Presentation 2022-03-02
A Study on Improvement of Recognition Accuracy and Speed-up of SOM-based Classification System
Shun Tasaka, Hiroomi Hikawa,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper discusses a new type of image classifier called class-SOM, which is based on self-organizing map (SOM). The class-SOM is made of multiple SOMs, each of which is trained with the same class images. Then the SOMs in the class-SOM are replaced with hierarchical SOMs. Since the hierarchical SOM is made of the smaller SOMs, the learning time is expected to be shortened. The proposed system was implemented using multithreading. Experiments with Fashion-MNIST revealed that the proposed classifier provided 83.0% of accuracy with 168 second of training time. Meanwhile the conventional $8 times 8$~SOM-Hebb classifier required 1575 second for training and provided 71.44% accuracy. Accuracy of the classification was finally tuned up to be 85.82%.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Self-Organizing Map / Neural Network / Machine Learning / Classification Systems / Image Recognition
Paper # NC2021-46
Date of Issue 2022-02-23 (NC)

Conference Information
Committee MBE / NC
Conference Date 2022/3/2(3days)
Place (in Japanese) (See Japanese page)
Place (in English) Online
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Ryuhei Okuno(Setsunan Univ.) / Rieko Osu(Waseda Univ.)
Vice Chair Junichi Hori(Niigata Univ.) / Hiroshi Yamakawa(Univ of Tokyo)
Secretary Junichi Hori(Osaka Electro-Communication Univ) / Hiroshi Yamakawa(ATR)
Assistant Jun Akazawa(Meiji Univ. of Integrative Medicine) / Emi Yuda(Tohoku Univ) / Nobuhiko Wagatsuma(Toho Univ.) / Tomoki Kurikawa(KMU)

Paper Information
Registration To Technical Committee on ME and Bio Cybernetics / Technical Committee on Neurocomputing
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Study on Improvement of Recognition Accuracy and Speed-up of SOM-based Classification System
Sub Title (in English)
Keyword(1) Self-Organizing Map
Keyword(2) Neural Network
Keyword(3) Machine Learning
Keyword(4) Classification Systems
Keyword(5) Image Recognition
1st Author's Name Shun Tasaka
1st Author's Affiliation Kansai University(Kansai Univ.)
2nd Author's Name Hiroomi Hikawa
2nd Author's Affiliation Kansai University(Kansai Univ.)
Date 2022-03-02
Paper # NC2021-46
Volume (vol) vol.121
Number (no) NC-390
Page pp.pp.1-4(NC),
#Pages 4
Date of Issue 2022-02-23 (NC)