摘要
提出了一种基于多分类投影极速学习机的快速人脸识别方法.首先采用2DGabor小波提取所有人脸样本图像的人脸特征,然后将学习样本的人脸特征用于训练多分类投影向量机,最后将训练好的多分类投影极速学习机用于分类.采用CMU-PIE和ORL人脸数据库进行了对比实验,大量实验结果证实所提方法的识别正确率和速度均优于极速学习机和支持向量机方法.
A novel face recognition method based on Multi-class Projection Extreme Learning Machine（MPELM） is proposed in this paper.First,all face images are decomposed by using 2D Gabor.Then the face feature as study samples to MPELM classifier are applied.Finally,MPELM classifier are used for classification,then face recognition results will be obtained.The experiments are performed on the CMU-PIE and ORL face databases,and the experimental results indicate that the proposed method has better performance than ELM-based method and SVM-based method.
出处
《微电子学与计算机》
CSCD
北大核心
2012年第7期13-17,共5页
Microelectronics & Computer
基金
国家自然科学基金项目(60972146)
中国博士后科学基金(20110491661)
华为创新基金项目(IRP-2011-03-21)
关键词
极速学习机
人脸识别
GABOR小波
支持向量机
extreme learning machine
face recognition
Gabor wavelet
support vector machine
