论文部分内容阅读
以子波滤波器组在矩阵稀疏化中的应用为背景，给出了设计优化（两带）子波滤波器组需要满足的一般性条件和优化子波滤波器组的两种设计方法，它们均是简单的线性规划问题．计算了误差对子波滤波器正交性的影响，给出了一个修正的优化算法，使数值精确度有明显提高．仿真结果表明，优化子波滤波器组在矩阵稀疏化和去相关性方面优于Ｄａｕｂｅｃｈｉｅｓ子波滤波器组．
Based on the application of wavelet filter banks in matrix thinning, two general design conditions for optimizing (two-band) wavelet filter banks and two design methods for optimizing the wavelet filter banks are given. All are simple linear programming problems. The effect of error on the orthogonality of the wavelet filter is calculated, and a modified optimization algorithm is given to improve the numerical accuracy significantly. The simulation results show that the optimized wavelet filter bank is better than the Daubechies wavelet filter bank in matrix sparsification and decorrelation.