改进非负矩阵分解的神经网络人脸识别

发布时间：2018-12-14 20:12

【摘要】：为了提高人脸识别正确率,提出基于改进非负矩阵分解的神经网络人脸识别算法。首先利用改进的非负矩阵分解对人脸图像进行特征提取,提高非负矩阵分解速度。接着将提取出的特征信息作为神经网络学习入口进行特征训练,由于神经网络在学习过程中,容易出现局部最小值且收敛速度慢等问题,为此采用改进的遗传算法对神经网络进行优化处理,获得最终的人脸识别结果。实验结果表明:利用改进的非负矩阵分解方法能够降低神经网络的分类训练负荷量和运算量,提高人脸识别识别率。通过和各种方法比较可知,本方法的人脸识别率都较高。本方法人脸特征分解速度快,提高了神经网络训练前期精度和收敛速度,使得人脸识别正确率高。当特征向量个数达到40以上时,人脸识别正确率保持95%以上。
[Abstract]:In order to improve the accuracy of face recognition, a neural network face recognition algorithm based on improved nonnegative matrix decomposition is proposed. Firstly, the improved non-negative matrix decomposition is used to extract the features of the face image to improve the speed of the non-negative matrix decomposition. Then the extracted feature information is used as the learning entrance of neural network for feature training. Because of the problem of local minimum and slow convergence rate in the learning process of neural network, The improved genetic algorithm is used to optimize the neural network and the final face recognition results are obtained. The experimental results show that the improved nonnegative matrix decomposition method can reduce the training load and operation load of the neural network classification and improve the face recognition rate. Compared with other methods, the face recognition rate of this method is higher. The method improves the accuracy and convergence speed of neural network training and makes the face recognition accuracy high. When the number of feature vectors reaches more than 40, the correct rate of face recognition is over 95%.
【作者单位】：长春信息技术职业学院计算机系;长春理工大学计算机科学技术学院;
【基金】：吉林省科技发展计划项目(No.20130303011GX,No.20140204050GX)~~
【分类号】：TP391.41;TP183

【相似文献】