Modified high-order neural network for invariant pattern recognition

被引：32

作者：

Artyomov, E ^{[1
]}

Yadid-Pecht, O ^{[1
]}

机构：

[1] Ben Gurion Univ Negev, VLSI Syst Ctr, IL-84105 Beer Sheva, Israel

来源：

PATTERN RECOGNITION LETTERS | 2005年 / 26卷 / 06期

关键词：

HONN; high-order neural networks; invariant pattern recognition; binary patterns;

D O I：

10.1016/j.patrec.2004.09.029

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

A modification for high-order neural networks (HONN) is described. The proposed modified HONN takes into account prior knowledge of the binary patterns that must be learned. This significantly reduces hence computation time as well as memory requirements for network configuration and weight storage. An "approximately equal triangles" scheme for weight sharing is also proposed. These modifications enable the efficient computation of HONNs for image fields of greater that 100 x 100 pixels without any loss of pattern information. (c) 2004 Elsevier B.V. All rights reserved.

引用

页码：843 / 851

页数：9

共 14 条

[11]

SPIRKOVSKA L, 1990, P IJCNN 90, V1, P21

[12] A NOTE ON A HIGHER-ORDER NEURAL-NETWORK FOR DISTORTION-INVARIANT PATTERN-RECOGNITION [J].

TAKANO, M ;

KANAOKA, T ;

SKRZYPEK, J ;

TOMITA, S .

PATTERN RECOGNITION LETTERS, 1994, 15 (06) :631-635

[13] GENERATION OF MOMENT INVARIANTS AND THEIR USES OF CHARACTER-RECOGNITION [J].

WONG, WH ;

SIU, WC ;

LAM, KM .

PATTERN RECOGNITION LETTERS, 1995, 16 (02) :115-123

[14] Invariant pattern recognition: A review [J].

Wood, J .

PATTERN RECOGNITION, 1996, 29 (01) :1-17

← 1 2 →