摘要
本文在分形函数逼近系统中引入基数B样条小波尺度函数 ,提出基于小波的分形函数逼近算法 ,该算法可充分利用尺度函数和分形函数在多尺度和多分辨性上的一致性。逼近参数的求解可转化为凸二次规划问题。实验结果表明 ,这种新算法的逼近效果明显优于基于多项式的分形函数逼近 ,且在数据压缩。
This paper introduces the B spline wavelet scaling function into fractal function approximation system and proposes a fractal function approximation algorithm based on wavelets. The algorithm fully exploits the consistency between the fractal and scaling functions in multi scale and multi resolution . Parameters of the approximation can be determined by solving a convex quadratic programming problem. Experimental results show that the degree of approximation of this new algorithm has obviously outperformed that of polynomial-based ones and has high potential in data compression, especially in image compression.
出处
《兵工学报》
EI
CAS
CSCD
北大核心
1999年第4期610-615,共6页
Acta Armamentarii
关键词
小波尺度函数
分形函数
逼近
信号处理
压缩编码
B-spline, wavelet scaling function, fractal function,quadratic programming, approximation
