基于小波变换和支持向量机的纹理图像分类研究

Texture Image Classification Based on Wavelet Transform and Support Vector Machine

如今在充满着视觉信息的现实世界中,纹理图像由于蕴含着大量的有用信息越来越受到研究者的关注。提出了一种基于小波变换和支持向量机的纹理图像分类方法,即在分析不同纹理图像之间在内容上的显著差异的基础上,先采用双树复小波变换对图像进行基于纹理特征的提取操作,再利用支持向量机对不同纹理图像进行分类。提出的方法在包含200幅不同纹理图像的图像集上进行实验验证,获得了较好的分类结果。

Nowadays, we have witnessed the exponential growth of visual information with the development of computer technologyin the real world. Among the large number of visual information, texture images are getting more and more attentions for research-ers since they contain many useful information. In this paper, we propose a new texture image classification method based on wave-let transform and Support Vector Machine(SVM). Firstly, based on the analysis of the significant differences among texture images,the texture features of these images are extracted by using the Dual Tree-Complex Wavelet Transform(DT-CWT). And then, theseimages are classified by applying SVM. Some experiments are conducted on an image set which includes 200 images, and the re-sults confirm that the proposed method obtains a better classification performance.