摘要
提出了一种基于遗传算法的协同神经网络中参数的优化算法 ,该算法利用遗传算法的全局最优搜索能力 ,在协同神经网络的参数空间搜索最优解 .对从“车牌识别系统”中采集得到的数字样本进行的测试表明 :优化算法能有效提高协同神经网络的识别性能 ,使识别率达到了较为实用的水平 (98.4% ) .另外 ,还对协同神经网络中各个参数在识别过程中的作用进行了讨论 .
An optimization algorithm based on generic algorithm for parameters of synergetic neural network was proposed, which searched the optimal solutions in the parameters space of synergetic neural network using the globally optimal searching ability of genetic algorithm. The test upon the numeral samples from 'Car License Recognition System' shows that the optimization algorithm is able to improve the recognition performance of synergetic neural network effectively, and make the recognition rate reach the level of real application (98.4%). Additionally, the functions of parameters of synergetic neural network in object recognition were discussed.
出处
《红外与毫米波学报》
SCIE
EI
CAS
CSCD
北大核心
2001年第3期215-218,共4页
Journal of Infrared and Millimeter Waves
基金
国家自然科学基金 (编号 6 9772 0 0 2 )资助项目&&
关键词
协同神经网络
神经网络优化
遗传算法
目标识别
synergetic neural network
optimization of neural network
genetic algorithm
target recognition