一种渐晕纹理图像自动分类方法

传统纹理分类方法对光照比较敏感，不均匀的光照分布（如渐晕）会在很大程度上影响纹理分类的准确率．为解决此类问题，针对渐晕纹理图像，提出了一种纹理图像自动分类方法；在利用小波包提取纹理指数算法的基础上，根据渐晕系数自动调整各小波包分解系数，从而消除了渐晕现象对纹理特征指数的影响，最终提高了纹理分类的准确率．仿真实验结果表明，利用此方法对渐晕纹理图像进行分类，准确率有了较大程度的提高，取得了比较理想的分类效果．

An Automatic Classification Approach to Vignetting Texture Images

Since traditional texture classification methods are usually sensitive to lighting condition, non-uniform light distribution, such as vignetting, will greatly reduce the classification accuracy of texture images. To solve this problem, this paper presented a new approach to automatic classification of vignetting texture images. By extracting texture features with the wavelet packet decomposition algorithm, vignetting coefficients were utilized to adjust the wavelet packet coefficients obtained, thus eliminating the effect of vignetting on texture features, and consequently improving texture classification accuracy. Experimental results show that the approach proposed in this paper can significantly improve classification accuracy and achieve ideal texture classification effect.