基于无抽样小波变换和MCE训练的纹理分类

提出了一种新的纹理分类的方法,该方法把基于无抽样小波变换的特征提取器和基于欧几里得距离的分类器进行了合并。把方差、偏态系数、峰态系数、三者的联合及谱直方图作为描述纹理图像不相重叠的图像窗的特征。一个使用线性转换矩阵的特征提取器对分类导向的特征做进一步的提取。利用基于欧几里得距离的分类器,每个纹理图像不相重叠的图像窗被确定到属于它的那一类。基于最小分类错误训练方法的特征提取器和分类器设计的合并使分类错误达到了最小化。使用该方法对25类BrodTex纹理图像进行了评估,分类精确度达到90%以上。

Texture classification based on undecimated wavelet transform and MCE training

This paper proposes a new method for texture of classification by incorporating the design of undecimated wavelet transform based feature extractor with the design of an Euclidean distance based classifier.Variance,skewness,kurtosis,the combination of the above three statistical moments and spectral histograms at the outputs of undecimated wavelet decomposition are used to characterize each nonoverlapping window of the texture image.A feature extractor using linear transformation matrix is further employed to extract the classification-oriented features.With an Euclidean distance based classifier,each nonoverlapping window of the texture image is then assigned to its corresponding category.Minimization of the classification error is achieved by incorporating the design of the feature extractor with the design of the classifier based on Minimum Classification Error(MCE)training method.The proposed method has been evaluated on the classification of 25 BrodTex texture,and more than 90% classification accuracy has been achieved.