Computer Interaction
Computer Vision and Pattern Recognition
Computer Networks and Communications

User-Defined Keywords

Deep learning
Text detection
Unstructured data

Access to Document

10.1145/3318299.3318327

Cite this

@inproceedings{0214f9b3154a4405a971b38e1c104bca,

title = "MHDT: A Deep-Learning-Based Text Detection Algorithm for Unstructured Data in Banking",

abstract = "Text detection in natural scene images becomes highly demanded for unstructured data in banking. In this paper, we propose a new deep learning algorithm called MSER, Hu-moment and Deep learning for Text detection (MHDT) based on Maximum Stable Extremal Regions (MSER) and Hu-moment features. Firstly, we extract MSERs as candidate characters. Secondly, a character classifier is introduced with Hu-moment features to reduce the number of input for clustering. After single linkage clustering, a text classifier trained from a Deep Brief Network is used to delete non-text. The proposed algorithm is evaluated on the ICDAR database, and the experimental results show that the proposed algorithm yields high precision and recall rate.",

keywords = "Deep learning, Text detection, Unstructured data",

author = "Shenglan Ma and Lingling Yang and Hao Wang and Hong Xiao and Dai, {Hong Ning} and Shuhan Cheng and Tongsen Wang",

note = "Funding Information: This work is partially funded by the Fujian Fumin Foundation and is supported by the National Natural Science Foundation of China under Grant (No. 61672170 and No. 61871313), the Science and Technology Planning Project of Guangdong Province (No. 2017A050501035), and Science and Technology Program of Guangzhou (No. 201807010058). Publisher Copyright: {\textcopyright} 2019 Copyright is held by the owner/author(s). Publication rights licensed to ACM.; 11th International Conference on Machine Learning and Computing, ICMLC 2019 ; Conference date: 22-02-2019 Through 24-02-2019",

year = "2019",

month = feb,

day = "22",

doi = "10.1145/3318299.3318327",

language = "English",

isbn = "9781450366007",

pages = "295--300",

booktitle = "ICMLC '19: Proceedings of the 2019 11th International Conference on Machine Learning and Computing",

publisher = "Association for Computing Machinery (ACM)",

address = "United States",

url = "https://dl.acm.org/doi/proceedings/10.1145/3318299",

}

Ma, S, Yang, L, Wang, H, Xiao, H, Dai, HN, Cheng, S & Wang, T 2019, MHDT: A Deep-Learning-Based Text Detection Algorithm for Unstructured Data in Banking. in ICMLC '19: Proceedings of the 2019 11th International Conference on Machine Learning and Computing. Association for Computing Machinery (ACM), pp. 295-300, 11th International Conference on Machine Learning and Computing, ICMLC 2019, Zhuhai, China, 22/02/19. https://doi.org/10.1145/3318299.3318327

MHDT: A Deep-Learning-Based Text Detection Algorithm for Unstructured Data in Banking. / Ma, Shenglan; Yang, Lingling; Wang, Hao et al.
ICMLC '19: Proceedings of the 2019 11th International Conference on Machine Learning and Computing. Association for Computing Machinery (ACM), 2019. p. 295-300.

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

TY - GEN

T1 - MHDT: A Deep-Learning-Based Text Detection Algorithm for Unstructured Data in Banking

AU - Ma, Shenglan

AU - Yang, Lingling

AU - Wang, Hao

AU - Xiao, Hong

AU - Dai, Hong Ning

AU - Cheng, Shuhan

AU - Wang, Tongsen

N1 - Funding Information: This work is partially funded by the Fujian Fumin Foundation and is supported by the National Natural Science Foundation of China under Grant (No. 61672170 and No. 61871313), the Science and Technology Planning Project of Guangdong Province (No. 2017A050501035), and Science and Technology Program of Guangzhou (No. 201807010058). Publisher Copyright: © 2019 Copyright is held by the owner/author(s). Publication rights licensed to ACM.

PY - 2019/2/22

Y1 - 2019/2/22

N2 - Text detection in natural scene images becomes highly demanded for unstructured data in banking. In this paper, we propose a new deep learning algorithm called MSER, Hu-moment and Deep learning for Text detection (MHDT) based on Maximum Stable Extremal Regions (MSER) and Hu-moment features. Firstly, we extract MSERs as candidate characters. Secondly, a character classifier is introduced with Hu-moment features to reduce the number of input for clustering. After single linkage clustering, a text classifier trained from a Deep Brief Network is used to delete non-text. The proposed algorithm is evaluated on the ICDAR database, and the experimental results show that the proposed algorithm yields high precision and recall rate.

AB - Text detection in natural scene images becomes highly demanded for unstructured data in banking. In this paper, we propose a new deep learning algorithm called MSER, Hu-moment and Deep learning for Text detection (MHDT) based on Maximum Stable Extremal Regions (MSER) and Hu-moment features. Firstly, we extract MSERs as candidate characters. Secondly, a character classifier is introduced with Hu-moment features to reduce the number of input for clustering. After single linkage clustering, a text classifier trained from a Deep Brief Network is used to delete non-text. The proposed algorithm is evaluated on the ICDAR database, and the experimental results show that the proposed algorithm yields high precision and recall rate.

KW - Deep learning

KW - Text detection

KW - Unstructured data

UR - http://www.scopus.com/inward/record.url?scp=85066464992&partnerID=8YFLogxK

U2 - 10.1145/3318299.3318327

DO - 10.1145/3318299.3318327

M3 - Conference proceeding

AN - SCOPUS:85066464992

SN - 9781450366007

SP - 295

EP - 300

BT - ICMLC '19: Proceedings of the 2019 11th International Conference on Machine Learning and Computing

PB - Association for Computing Machinery (ACM)

T2 - 11th International Conference on Machine Learning and Computing, ICMLC 2019

Y2 - 22 February 2019 through 24 February 2019

ER -

MHDT: A Deep-Learning-Based Text Detection Algorithm for Unstructured Data in Banking

Abstract

Conference

Scopus Subject Areas

User-Defined Keywords

Access to Document

Other files and links

Fingerprint

Cite this