为了在高信噪比条件下来对静态彩色图像进行高倍压缩，提出了一种具有自适应功能的多维离散余弦变换图像编码算法。该算法首先将图像分割成互不覆盖的 8× 8子块;然后抽取每个子块的统计特征，再根据子块的方差和均值来判断子块内部信息的丰富程度，其中对于内部信息丰富的子块，可利用三维DCT编码来消除块内相邻像素间、色彩空间Y、U、V各分量间的相关性，而对于内部信息不丰富的子块，则利用四维DCT编码来消除块内相邻像素间、色彩空间Y、U、V各分量间、相邻子块间的相关性;最后对变换系数进行量化编码。实验结果证明，此算法可以提供比JEPG更高的压缩率和更高的平均峰值信噪比

An adaptive multi-dimensional discrete cosine transform (multi-D DCT) algorithm are proposed in order to obtain high compression for still color image under high signal-to-noise ratio. In this algorithm, the original image is splitted into non-over-lapped 8х8 blocks. Then the statistic characteristic of each block is calculated and assess the abundance of inner information in each block based on averages and variances. To the block of which the inner information is abundant, correlations between adjacent pixels of each block and YUV channels of color space are exploited through 3-dimensional discrete cosine transform(3-DDCT);To the block ofwhich the inner information is not, correlations among adjacent pixels of each block ,YUV channels of color space and adjacent blocks are exploited through 4-dimensional discrete cosine transform(4-D DCT). The transform coefficients are coded by quantization further. Experimental results have proved that the proposed algorithm can achieve higher compression ratio and ThePSNR(average Peak Signal-Noise-Ratio) of reconstructed image is higher than by JPEG.