Research on low-quality text image detection and recognition based on 
EDSR and improved ASTER

Yan Wan; Ziang Zou; Li Yao

doi:10.1117/12.2685651

15 August 2023 Research on low-quality text image detection and recognition based on EDSR and improved ASTER

Yan Wan, Ziang Zou, Li Yao

Proceedings Volume 12719, Second International Conference on Electronic Information Technology (EIT 2023); 127190Z (2023) https://doi.org/10.1117/12.2685651
Event: Second International Conference on Electronic Information Technology (EIT 2023), 2023, Wuhan, China

Abstract

This paper proposes an improved ASTER-based low-quality text image recognition algorithm to address the issue of blurry, low-resolution, and poor-quality text image recognition.The basic idea is to use super-resolution reconstruction technology EDSR to convert low-resolution text images into high-resolution text images, thereby enriching text feature information. In addition, the DCNV2 module is added to ASTER's feature extraction network to focus on the extraction of text feature information. Compared with previous methods, the method proposed in this paper significantly improves the recognition rate on seven public datasets, especially in the SVT and SVTP text image datasets, which have poor quality.

Citation Download Citation

Yan Wan, Ziang Zou, and Li Yao "Research on low-quality text image detection and recognition based on EDSR and improved ASTER", Proc. SPIE 12719, Second International Conference on Electronic Information Technology (EIT 2023), 127190Z (15 August 2023); https://doi.org/10.1117/12.2685651

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available