Education, Science, Technology, Innovation and Life
Open Access
Sign In

Study on Text Irregular Image Algorithm Based on Convolutional Neural Network

Download as PDF

DOI: 10.23977/autml.2023.040109 | Downloads: 30 | Views: 608


Aisi Luo 1, Zeyu Jia 1, Hongyu Fu 1, Sheng Han 1


1 Hebei University of Science & Technology, Yuxiang Street, Shijiazhuang, 050024, China

Corresponding Author

Aisi Luo


In view of character defects due to irregular interference of text images, this paper proposed a restoration algorithm model based on convolutional neural network and key point detection, implements restoration training on the defective character areas. By studying and analyzing the characteristics of ancient text images of different styles, a digital text image database was established. Based on the Chinese character regional positioning technology and key point detection technology, the restoration effect was evaluated and the training test was conducted after restoration of defective strokes.


Key point detection, convolutional neural network, Chinese character regional positioning technology, optical character recognition


Aisi Luo, Zeyu Jia, Hongyu Fu, Sheng Han, Study on Text Irregular Image Algorithm Based on Convolutional Neural Network. Automation and Machine Learning (2023) Vol. 4: 57-62. DOI:


[1] Xue Wenjuan. The significance of stone tablets to the study of local history and culture. World of Antiquity, 2014 (01): 64-65+49.
[2] Ilya Blayvas, Alfred M. Bruckstein, and Ron Kimmel. Efficient computation of adaptivethreshold surfaces forimage binarization In 2001 IEEEComputerSociety Conference on Computer Vision andPattern Recognition (CVPR 2001) 2001: 737-742.
[3] Yibing Yang and Hong Yan. An adaptive logical method for binarization of degradeddocument images Pattern Recognition, 2000, 33 (5): 787–807.
[4] Moon-Soo Chang, Sun-Mee Kang, Woo-Sik Rho, Heok-Gu Kim, and Duck-Jin Kim. Im-proved binarization algorithm for document image by histogram and edge detection.In ICDAR. 1995: 636-639.
[5] Xiangyun Ye and Mohamed Cheriet and Ching Y. Suen. Stroke-model-based characterextraction from gray-level document images. IEEETransactions on Image Processing, 2001, 10 (8): 1152–1161.
[6] Junsong Zhang, Guohong Mao, Hongwei Lin, Jinhui Yu, and Changle Zhou. Outlinefont generating from images of ancient chinese calligraphy. T. Edutainment, 2011, 5: 122-131.
[7] Zhang Galun. Let cultural heritage be "immortalized" under digital technology. Science & Technology Daily, 2021-09-17 (008).
[8] SONG Junhua. Some thoughts on the digital protection of intangible cultural heritage. Cultural Heritage, 2015 (02): 1-8+157.

Downloads: 1425
Visits: 62661

Sponsors, Associates, and Links

All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.