On the Study of Image Characters Location, Segmentation and Pattern Recognition using LS-SVM

Author

Ge, Guangying ; Xu, Jianjian ; Wang, Minghong

Author_Institution

Liaocheng Univ.

Volume

2

fYear

0

fDate

0-0 0

Firstpage

9650

Lastpage

9654

Abstract

Based on pattern recognition theory and support vector machine (SVM) network technology, automatic detection, location, segmentation and recognition of license plates characters are discussed in detail. A new method is presented to adopt gray-grads, shape and posture, vision model and so on. And a new multi-sorts classification method - binary exponent classification is proposed. By comparing least squares support vector machines (LS-SVM) with BP neural network in vehicle license plates pattern recognition and classification, Experimental results showed that SVM is more convenient, fast and nicely of judgment, and will be the mainstream of the future development in the application of finite sample data pattern recognition

Keywords

backpropagation; character recognition; image classification; image segmentation; least squares approximations; neural nets; object detection; support vector machines; back propagation neural network; binary exponent classification; gray grads; image character location; image segmentation; least squares support vector machines; license plate characters; multisorts classification; object detection; pattern recognition; posture; shape; support vector machine; vision model; Character recognition; Image segmentation; Least squares methods; Licenses; Neural networks; Pattern recognition; Shape; Support vector machine classification; Support vector machines; Vehicles; Least Squares Support Vector Machines (LS-SVM); characters pattern recognition and classification; license plates location and segmenting;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Control and Automation, 2006. WCICA 2006. The Sixth World Congress on

Conference_Location

Dalian

Print_ISBN

1-4244-0332-4

Type

conf

DOI

10.1109/WCICA.2006.1713875

Filename

1713875