摘要
在LSVM算法的基础上提出了利用LASVM算法来进行人脸识别的方法。首先利用MEGM度量学习算法转换CAS—PEAL—R1人脸共享数据库的样本数据和测试数据的特征空间，然后采用LSVM算法程序训练测试转换后的人脸数据并进行分类识别，最高准确率可达到98％。实验表明：利用LASVM算法进行人脸识别，不但识别速度快而且识别效率也高，具有较高的可行性。
On the basis of LASVM algorithm, a method of using LASVM algorithm to distinguish human faces is proposed. MEGM measure learning algorithm is firstly applied by LASVM algorithm to transfer characteristic space of training sample data and testing data in CAS-PEAL-RI face sharing database, and then, LSVM algorithm program is used for training the tested and converted face data and further categorically recognizing these data, the highest accuracy rate can reach to 98%. The experiment indicates that the face recognition method by making use of LASVM algorithm is of high speed, high efficiency and higher feasibility.
出处
《自动化信息》
2011年第10期36-38,共3页
Automation Information
