摘要
针对分形图像压缩中矩形划分计算量太大的问题,提出了一种混合分类方法并将其应用于图像的矩不变量,得到了一种基于矩形划分的快速分形编码方法。实验表明,该方法相对于全局搜索,在压缩比和解码质量略有下降的基础上,能极大地提高分形编码速度;与均匀分类方法相比,混合分类法可进一步提高分形编码速度并改善解码图像质量,可以在一定的条件下取得压缩比优势。
Aiming at the question that the computational complexity of HV partition is great in fractal image compressing, a hybrid classification method was proposed and it was used on mass center of image. Then a fast method for fractal coding were obtained based on HV partition. Experimental results indicate that in contrast with exhaustive search hybrid classification method can improve the speed of fractal coding much more on foundation of lower compression ratio and worse quality of decoded image. In contrast with original uniform classification, hybrid classification method can improve the speed of fractal coding and the quality of decoded image, moreover in some condition it can achieve higher compression ratio.
出处
《计算机应用研究》
CSCD
北大核心
2007年第7期64-66,共3页
Application Research of Computers
基金
国家自然科学基金资助项目(10071060)
西北工业大学研究生创业种子基金资助项目(Z200656)
关键词
分形图像压缩
迭代函数系统
矩形划分
自适应分类
混合分类
fractal image compression
iterated function system
HV partition
adaptive classification
hybrid classification