基于PCA模型的人脸图像纹理特征高精度识别方法

High-precision Recognition Method for Face Image Texture Features Based on PCA Model

针对传统人脸图像纹理特征识别方法中存在的计算量大,样本训练与测试时间较长,识别正确率较低等问题,提出一种新的基于PCA模型的人脸图像纹理特征高精度识别方法。在人脸图像预处理过程中,采用Gabor滤波器确定人脸图像训练样本中的双眼位置,结合卷积运算与人脸几何模型从图像中分割出目标人脸区域,并对分割得到的图像进行规范化处理;采用PCA模型对预处理后的图像进行降维与特征向量提取,并根据选取的主要纹理特征以及欧式距离近似度量结果,实现人脸图像纹理特征高精度识别。实验结果表明,所提方法的识别准确度高于实验对比方法,且样本训练时间与测试时间明显缩短,具有较好的鲁棒性。

Aiming at the problems existing in traditional face image texture feature recognition methods,such as large amount of computation,long training and testing time,low recognition accuracy,a new high-precision face image texture feature recognition method based on PCA model is proposed.In the process of face image preprocessing,Gabor filter is used to determine the binocular position of the training sample of face image,convolution operation and face geometry model are combined to segment the target face area from the image,and normalize the segmented image;PCA model is used to reduce the dimension of the preprocessed image and extract the feature vectors,and according to the selected main texture features.The feature and Euclidean distance approximation measure results are used to realize high-precision recognition of face image texture features.The experimental results show that the recognition accuracy of the proposed method is higher than that of the experimental comparison method,and the training time and testing time of the samples are significantly shortened,which has better robustness.