基于网格搜索算法的PCNN模型参数自适应  

Self-adaption of model parameters of PCNN based on grid search method

作　　者：李瀚[1] 杨晓峰 邓红霞[1] 常莎[1] 李海芳[1] 

机构地区：[1]太原理工大学计算机科学与技术学院,山西晋中030600 [2]山西建筑职业技术学院计算机工程系,山西晋中030600

出　　处：《计算机工程与设计》2017年第1期192-197,共6页Computer Engineering and Design

基　　金：山西省自然科学(青年科技研究)基金项目(2014021022-5);国家自然科学基金项目(61472270)

摘　　要：为提高利用脉冲耦合神经网络(pulse coupled neural network,PCNN)模型进行人脸识别时的准确率,并解决利用PCNN进行人脸识别时,模型中多个参数需凭经验设定的问题,提出一种基于脉冲发放强度的PCNN(QD-PCNN)模型和改进的网格搜索算法。QD-PCNN模型在简化的PCNN模型基础上,引入脉冲发放强度,细化模型的输出。改进的网格搜索算法在进行参数寻优时,根据识别对象,在较大范围内搜索,在得到的寻优结果附近区域进行精确搜索。在实验中,将通过改进的网格搜索法得到的参数组合运用到QD-PCNN模型中进行人脸识别,实验结果验证了该方法的有效性。To improve the accuracy of face recognition using pulse coupled neural network （PCNN） model and solve the problem that the parameters of PCNN model must be set with experience, the PCNN model based on pulse intensity （QD-PCNN） and the improved grid search method were proposed. In the QI）-PCNN, the conception of pulse intensity was proposed, which made the outputs of the model more accurate. When the improved grid search method was used to find the suitable parameters, the parameters were searched in a large space, and the found parameters were searched precisely according to the objects to be recognized. In the experimental process, the parameters obtained through improved grid search method were applied to QD-PCNN model to recognize faces. Results show the efficiency of this method.

关 键 词：脉冲耦合神经网络 网格搜索法 参数寻优 人脸识别 脉冲发放强度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]