Holistic Persian handwritten word recognition using convolutional neural network

被引：0

作者：

Zohrevand A. ^{[1
]}

Imani Z. ^{[1
]}

机构：

[1] Computer Engineering Department, Kosar University of Bojnord, Bojnord

来源：

International Journal of Engineering, Transactions B: Applications | 2021年 / 34卷 / 08期

关键词：

Convolutional neural network; End-to-end learning method; Persian handwritten dataset; Persian handwritten word recognition; Transfer learning;

D O I：

10.5829/ije.2021.34.08b.24

中图分类号：

学科分类号：

摘要：

Due to the cursive-ness and high variability of Persian script, the segmentation of handwritten words into sub-words is still a challenging task. These issues could be addressed in a holistic approach by sidestepping segmentation at the character level. In this paper, an end-to-end holistic method based on deep convolutional neural network is proposed to recognize off-line Persian handwritten words. The proposed model uses only five convolutional layers and two fully connected layers for classifying word images effectively, which can lead to a substantial reduction in the required parameters. The effect of various pooling strategies is also investigated in this paper. The primary goal of this article is to ignore handcrafted feature extraction and to attain a generalized and stable word recognition system. The presented model is assessed using two famous handwritten Persian word databases called Sadri and IRANSHAHR. The recognition accuracies were obtained at 98.6% and 94.6%, on Sadri and IRANSHAHR datasets respectively, and outperformed the state-of-the-art methods. © 2021 Materials and Energy Research Center. All rights reserved.

引用

页码：2028 / 2037

页数：9

共 34 条

[1]

Das D., Nayak D.R., Dash R., Majhi B., Zhang Y.-D., Hwordnet: A holistic convolutional neural network approach for handwritten word recognition, IET Image Processing, 14, 9, pp. 1794-1805, (2020)

[2]

Imani Z., Ahmadyfard Z., Zohrevand A., Holistic farsi handwritten word recognition using gradient features, Journal of AI and Data Mining, 4, 1, pp. 19-25, (2016)

[3]

Akbari Y., Jalili M.J., Sadri J., Nouri K., Siddiqi I., Djeddi C., A novel database for automatic processing of persian handwritten bank checks, Pattern Recognition, 74, pp. 253-265, (2018)

[4]

Ye M., Viola P., Raghupathy S., Sutanto H., Li C., Learning to group text lines and regions in freeform handwritten notes, Ninth International Conference on Document Analysis and Recognition (ICDAR 2007), 1, pp. 28-32, (2007)

[5]

Razzak I., Kamran I., Naz S., Deep analysis of handwritten notes for early diagnosis of neurological disorders, 2020 International Joint Conference on Neural Networks (IJCNN), pp. 1-6, (2020)

[6]

Vajda S., Roy K., Pal U., Chaudhuri B.B., Belaid A., Automation of indian postal documents written in bangla and english, International Journal of Pattern Recognition and Artificial Intelligence, 23, pp. 1599-1632, (2009)

[7]

Venu G., Xue H., Fast handwriting recognition for indexing historical documents, First International Workshop on Document Image Analysis for Libraries, 2004. Proceedings, pp. 314-320, (2004)

[8]

Sanchez J.A., Romero V., Toselli A.H., Villegas M., Vidal E., A set of benchmarks for handwritten text recognition on historical documents, Pattern Recognition, 94, pp. 122-134, (2019)

[9]

Bhowmik S., Malakar S., Sarkar R., Basu S., Kundu M., Nasipuri M., Off-line bangla handwritten word recognition: A holistic approach, Neural Computing and Applications, 31, 10, pp. 5783-5798, (2019)

[10]

Abbaszadeh Arani S.A.A., Kabir E., Ebrahimpour R., Combining rtl and ltr hmms to recognise handwritten farsi words of small‐and medium‐sized vocabularies, IET Computer Vision, 12, 6, pp. 925-932, (2018)

← 1 2 3 4 →