基于图像平滑度的WBCT算法研究与改进

The Research and Improvement of WBCT Algorithm Based on Image Smoothness

由于WBCT压缩算法对于不同平滑度图像无差别滤波,使得对于较平滑的图像,其恢复图像效果要低于一般小波编码方法,为此,本文针对不同平滑度的图像,通过定义图像平滑度,对图像进行分类,利用多方向多尺度临界采样,进行不同程度的方向滤波。笔者利用变换后的小波系数的特点,采用SPIHT算法实现嵌入式编码,从而改进了WBCT压缩算法。实验证明,本文算法比WBCT算法能更有效地提取图像的边缘和纹理等几何特征。

As WBCT（Wavelet-based Contourlet Transform） compression algorithm filter image for different smoothness with the same way, for smooth image, the effect of restored image from WBCT algorithm is lower than the general wavelet coding. To overcome this disadvantage, image smoothness is defined in this article, and images are classified by the image smoothness. For different smoothness of images, the improved algorithm achieves different degrees of directional filtering through multidirectional and multi- scale transform with critical sampling, then use the SPIHT algorithm to achieve the embedded coding Of images based on characteristics of wavelet coefficients. Experiments show that the proposed algorithms can extract the geometric features such as edges and textures more efficiently than WBCT.