云环境中考虑隐私保护的人脸图像识别

针对目前使用计算机进行人脸识别往往效率较差,且未考虑到隐私保护等问题,提出一种新的云环境中考虑隐私保护的人脸图像识别方法.对人脸图像进行简化处理以保护主要信息,并建立人脸图像主成分分析(PCA)数学模型,采用局部二值模式(LBP)方法获取人脸图像纹理特征,采用保局投影(LPP)方法选择人脸特征,并选取最常见的分数法对选择后的特征进行稳定性度量,引入深度网络法进行人脸图像的识别.结果表明,所提出的改进方法在保证人员隐私的基础上,具有较高的识别精度与识别效率.

Face image recognition based on privacy protection in cloud environment

Aiming at the problem that when the computer is used for the face recognition at present, the efficiency is usually poor and the privacy protection is not considered, a new face image recognition method with considering the privacy protection in the cloud environment was proposed. The simplifying treatment for the face image was performed to protect the main information, and the principal component analysis(PCA)mathematical model for the face image was established. In addition, the local binary pattern(LBP)method was used to get the texture features of face image, and the locality preserving projection(LPP)method was adopted for the feature selection of face image. Through selecting the most common fraction method, the stability measurement of feature after the selection was conducted, and the deep network method was introduced to perform the identification of face image. The results show that the proposed method has higher recognition accuracy and recognition efficiency on the basis of ensuring the personnel privacy.