一种基于稀疏矩阵的高效图像压缩算法

针对稀疏矩阵图像提出了一种低复杂度智能无损压缩算法.首先通过图像分块,采用列坐标和标志位相结合表示非常规像素位置;然后基于图像特性分析,得出非常规像素列坐标相对值的概率分布,从而根据非常规像素位置分布的连续性,对其相对列坐标提出一种基于Rice编码的智能码字分配方法.并且,针对非常规像素灰度值分布的连续性,引入预测的方式去除统计相关,对预测冗余采用简单有效的Golomb编码.实验结果表明,该算法压缩效率高,且结构简单,易于硬件实现.

An efficient image compression algorithm based on sparse matrix

Aimed at sparse matrix image,Low-complexity Intelligent Lossless Compression(LILC) algorithm was proposed.By blocking the image,column coordinate was combined with the flag bit to express the position of the non-regular pixels.On the basis of the analysis of image characteristic,the probability distribution of the column relative coordinate was derived.A Rice-based intelligent codes assigned in column relative coordinate was proposed according to the continuity of the position of non-regular pixels.According to the continuity of the gray-level of non-regular pixels,the prediction was used to do the statistic correlation.Golomb code was simple and efficient for image prediction residuals.The experiment results showed that the algorithm was characterized by high compression efficiency,simple structure and easy for hardware implementation.