Classification of handwritten Chinese numbers with convolutional neural networks

Rasoul Ameri; Ali Alameer; Saideh Ferdowsi; Vahid Abolghasemi; Kianoush Nazarpour

doi:10.1109/IPRIA53572.2021.9483557

Classification of handwritten Chinese numbers with convolutional neural networks

Rasoul Ameri, Ali Alameer, Saideh Ferdowsi, Vahid Abolghasemi, Kianoush Nazarpour

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Deep learning methods have become the key ingredient in the field of computer vision; in particular, convolutional neural networks (CNNs). Appropriating the network architecture and data pre-processing have significant impact on performance. This paper focuses on the classification of handwritten Chinese numbers. Firstly, we applied various methods of pre-processing to our collected image dataset. Secondly, we customised a CNN-based architecture with minimal number of layers and parameters specifically for the task. Experimental results showed that our proposed methods provides superior classification rate of 99.1%. Our results also show that the proposed method has competitive performance compared to smaller neural networks with fewer parameters, e.g. Squeezenet and deeper networks with a larger size and number of parameters, e.g., pre-trained GoogLeNet and MobileNetV2.

Original language	English
Title of host publication	Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021
Publisher	Institute of Electrical and Electronics Engineers
ISBN (Electronic)	9781665426596
ISBN (Print)	9781665426602
DOIs	https://doi.org/10.1109/IPRIA53572.2021.9483557
Publication status	Published - 26 Jul 2021
Event	5th International Conference on Pattern Recognition and Image Analysis - Kashan, Iran, Islamic Republic of Duration: 28 Apr 2021 → 29 Apr 2021

Publication series

Name	Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021
Publisher	IEEE
ISSN (Electronic)	2049-3630

Conference

Conference	5th International Conference on Pattern Recognition and Image Analysis
Abbreviated title	IPRIA 2021
Country/Territory	Iran, Islamic Republic of
City	Kashan
Period	28/04/21 → 29/04/21

Keywords / Materials (for Non-textual outputs)

Chinese number classification
convolutional neural network
deep learning
hand written recognition
image processing

Access to Document

10.1109/IPRIA53572.2021.9483557Licence: All Rights Reserved

https://ieeexplore.ieee.org/document/9483557Licence: All Rights Reserved

Cite this

Ameri, R., Alameer, A., Ferdowsi, S., Abolghasemi, V., & Nazarpour, K. (2021). Classification of handwritten Chinese numbers with convolutional neural networks. In Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021 Article 9483557 (Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021). Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/IPRIA53572.2021.9483557

Ameri, Rasoul ; Alameer, Ali ; Ferdowsi, Saideh et al. / Classification of handwritten Chinese numbers with convolutional neural networks. Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021. Institute of Electrical and Electronics Engineers, 2021. (Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021).

@inproceedings{0697aa5e63c4414a9dbaf7979c67b150,

title = "Classification of handwritten Chinese numbers with convolutional neural networks",

abstract = "Deep learning methods have become the key ingredient in the field of computer vision; in particular, convolutional neural networks (CNNs). Appropriating the network architecture and data pre-processing have significant impact on performance. This paper focuses on the classification of handwritten Chinese numbers. Firstly, we applied various methods of pre-processing to our collected image dataset. Secondly, we customised a CNN-based architecture with minimal number of layers and parameters specifically for the task. Experimental results showed that our proposed methods provides superior classification rate of 99.1%. Our results also show that the proposed method has competitive performance compared to smaller neural networks with fewer parameters, e.g. Squeezenet and deeper networks with a larger size and number of parameters, e.g., pre-trained GoogLeNet and MobileNetV2. ",

keywords = "Chinese number classification, convolutional neural network, deep learning, hand written recognition, image processing",

author = "Rasoul Ameri and Ali Alameer and Saideh Ferdowsi and Vahid Abolghasemi and Kianoush Nazarpour",

note = "Publisher Copyright: {\textcopyright} 2021 IEEE.; 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021 ; Conference date: 28-04-2021 Through 29-04-2021",

year = "2021",

month = jul,

day = "26",

doi = "10.1109/IPRIA53572.2021.9483557",

language = "English",

isbn = "9781665426602",

series = "Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021",

publisher = "Institute of Electrical and Electronics Engineers",

booktitle = "Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021",

address = "United States",

}

Ameri, R, Alameer, A, Ferdowsi, S, Abolghasemi, V & Nazarpour, K 2021, Classification of handwritten Chinese numbers with convolutional neural networks. in Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021., 9483557, Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021, Institute of Electrical and Electronics Engineers, 5th International Conference on Pattern Recognition and Image Analysis, Kashan, Iran, Islamic Republic of, 28/04/21. https://doi.org/10.1109/IPRIA53572.2021.9483557

Classification of handwritten Chinese numbers with convolutional neural networks. / Ameri, Rasoul; Alameer, Ali; Ferdowsi, Saideh et al.
Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021. Institute of Electrical and Electronics Engineers, 2021. 9483557 (Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Classification of handwritten Chinese numbers with convolutional neural networks

AU - Ameri, Rasoul

AU - Alameer, Ali

AU - Ferdowsi, Saideh

AU - Abolghasemi, Vahid

AU - Nazarpour, Kianoush

PY - 2021/7/26

Y1 - 2021/7/26

N2 - Deep learning methods have become the key ingredient in the field of computer vision; in particular, convolutional neural networks (CNNs). Appropriating the network architecture and data pre-processing have significant impact on performance. This paper focuses on the classification of handwritten Chinese numbers. Firstly, we applied various methods of pre-processing to our collected image dataset. Secondly, we customised a CNN-based architecture with minimal number of layers and parameters specifically for the task. Experimental results showed that our proposed methods provides superior classification rate of 99.1%. Our results also show that the proposed method has competitive performance compared to smaller neural networks with fewer parameters, e.g. Squeezenet and deeper networks with a larger size and number of parameters, e.g., pre-trained GoogLeNet and MobileNetV2.

AB - Deep learning methods have become the key ingredient in the field of computer vision; in particular, convolutional neural networks (CNNs). Appropriating the network architecture and data pre-processing have significant impact on performance. This paper focuses on the classification of handwritten Chinese numbers. Firstly, we applied various methods of pre-processing to our collected image dataset. Secondly, we customised a CNN-based architecture with minimal number of layers and parameters specifically for the task. Experimental results showed that our proposed methods provides superior classification rate of 99.1%. Our results also show that the proposed method has competitive performance compared to smaller neural networks with fewer parameters, e.g. Squeezenet and deeper networks with a larger size and number of parameters, e.g., pre-trained GoogLeNet and MobileNetV2.

KW - Chinese number classification

KW - convolutional neural network

KW - deep learning

KW - hand written recognition

KW - image processing

UR - http://www.scopus.com/inward/record.url?scp=85112086514&partnerID=8YFLogxK

U2 - 10.1109/IPRIA53572.2021.9483557

DO - 10.1109/IPRIA53572.2021.9483557

M3 - Conference contribution

AN - SCOPUS:85112086514

SN - 9781665426602

T3 - Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021

BT - Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021

PB - Institute of Electrical and Electronics Engineers

T2 - 5th International Conference on Pattern Recognition and Image Analysis

Y2 - 28 April 2021 through 29 April 2021

ER -

Ameri R, Alameer A, Ferdowsi S, Abolghasemi V, Nazarpour K. Classification of handwritten Chinese numbers with convolutional neural networks. In Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021. Institute of Electrical and Electronics Engineers. 2021. 9483557. (Proceedings of the 5th International Conference on Pattern Recognition and Image Analysis, IPRIA 2021). doi: 10.1109/IPRIA53572.2021.9483557

Classification of handwritten Chinese numbers with convolutional neural networks

Abstract

Publication series

Conference

Keywords / Materials (for Non-textual outputs)

Access to Document

Other files and links

Fingerprint

Cite this