DocumentCode :
3268925
Title :
A Study on Hybrid Approaches in Image Recognition System
Author :
Duan, Bobo ; Yang, Chunyang ; Liu, Wei ; Wen, Xuezhi ; Yuan, Huai
Author_Institution :
Northeastern Univ., Shenayng
fYear :
2007
fDate :
13-15 June 2007
Firstpage :
1072
Lastpage :
1076
Abstract :
Hybrid approaches have been applied widely in image recognition system to adapt the increasing requirement of real-time and high accuracy and there are a lot of successful hybrid image recognition algorithms. To use hybrid approaches more effectively, some basic research about it is very necessary. This paper analyzes current typical hybrid samples, then presents basic hybrid image recognition patterns and evaluates their performance about real-time, accuracy, robustness, expandability and flexibility. At last a developing hybrid image recognition system would be introduced as a sample.
Keywords :
image recognition; robust control; accuracy; expandability; flexibility; hybrid approaches; image recognition system; robustness; Automotive engineering; Classification tree analysis; Face recognition; Feedforward neural networks; Handwriting recognition; Image recognition; Neural networks; Pattern recognition; Real time systems; Robustness; Accuracy; Expandability; Flexibility; Hybrid; Image Recognition System; Real-Time; Robustness;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Vehicles Symposium, 2007 IEEE
Conference_Location :
Istanbul
ISSN :
1931-0587
Print_ISBN :
1-4244-1067-3
Electronic_ISBN :
1931-0587
Type :
conf
DOI :
10.1109/IVS.2007.4290259
Filename :
4290259
Link To Document :
بازگشت