Privacy-protected outsourcing image feature extraction and classification method

In order to solve the problem that some existing outsourcing feature extraction methods cannot prevent facial information leakage or face recognition operation is very time-consuming after face encryption, the invention provides an outsourcing calculation method for privacy protection of image feature extraction and feature classification by using a cloud. The RGB image is converted to a YCbCr color space at a client, and pixel value matrixes of R, B, Cb and Cr color channels of the RGB image are obtained; 2, encrypting the matrix in the step 1 by using a one-time pad type key, and sending the encrypted matrix to a cloud end, and 3, extracting the encrypted matrix in the step 2 by the cloud end, inputting the matrix into a neural network or an integrated classifier model for training, and obtaining a result; and step 4, the cloud returns a classification result to the client. 针对现有的一些外包特征提取方法无法防止面部信息泄露或加密面部后人脸识别操作非常耗时的问题，提出了一种利用云端进行图像特征提取及特征分类的隐私保护的外包计算方法，包括以下步骤：步骤一，在客户端将RGB图像转换到YCbCr色彩空间，获得其R、B、Cb、Cr色彩通道的像素值矩