Journal of Transportation Systems Engineering and Information Technology

Journal of Transportation Systems Engineering and Information Technology ›› 2003, Vol. 3 ›› Issue (4): 34-37 .

• ITS System and Technology • Previous Articles     Next Articles

Vehicle Recognition Based on 1 -- SVM

SUN De- shan 1，2 ,WU Jin-pei3   

  1. 1. School of Mathematical Science and Computing Technology, Central South University, Changsha 410075, China；2.Department of Mathematics, Liaoning Normal University, Dalian 116029, China；3.Institute of Intelligent Technology & System, Wuyi University, Jiangmen 529020, China
  • Received:2003-09-17 Revised:1900-01-01 Online:2003-11-01 Published:2003-11-01

一类支持向量机在车辆识别中的应用

孙德山1，2，吴今培3   

  1. 1.中南大学数学科学与计算技术学院，长沙410075； 2.辽宁师范大学数学系，大连，116029；3. 五邑大学智能技术与系统研究所，江门529020

PDF

Abstract: The Support Vector Machine (SVM) has shown excellent performance
in practice as a classification methodology, which were originally designed for binary classification. How to effectively extend SVM for multi-class classification is still an on-going research issue. Oftentimes multi-class classification problem have been treated as a series of binary problems in the SVM paradigm, but it is computationally more expensive to solve multi-class problems. In this paper, a new multi-class classification method is proposed based on one-class support vector machine (1-SVM), and then applies the method to vehicle recognition. The results of simulation experiments at vehicle database show that the proposed method is effective and feasible.

Key words: 1-SVM, kernel function, multi-class classification, vehicle recognition

摘要： 支持向量机分类方法已经在实际应用中显示了良好的学习性能，其最初是针对二值分类问题提出的.如何有效地将支持向量机推广到多值分类中一直是人们关注的课题.通常的多值分类问题是一系列二值分类来实现，可是这将导致较高的计算复杂性.本文将一类支持向量机推广到多值分类情况，并将其应用于车辆识别中.仿真实验结果表明了所给方法的可行性及有效性.

关键词: 一类支持向量机, 核函数, 多值分类, 车辆识别