基于最小二乘支持向量机的车牌字符特征分类研究

LSSVM-based License Plate Character Feature Classification

最小二乘支持向量机是一种新的有效的机器学习算法。论文介绍了最小二乘支持向量机模型,研究了最小二乘支持向量机算法和经典的多类分类算法,提取车牌字符的奇异值特征,将奇异值系数特征作为最小二乘支持向量机的输入进行训练和分类。实验采用LS-SVM工具箱,得到了较好的结果。

Least Square Support Vector Machine（LSSVM） is a kind of novel machine learning method. This paper in troduces the LSSVM model, LSSVM algorithm and the classic multiple classification algorithm is also studied in this paper. The singular value feature of license plate characters is extracted, then LSSVM is used to train these features and to classify. Using the LSSVM toolbox, the experimental results demonstrate the efficency of the proposed approach.