摘要
该文提出了一种用神经网络算法来设计二维线性相位数字滤波器的新方法。通过分析二维FIR 线性相位滤波器的幅频响应特性,建立了神经网络算法。根据给定的幅频响应指标,按该算法可获得滤波器系数。为保证该算法的稳定性,提出并证明了该算法的收敛定理。文中给出了圆对称和矩形对称二维低通线性相位FIR 数字滤波器优化设计实例。计算机仿真结果表明由该方法设计的二维数字滤波器,通带和阻带范围波动小,所需计算量非常少,稳定性强,因而是一种优异的设计方法。
This paper provides a new design approach based on a Neural Networks Algorithm（NNA）. According to the amplitude-frequency response characteristics of 2-D FIR linear-phase filters ,the NNA is established .Using the NNA,the designed filter coefficients can be obtained from the specified amplitude-frequency responses.To ensure stability of the NNA, the convergence theorem of the NNA is presented and proved. Two examples including circularly-symmetric and quadrately-symmetric 2-D lowpass linear-phase FIR filtsrs are also given to illustrate the effectiveness of the NNA-based design approach,and the results show that the ripple is considerably small in passband and in stopband,and the NNA-based method is of strong stability and requires significantly little amount of computations.Therefore,the optimal design approach is effective and excellent in the design field of 2-D linear phase FIR digital filters.
出处
《电子与信息学报》
EI
CSCD
北大核心
2005年第11期1755-1759,共5页
Journal of Electronics & Information Technology
基金
国家自然科学基金(50277010)高等学校博士学科点专项科研基金(20020532016)
关键词
二维数字滤波器
线性相位
神经网络
收敛定理
优化设计
2-D digital filters, Linear-phase, Neural network, Convergence theorem, Optimal design
