基于改进SICA的人脸特征提取方法研究

独立分量分析是一种有效的人脸特征提取方法。考虑到人脸样本的对称性,本文采用对称独立分量分析的方法对人脸样本进行特征提取。为了提高独立分量分析法表征人脸特征空间的能力,本文采用遗传算法对特征空间进行选择优化,获得最优的人脸特征子集。仿真实验表明,本文提出方法的识别率明显的好于独立分量分析方法的识别率。

Independent Component Analysis(ICA) is presented as an efficient face feature extraction method. Considering of the symmetry of face samples,Symmetric Independent component analysis(SICA) is proposed to extract face feature.In order to improve the ability of ICA,a genetic algorithm is represented to select optimal independent components(ICs) to reconstruct new feature space in this paper.Experimental results show that the GA-SICA method performs better than ICA method.