基于独立成分分析和核向量机的人脸识别

Face Recognition Based on Independent Component Analysis and Core Vector Machines

提出利用独立成分分析提取人脸特征并用核向量机进行识别的方法。独立成分分析能更本质地描述图像特征,通过选择合适的特征个数达到较高的识别准确率。利用核向量机进行分类判决,可以快速地对大样本数据进行准确分类,产生较少的支持向量。实验证明了该方法的可行性和有效性,在ORL人脸数据库上达到了94.38%的准确率。

This paper proposes an algorithm which adopts Independent Component Analysis（1CA） to extract face feature and Core Vector Machines（CVM） to recognize. ICA is used to extract statistical independent feature and a good result can be received by selecting right feature numbers. CVM is used to classify the feature and it can handle large data sets more quickly. Experimental results show that the algorithm is feasible and effective for face recognition, and ils accuracy is 94.38% on ORL.