Key point differential privacy-driven face image privacy protection method

Abstract

The invention discloses a key point differential privacy-driven face image privacy protection method, which comprises the following steps of: firstly, carrying out data preprocessing, and constructing a key point identification anonymous space; then constructing an anonymous face generative adversarial network structure, and determining an anonymous face generative adversarial network optimization objective function; obtaining an improved differential privacy algorithm applied to anonymous face key point identification; and finally, training the anonymous face generative adversarial network through the preprocessed data set, and outputting a final result. According to the method, the face key point structure of the face in the image is modified to realize face identity anonymity, better data availability and visual effect are obtained, the quality of the generated image is higher, the original non-identity attribute can be kept, and any attribute tag is not needed.

Description

technical field [0001] The invention belongs to the field of image privacy protection, and in particular relates to a face image privacy protection method driven by key point differential privacy. With the rapid development of the big data era, face data is one of the most easily leaked personal data, and its privacy security is seriously threatened. The present invention proposes a face image privacy protection technology driven by face key points. By using the differential privacy theory to anonymize the face key point identification, changing its facial geometric structure to realize the anonymization of face images to protect Identity Information. Background technique [0002] With the explosive growth of diversified Internet services and social network platforms, large-scale users interact with massive amounts of information on social network platforms, resulting in a huge amount of shared data mainly based on visual media in social networks. With the explosive growth...

AI Technical Summary

Problems solved by technology

[0004] All the problems and dilemmas mentioned above can be boiled down to one question: given an image of a face, how can we create another image with a similar appearance and the same background, while the real identity is hidden, and the face detector can still Work? Traditional anonymization techniques are mainly based on obfuscation and always significantly change the original face image

Most of its current deep learning methods generate images that are not realistic enough or use other people's ids to fuse with their own non-identity information to achieve anonymity, but there will be suspicion of violating other people's privacy

Images