SHAPE-RECOGNITION USING THE KOHONEN SELF-ORGANIZING FEATURE MAP

被引：3

作者：

SARKARIA, SS ^{[1
]}

HARGET, AJ ^{[1
]}

CLARIDGE, E ^{[1
]}

机构：

[1] UNIV BIRMINGHAM,SCH COMP STUDIES,BIRMINGHAM B15 2TT,W MIDLANDS,ENGLAND

来源：

PATTERN RECOGNITION LETTERS | 1992年 / 13卷 / 03期

关键词：

FEATURE MAP; SELF-ORGANIZATION; PERCEPTUAL SIGNIFICANCE; CLASSIFICATION;

D O I：

10.1016/0167-8655(92)90059-9

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper presents an account of a study which shows that the Kohonen self-organising algorithm can be applied in the area of shape recognition. The results show that unsupervised training produces a Kohonen feature map capable of recognition and generalisation.

引用

页码：189 / 194

页数：6

共 6 条

[1] THE NEURAL PHONETIC TYPEWRITER
KOHONEN, T
[J]. COMPUTER, 1988, 21 (03) : 11 - 72
[2] KOHONEN T, 1988, SELF ORG ASS MEMORY, P119
[3] Lippmann R. P., 1988, Computer Architecture News, V16, P7, DOI [10.1109/MASSP.1987.1165576, 10.1145/44571.44572]
[4] 3-DIMENSIONAL OBJECT RECOGNITION FROM SINGLE TWO-DIMENSIONAL IMAGES
LOWE, DG
[J]. ARTIFICIAL INTELLIGENCE, 1987, 31 (03) : 355 - 395
[5] Rumelhart D.E., 1986, PARALLEL DISTRIBUTED, P319
[6] SARKARIA SS, 1990, THESIS ASTON U BIRMI

← 1 →