摘要
为解决离散过程神经网络的训练问题,提出了两种基于数值积分的离散过程神经网络训练算法.分别采用三次样条积分和抛物插值积分直接处理隐层离散样本和权值的时域聚合运算,输出层采用普通神经元,采用L-M(Levenberg-Marquard)算法实现网络参数的调整.以模糊图像的恢复为例,实验结果表明,两种训练方法的性能比较接近,但都优于基于沃尔什变换的离散过程神经网络和基于样条差值函数的离散过程神经网络,从而揭示出数值积分方法在提升离散过程神经网络性能和应用方面具有一定潜力。
To address the training problem of discrete process neural networks, two training algorithms based on numeri- cal integration were proposed. The cubic spline integration and the parabolic interpolation integration were used in the hid- den layer to deal with the time-domain aggregation of discrete samples and weights. The classical neurons were used in out- put layer. In order to improve the convergence ability of the network, the Levenberg-Marquard algorithm was employed to ad- just the networks＇ parameters. The effeeiveness of the proposed algorithms was testified by applying the network to the resto- ration of a fuzzy image. Experimental results show that the performance of the two algorithms is relatively close and is superior to the Walsh transformation-based discrete process neural networks and spline function-based diserete process neural networks in both approximation ability and effect of image restoration, which reveals that the proposed methods of numerical integration have some potential in the performance improvement and application extension of discrete process neural networks.
出处
《信号处理》
CSCD
北大核心
2013年第9期1182-1189,共8页
Journal of Signal Processing
基金
国家自然科学基金(61170132)
关键词
离散过程神经网络
数值积分
三次样条积分
抛物插值积分
图像恢复
discrete process neural networks
numerical integration
cubic spline integration
parabolic interpolation in-tegration
image restoration
