摘要
提出一种基于圆盘旋转匹配的新型分形压缩算法。首先用Carotid-Kundalini函数生成分形集,并用logistic映射生成量化表量化生成的分形集来增加压缩字典的变化性;接着将生成的分形集分割为4×4的小图像块,将其转换为圆盘后进行旋转;再将旋转后的图像块重新变换为4×4正方形块,得到固定可通用的数据字典。扩充了原有的几何变换类型,打破了图像与字典之间的一一对应关系。实验结果表明,本算法简单有效,具有良好的压缩效果和高质量的重建图像,且解码的速度快。
At first the Carotid-Kundalini function was employed to generate the fractal set and then the set was quantized by the quantized table created by the Logistic mapping to increase the variety of the compression dictionary. Each curve in the set was divided into 4 × 4 square block and transformed into disk and then rotated, after that it was retransformed into 4 × 4 square block to form the dictionary used for coding. Thus a static and universal dictionary was attained. Meanwhile the number of geo- metrical transformation was largely extended and the problem that the image and digital dictionary should be corresponded with each other could be solved. The result shows that this algorithm is simply, effective and can gain fine compression effect, good decoded image and fast decoding speed.
出处
《计算机应用研究》
CSCD
北大核心
2007年第10期188-189,205,共3页
Application Research of Computers
基金
国家自然科学基金资助项目(69973033)
浙江省自然科学基金资助项目(199046)
