Title :
Design of 2D digital filters using BP neural networks
Author :
Wang, Xiao-Hua ; He, Yi-Gang ; Li, Hong ; Peng, Yu-Lou ; Li, Mao-Jun
Author_Institution :
Electri. & Inf. Eng. Coll., Changsha Univ. of Sci. & Technol., China
Abstract :
This paper provides a new approach for the design of 2D FIR linear-phase digital filters based on a parallel back-propagation neural-networks algorithm (PBPNNA). According to the frequency response characteristic of a 2D linear-phase FIR filter, a compact expression for the transfer function is derived, and a new PBPNNA is established for designing 2D FIR linear-phase filters. The convergence theorem is presented and proved to illustrate the stability of the PBPNNA. Design examples are also given to illustrate the effectiveness of the proposed design approach.
Keywords :
FIR filters; backpropagation; convergence; electronic engineering computing; neural nets; transfer functions; 2D digital filter design; 2D linear-phase FIR filter; BP neural network; convergence theorem; frequency response characteristics; parallel back-propagation; transfer function; Algorithm design and analysis; Convergence; Design engineering; Digital filters; Finite impulse response filter; Neural networks; Nonlinear filters; Signal processing algorithms; Stability; Transfer functions; 2-D Linear-phase digital filters; back-propagation neural network; convergence theorem; stability;
Conference_Titel :
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location :
Guangzhou, China
Print_ISBN :
0-7803-9091-1
DOI :
10.1109/ICMLC.2005.1527765
