Image privacy protection method and system based on generative adversarial network

Abstract

The invention provides an image privacy protection method and system based on a generative adversarial network. The method comprises the following steps: acquiring an original picture data set containing user privacy data; taking an original picture data set as a training data set, and performing joint training on the nonlinear mapping module, the generator module and the discriminator module; differential privacy is added in the joint training process of the nonlinear mapping module, the generator module and the discriminator module, so that the privacy of the original picture can be protected from being leaked by the picture generated by the generator; and for the trained nonlinear mapping module and generator module, continuously changing a random input variable of the nonlinear mapping module and a noise characteristic variable of the generator module to obtain an expected number of privacy protection picture data sets. The image privacy protection method and system based on the generative adversarial network are high in privacy protection, high in availability after image privacy processing and simple in image privacy protection.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to an image privacy protection method and system based on a generative confrontation network. Background technique [0002] The rapid development of machine learning makes it one of the most effective tools in the field of artificial intelligence, but the training process of the algorithm often requires a large amount of user data, which brings great risks of privacy leakage to users. With the vigorous promotion of Internet technology, the information available for remote monitoring and acquisition is constantly increasing. In addition to commercial buildings and public places, video surveillance equipment in private homes is also widely used in surveillance, entertainment and communication. These videos and photos usually contain easily identifiable face images, and the face area in the image can represent personal identity information , so it is a seriousl...

AI Technical Summary

Problems solved by technology

In order to prevent the leakage of face identity information in this method, the usual specific method is to modify the face area in the original image, resulting in low recognition accuracy of the face recognition model, thereby protecting identity privacy

At present, the most commonly used deanonymization algorithm is the K-Same series algorithm, which will cause artifacts in private images

There is also a method of combining the active appearance model with the K-Same algorithm. By minimizing the cost function of attribute retention and de-identification, the face is difficult to recognize after fusion, thereby protecting the privacy of face attributes. The privacy image generated by the method is not visually natural enough

Images