Unsupervised Image Classification Using Multi-Autoencoder and K-means++

Authors
Shingo [email protected]
Graduate School of Sciences and Technology for Innovation, Yamaguchi University, Tokiwadai 2-16-1 Ube, Yamaguchi 755-8611, Japan
Kyoichiro Kobayashi
Department of Information Science and Engineering, Faculty of Engineering, Yamaguchi University, Tokiwadai 2-16-1 Ube, Yamaguchi 755-8611, Japan
Masanao [email protected]
Graduate School of Sciences and Technology for Innovation, Yamaguchi University, Tokiwadai 2-16-1 Ube, Yamaguchi 755-8611, Japan
Takashi [email protected]
Graduate School of Sciences and Technology for Innovation, Yamaguchi University, Tokiwadai 2-16-1 Ube, Yamaguchi 755-8611, Japan
Available Online 30 June 2018.
DOI
https://doi.org/10.2991/jrnal.2018.5.1.17
Keywords
neural network; deep autoencoder; K-means++; clustering
Abstract
Supervised learning algorithms such as deep neural networks have been actively applied to various problems. However, in image classification problem, for example, supervised learning needs a large number of data with correct labels. In fact, the cost of giving correct labels to the training data is large; therefore, this paper proposes an unsupervised image classification system with Multi-Autoencoder and K-means++ and evaluates its performance using benchmark image datasets.

Copyright
Copyright © 2018, the Authors. Published by ALife Robotics Corp. Ltd.
Open Access
This is an open access article under the CC BY-NC license (http://creativecommons.org/licences/by-nc/4.0/).

Download article (PDF)