M集在分形图像压缩编码中的应用

Fractal Image Compression Encoding Using M Collection

目的构造一个固定的压缩字典,改变传统的一幅图像对应一个压缩字典的分形图像压缩方法,解决Mandelbrot图像在分形图像压缩算法中的应用问题.方法采用函数f(z),改变参数z,生成不同的曲线,用灰度值量化规则进行量化,得到许多幅图像块,可以构成丰富的压缩字典,编码时将父块进行自适应合并分割,与压缩字典中的图像块进行匹配,选出满足条件的图像块,再对该图像块进行编码;解码时读取压缩字典,重建图像.结果该算法编码过程中生成丰富的压缩字典,所以解码图像质量高,并且比传统分形图像压缩算法压缩比高,解码速度快.结论该算法减少了搜索时间.实验证明本算法实现简单、可行,具有良好的压缩效果和高质量的重建图像.

A fixed compression dictionary is created to change the fact that an image has one-to-one compression dictionary and advance the application of the Mandelbrot image in fractal image compression algorithm. The function with different z can create different curves. The paper uses the rule to quantify the curves and then obtain many image blocks, thus abundant compression dictionary is built. During the encoding, the domain blocks are divided up in manner of self-adopt combination and match the image blocks in the compression dictionary, and then the image block which meets the demand will be encoded. During the decoding, reading from the compression dictionary, the former image should be restructured. Using this algorithm, a better result can be obtained because of the abundant compression dictionary. The experiment proves that the algorithm, which has fine compression effect and perfect decoded image, is timesaving, simple and feasible.