Combination of Convolutional Neural Network Architecture and its Learning Method for Rotation-Invariant Handwritten Digit Recognition

被引：4

作者：

Urazoe, Kazuya ^{[1
]}

Kuroki, Nobutaka ^{[1
]}

Hirose, Tetsuya ^{[2
]}

Numa, Masahiro ^{[1
]}

机构：

[1] Kobe Univ, Dept Elect & Elect Engn, Nada Ku, 1-1 Rokkodai Cho, Kobe, Hyogo 6578501, Japan

[2] Osaka Univ, Div Elect Elect & Informat Engn, 2-1 Yamadaoka, Suita, Osaka 5650871, Japan

来源：

IEEJ TRANSACTIONS ON ELECTRICAL AND ELECTRONIC ENGINEERING | 2021年 / 16卷 / 01期

关键词：

handwritten digit recognition; convolutional neural network; rotation invariance; data augmentation; MNIST;

D O I：

10.1002/tee.23278

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

This letter presents several combinations of a convolutional neural network (CNN) and its learning method for rotation-invariant digit recognition. Rotation data augmentation is widely used for improving rotation invariance. Data augmentation commonly assigns the same label to all augmented images of the same source. However, this learning method causes some collisions between original and rotated digits. Thus, this letter presents three types of rotation-invariance learning methods and applies them to five popular CNN architectures. Experimental results indicate that multi-task learning on ResNet-50 is the best combination. (c) 2020 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

引用

页码：161 / 163

页数：3

共 50 条

[21] HANDWRITTEN DIGIT RECOGNITION - APPLICATIONS OF NEURAL NETWORK CHIPS AND AUTOMATIC LEARNING
LECUN, Y
JACKEL, LD
BOSER, B
DENKER, JS
GRAF, HP
GUYON, I
HENDERSON, D
HOWARD, RE
HUBBARD, W
IEEE COMMUNICATIONS MAGAZINE, 1989, 27 (11) : 41 - 46
[22] Handwritten digit recognition of an integrated neural network
Luo, Xiaobin
Dong, Shoubin
Jin, Lianwen
Xu, Hao
Jisuanji Gongcheng/Computer Engineering, 2002, 28 (08):
[23] Handwritten digit recognition based on a neural SVM combination
Nemmour H.
Chibani Y.
International Journal of Computers and Applications, 2010, 32 (01) : 104 - 109
[24] Image Augmentation by Blocky Artifact in Deep Convolutional Neural Network for Handwritten Digit Recognition
Shopon, Md
Mohammed, Nabeel
Abedin, Md Anowarul
2017 IEEE INTERNATIONAL CONFERENCE ON IMAGING, VISION & PATTERN RECOGNITION (ICIVPR), 2017,
[25] Bangla Handwritten Character and Digit Recognition Using Deep Convolutional Neural Network on Augmented Dataset and Its Applications
Huda, Hasibul
Fahad, Md Ariful Islam
Islam, Moonmoon
Das, Amit Kumar
PROCEEDINGS OF THE 2022 16TH INTERNATIONAL CONFERENCE ON UBIQUITOUS INFORMATION MANAGEMENT AND COMMUNICATION (IMCOM 2022), 2022,
[26] Rotation-invariant convolutional neural networks for galaxy morphology prediction
Dieleman, Sander
Willett, Kyle W.
Dambre, Joni
MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY, 2015, 450 (02) : 1441 - 1459
[27] Rotation-Invariant Gait Identification with Quaternion Convolutional Neural Networks
Jing, Bowen
Prabhu, Vinay
Gu, Angela
Whaley, John
THIRTY-FIFTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, THIRTY-THIRD CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE AND THE ELEVENTH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2021, 35 : 15805 - 15806
[28] An Efficient Handwritten Digit Recognition Based on Convolutional Neural Networks with Orthogonal Learning Strategies
Senthil, T.
Rajan, C.
Deepika, J.
INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE, 2022, 36 (01)
[29] Cascaded Heterogeneous Convolutional Neural Networks for Handwritten Digit Recognition
Wu, Chunpeng
Fan, Wei
He, Yuan
Sun, Jun
Naoi, Satoshi
2012 21ST INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR 2012), 2012, : 657 - 660
[30] Fuzzy hyperline segment neural network for rotation invariant handwritten character recognition
Kulkarni, UV
Sontakke, TR
Randale, GD
IJCNN'01: INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1-4, PROCEEDINGS, 2001, : 2918 - 2923

← 1 2 3 4 5 →