LDA及支持向量机的人脸识别算法研究

在主成分分析的基础上采用线性差别分析法对人脸图像进行特征提取,构造人脸的特征脸空间,在特征脸空间运用线性差别分析法进行人脸识别。在支持向量机方法理论基础上,利用LibSVM分类器对处理后的人脸图像进行分类,考虑到核函数参数对分类结果的影响,通过参数寻优及算法的改进将多类问题的分类简单化,并大大提高识别效率,在ORL人脸库的识别结果表明,本方法在特征参数个数的选取、识别效果等方面都有其独到的优越性,具有很好的可行性和实际意义。

Reseach of the Face Recognition Algorithm with LDA and SVM

The Linear Discriminant Analysis is applied to the face image feature extraction based on the principal component analysis to creat the face eigenface space,then the LDA algorithms are used for face recognition.Based on the theory of support vector machines,using LibSVM classifier to classify the handled face images,taking into account the impact of kernel function parameters on the classification results,the multi-class classification problems is simplified by parameter optimization and algorithm improvements,the identification efficiency is greatly improved.The identification of the ORL face database shows that this method has its own unique superiority in selecting the characteristic parameters,recognition results and so on.It has a good feasibility and practical significance.