Characterizing the SOM feature detectors under various input conditions

College

College of Computer Studies

Department/Unit

Computer Technology

Document Type

Conference Proceeding

Source Title

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Volume

11441 LNAI

First Page

144

Last Page

155

Publication Date

1-1-2019

Abstract

A classifier with self-organizing maps (SOM) as feature detectors resembles the biological visual system learning mechanism. Each SOM feature detector is defined over a limited domain of viewing condition, such that its nodes instantiate the presence of an object's part in the corresponding domain. The weights of the SOM nodes are trained via competition, similar to the development of the visual system. We argue that to approach human pattern recognition performance, we must look for a more accurate model of the visual system, not only in terms of the architecture, but also on how the node connections are developed, such as that of the SOM's feature detectors. This work characterizes SOM as feature detectors to test the similarity of its response vis-á-vis the response of the biological visual system, and to benchmark its performance vis-á-vis the performance of the traditional feature detector convolution filter. We use various input environments i.e. inputs with limited patterns, inputs with various input perturbation and inputs with complex objects, as test cases for evaluation. © Springer Nature Switzerland AG 2019.

html

Digitial Object Identifier (DOI)

10.1007/978-3-030-16142-2_12

Disciplines

Computer Sciences

Keywords

Self-organizing maps; Pattern recognition systems

Upload File

wf_yes

This document is currently not available here.

Share

COinS