摘要
鉴于纹理特征对于图像分类的良好性能,提出了结合离散傅里叶变换和排列组合熵的纹理特征分析方法。利用主成分分析方法对特征向量进行降维,再采用支持向量机方法对纹理图像进行分类,取得了较好的效果。
As the texture feature output has superior performance in image classification, a method based on discrete Fourier transform and permutation entropy is proposed. The method of principal components analysis is a- dopted to reduce the dimension of the eigenvector. The Support Vector Machine （SVM） is used to classify the texture image. The results show the methods perform well in texture image classification.
出处
《电子科技》
2011年第1期9-11,共3页
Electronic Science and Technology
关键词
离散傅里叶变换
排列组合熵
主成分分析
支持向量机
纹理图像
discrete Fourier transform
permutation entropy
principal components analysis
support vectormachine
texture image
