基于神经网络的货币识别研究  

Reseach on the Paper Currency Recognition Based on Neural Network Algorithm

作　　者：石文渊[1] 彭军民[1] 

机构地区：[1]西安理工大学自动化与信息工程学院,陕西西安710048

出　　处：《西安理工大学学报》2008年第2期201-205,共5页Journal of Xi'an University of Technology

摘　　要：货币识别研究的重点在于识别的准确和速度。采用结构相对简单的前馈神经网络,选取合适的货币采样特征为网络的输入特征,使用尺度共轭梯度(SCG)算法对网络进行快速高效训练,实现对货币的快速、准确识别。实验结果表明,该方案是可行、有效的。The priority given to research on paper currency recognition lies in the accuracy and velocity of recognition. In this paper, the authors adopt the feed -forward neural network with a comparatively easy structure, select the suitable paper currency sampling features to be the network inputting characteristics, and use Scaled Conjugate Gradient （SCG） algorithm to conduct the rapid and efficient training for the network, whereby the swift and accurate recognition of paper currency can be realized. The result of the experiment demonstrates that the proposal is feasible and effective.

关 键 词：前馈神经网络 尺度共轭梯度算法 货币识别 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]