基于DLBP、PCA和SVM算法的人脸识别

Face Recognition Based on DLBP, PCA and SVM Algorithms

为提高人脸图像的识别率,提出一种结合DLBP、PCA和SVM算法的人脸识别方法。首先对每幅人脸图像进行分块处理,对每一分块图像进行局部二值模式统计,选取其中出现频率较高的模式,将所有图像分块的LBP直方图衔接起来作为图像最终的纹理特征,然后应用主成分分析法(PCA)对所提取的纹理特征进行降维处理,最后使用支持向量机分类器来对图像进行分类识别。基于该方法构建实用、有效的人脸特征提取、选择和识别过程,并在ORL人脸数据库进行实验,结果表明,相较于之前的LBP+SVM、PCA+SVM、LBP+PCA+SVM等人脸识别算法,该方法能有效提高人脸图像的识别率。

In order to improve the recognition rate of face image, proposes a combination of DLBP, PCA and SVM face recognition algorithm. Firstly,each face image is divided into blocks, each block of the image of the local two value model selection statistics, which has higher frequency mode, LBP will block all image histogram link up as the image texture features, then uses principal component analysis（PCA） to reduce the dimension of the texture feature extraction, finally uses support vector machine classifier to classify and identify the image.This method is constructed based on the practical, effective facial feature extraction, selection and identification process, and carries out the experiment and in the ORL face database. The results show that compared to the previous LBP＋SVM, PCA＋SVM, LBP ＋PCA＋SVM face recognition algorithm, this method can effectively improve the recognition rate of face image.