Hypercolumn Model

A Combination Model of Hierarchical Self-Organizing Maps and Neocognitron for Image Recognition

Naoyuki Tsuruta, Rin-Ichiro Taniguchi, Makoto Amamiya

研究成果: ジャーナルへの寄稿記事

13 引用 (Scopus)

抄録

In this paper, a neural network model, the hypercolumn model (HCM), which is applicable to general image recognition, is proposed. The HCM is a combination model of hierarchical self-organizing maps (HSOM) and neocognitron (NC); it resolves the disadvantages of both the HSOM and the NC, and inherits all of the advantages of both models. The HSOM quantizes and nonlinearly maps an input space into a small dimensional feature map. Using this feature map, the HSOM can function as an unsupervised clustering method and can easily solve pattern recognition problems, which are hard for traditional linear methods or for multilayered perceptrons using the back-propagation learning method because of the complex shapes of the class boundaries. On the other hand, the NC is a multilayered network, in which layers to extract local features and layers to integrate shifted features alternate. Using this structure, the NC can reduce the dimensionality gradually and recognize shifted and scaled patterns. By combining the characteristics of the HSOM and the structure of the NC, the HCM can be made applicable to general image recognition problems which have high dimensional input data and in which the shapes of the class boundaries are quite complex because of the high abstraction level of the classification.

元の言語英語
ページ(範囲)49-61
ページ数13
ジャーナルSystems and Computers in Japan
31
発行部数2
DOI
出版物ステータス出版済み - 1 1 2000

Fingerprint

Image recognition
Image Recognition
Self organizing maps
Self-organizing Map
Unsupervised Clustering
Model
Local Features
Neural networks
Back Propagation
Perceptron
Clustering Methods
Neural Network Model
Alternate
Pattern Recognition
Dimensionality
Resolve
Backpropagation
High-dimensional
Pattern recognition
Integrate

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Information Systems
  • Hardware and Architecture
  • Computational Theory and Mathematics

これを引用

Hypercolumn Model : A Combination Model of Hierarchical Self-Organizing Maps and Neocognitron for Image Recognition. / Tsuruta, Naoyuki; Taniguchi, Rin-Ichiro; Amamiya, Makoto.

:: Systems and Computers in Japan, 巻 31, 番号 2, 01.01.2000, p. 49-61.

研究成果: ジャーナルへの寄稿記事

@article{8322f74bd01940ceb66917bdcf483105,
title = "Hypercolumn Model: A Combination Model of Hierarchical Self-Organizing Maps and Neocognitron for Image Recognition",
abstract = "In this paper, a neural network model, the hypercolumn model (HCM), which is applicable to general image recognition, is proposed. The HCM is a combination model of hierarchical self-organizing maps (HSOM) and neocognitron (NC); it resolves the disadvantages of both the HSOM and the NC, and inherits all of the advantages of both models. The HSOM quantizes and nonlinearly maps an input space into a small dimensional feature map. Using this feature map, the HSOM can function as an unsupervised clustering method and can easily solve pattern recognition problems, which are hard for traditional linear methods or for multilayered perceptrons using the back-propagation learning method because of the complex shapes of the class boundaries. On the other hand, the NC is a multilayered network, in which layers to extract local features and layers to integrate shifted features alternate. Using this structure, the NC can reduce the dimensionality gradually and recognize shifted and scaled patterns. By combining the characteristics of the HSOM and the structure of the NC, the HCM can be made applicable to general image recognition problems which have high dimensional input data and in which the shapes of the class boundaries are quite complex because of the high abstraction level of the classification.",
author = "Naoyuki Tsuruta and Rin-Ichiro Taniguchi and Makoto Amamiya",
year = "2000",
month = "1",
day = "1",
doi = "10.1002/(SICI)1520-684X(200002)31:2<49::AID-SCJ6>3.0.CO;2-8",
language = "English",
volume = "31",
pages = "49--61",
journal = "Systems and Computers in Japan",
issn = "0882-1666",
publisher = "John Wiley and Sons Inc.",
number = "2",

}

TY - JOUR

T1 - Hypercolumn Model

T2 - A Combination Model of Hierarchical Self-Organizing Maps and Neocognitron for Image Recognition

AU - Tsuruta, Naoyuki

AU - Taniguchi, Rin-Ichiro

AU - Amamiya, Makoto

PY - 2000/1/1

Y1 - 2000/1/1

N2 - In this paper, a neural network model, the hypercolumn model (HCM), which is applicable to general image recognition, is proposed. The HCM is a combination model of hierarchical self-organizing maps (HSOM) and neocognitron (NC); it resolves the disadvantages of both the HSOM and the NC, and inherits all of the advantages of both models. The HSOM quantizes and nonlinearly maps an input space into a small dimensional feature map. Using this feature map, the HSOM can function as an unsupervised clustering method and can easily solve pattern recognition problems, which are hard for traditional linear methods or for multilayered perceptrons using the back-propagation learning method because of the complex shapes of the class boundaries. On the other hand, the NC is a multilayered network, in which layers to extract local features and layers to integrate shifted features alternate. Using this structure, the NC can reduce the dimensionality gradually and recognize shifted and scaled patterns. By combining the characteristics of the HSOM and the structure of the NC, the HCM can be made applicable to general image recognition problems which have high dimensional input data and in which the shapes of the class boundaries are quite complex because of the high abstraction level of the classification.

AB - In this paper, a neural network model, the hypercolumn model (HCM), which is applicable to general image recognition, is proposed. The HCM is a combination model of hierarchical self-organizing maps (HSOM) and neocognitron (NC); it resolves the disadvantages of both the HSOM and the NC, and inherits all of the advantages of both models. The HSOM quantizes and nonlinearly maps an input space into a small dimensional feature map. Using this feature map, the HSOM can function as an unsupervised clustering method and can easily solve pattern recognition problems, which are hard for traditional linear methods or for multilayered perceptrons using the back-propagation learning method because of the complex shapes of the class boundaries. On the other hand, the NC is a multilayered network, in which layers to extract local features and layers to integrate shifted features alternate. Using this structure, the NC can reduce the dimensionality gradually and recognize shifted and scaled patterns. By combining the characteristics of the HSOM and the structure of the NC, the HCM can be made applicable to general image recognition problems which have high dimensional input data and in which the shapes of the class boundaries are quite complex because of the high abstraction level of the classification.

UR - http://www.scopus.com/inward/record.url?scp=0347074616&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0347074616&partnerID=8YFLogxK

U2 - 10.1002/(SICI)1520-684X(200002)31:2<49::AID-SCJ6>3.0.CO;2-8

DO - 10.1002/(SICI)1520-684X(200002)31:2<49::AID-SCJ6>3.0.CO;2-8

M3 - Article

VL - 31

SP - 49

EP - 61

JO - Systems and Computers in Japan

JF - Systems and Computers in Japan

SN - 0882-1666

IS - 2

ER -