DocumentCode :
1391726
Title :
Invariant recognition system for 2D objects based on three-stage BPNN structure
Author :
Kunchev, R. ; Topalova, I.
Author_Institution :
Lab. of Image Process., Darveniza Tech. Univ., Sofia, Bulgaria
Volume :
34
Issue :
10
fYear :
1998
fDate :
5/14/1998 12:00:00 AM
Firstpage :
974
Lastpage :
976
Abstract :
A new recognition system is presented which includes a three-stage backpropagation neural network structure with two-band coding of the DCT input spectrum space. The system is designed for invariant recognition of 2D objects. It enables a high recognition rate of objects with overlapping classes to be obtained
Keywords :
backpropagation; discrete cosine transforms; image recognition; neural nets; object recognition; 2D objects; DCT input spectrum space; backpropagation neural network structure; high recognition rate; invariant recognition system; overlapping classes; three-stage BPNN structure; two-band coding;
fLanguage :
English
Journal_Title :
Electronics Letters
Publisher :
iet
ISSN :
0013-5194
Type :
jour
DOI :
10.1049/el:19980603
Filename :
682943
Link To Document :
بازگشت