基于Contourlet的图像压缩算法

Image Compression Based on Contourlet Transform

Contourlet变换结合了方向滤波组,具备小波变换不能表达的多方向性,能够很有效地捕获自然图像的边缘轮廓信息。在JPEG2000的压缩标准中,它采用了小波变换和死区均匀量化。鉴于JPEG2000标准,本文提出一套新的编码方案——Contourlet变换和最佳量化器(Lloyd-Max量化器)。同样,本文也将死区量化应用到最佳量化器中。

Combining non-separable and directional filters banks, the Contourlet transform can effectively capture more edges and contours in natural images than wavelets due to its capability of representing directional information. In view of JPEG2000 compression standard, which uses the wavelet transform and dead-zone uniform quantizer, the paper proposes a new coding scheme based on Contourlet transform and optimal quantizer. Similarly, this article is applied to the dead-zone quantization in the optimal quantizer.