摘要
特征提取和分类器设计是人脸识别算法中的两个关键问题。提出一种基于二次小波变换、PCA算法与BP神经网络的人脸识别算法。该算法采用二次小波变换与PCA相结合的算法提取人脸图像的主要特征,并运用加入动量项的改进BP神经网络算法进行人脸图像分类识别。在MATLAB环境下,利用ORL人脸图像数据库进行了仿真实验,实验结果表明,该算法实现简单、识别速度快、识别率较高。
Feature extraction and classifier design are two key techniques of face recognition. A face recognition method of the combination of second wavelet transform and PCA algorithm based on improved BP neural network is proposed. The algorithm uses second wavelet transform and PCA algorithm to extract principal features of face image and adds into the momentum item to improve the BP algorithm for face image classification and recognition. The simulation experiment results on ORL face in MATLAB environment database demonstrate that the realization is simple,the algorithm has fast recognition speed and high recognition rate.
出处
《信息技术》
2015年第6期8-11,共4页
Information Technology
基金
国家自然科学基金(41120002/D020101)
