基于色彩向量模型的彩色图象编码算法

在彩色图像编码技术中, 普遍存在的一个问题就是未能充分利用个色彩分量间的相关性。本文提出了一种基于色彩向量模型的彩色图象压缩编码算法, 高效地去除了彩色图像各分量间的相关性。该算法是将彩色图像 3 个色彩分量中一个分量, 将该色彩分量分割成若干个大小相等, 但不重叠的子块; 然后构造出该色彩分量的每个子块与另两个色彩分量所对应的子块中的色彩向量; 这样对另两个色彩分量的压缩编码转化为对色彩向量的编码。解码则是通过简单的逆变换, 以得到重构的两个色彩分量。这样一幅彩色图像就可以用一个色彩分量和其子块分割并计算出表示其它两个色彩分量的色彩向量来表示。实验结果标明, 该算法易于实现, 具有较高的峰值噪比和压缩比, 且恢复图像也具有视觉无损的效果。

An algorithm for the color image coding by the color-vector model

It is a common phenomenon that is making little advantage of the correlation of color components in color image coding. An algorithm for the color image coding by the color- vector model is presented. In this algorithm, one of the three color components of a color image is segmented into several non- overlapping sub- regions in the same size. And then two color- vectors is structured by the correlation coefficients, which is ciphered out using chromatic m matrix analysis. The chromatic m matrixes are figured out by the m region between the component and the other two components in each sub- regions. Based on the algorithm,, the coding for the other two components becomes the coding of two color- vectors. During decoding, components are reconstructed by the simple inverse transform. Then a piece of color image is represented by one of the three color components of a color image and two color- vectors. Experimental results show that the algorithm has characteristics of higher coding and decoding rates, peak signal- to- noise ratio and compression ratio, so the reconstructed image is lossless in the visual perception.