Video analysis device, server, system and method for protecting identity privacy

Abstract

The invention discloses a video analysis device, a server, a system and a method for protecting identity privacy, and belongs to the technical field of privacy protection. The method comprises the steps that parameters of a selected pooling layer and a previous layer of a first convolutional neural network are fixed, and a second convolutional neural network is trained, so that the difference between a predicted identity and a real identity is as small as possible; adjusting parameters of a selected pooling layer and a previous layer of the first convolutional neural network, so that the difference between the predicted identity and the real identity is as large as possible; the first convolutional neural network after parameter adjustment is trained, so that the difference between a prediction result and a real result is as small as possible; and judging whether the difference value of the first convolutional neural network is smaller than a threshold value and the difference value ofthe second convolutional neural network is greater than the threshold value at the same time, If yes, ending. According to the method, the privacy network is introduced, an original neural network structure is not greatly modified, and dynamic balance is allowed according to the requirements for privacy and practicability while user data privacy is guaranteed.

Description

technical field [0001] The invention belongs to the technical field of privacy protection, and more specifically relates to a video analysis device, server, system and method for protecting identity privacy. Background technique [0002] With the accumulation of large-scale data, the development of the basic theory of deep learning and the huge improvement of hardware computing power, deep learning has become the mainstream technology in the field of data mining and artificial intelligence, especially in the field of image and voice, both in academia and industry. , text and other complex data. [0003] Video analysis technologies based on deep learning, such as intelligent fall detection based on deep learning, intelligent evaluation of Parkinson's disease motor function, abnormal behavior recognition, etc., generally adopt the following processing procedures: (1) use image collection equipment to collect image data; (2) Send the image data to the cloud server; (3) The ser...

AI Technical Summary

Problems solved by technology

However, the existing privacy protection schemes based on data encoding have the following outstanding problems: 1. There is a potential privacy leakage risk: the existing protection scheme does not consider the potential privacy leakage risk of the encoded data, that is, deep data mining on the encoded data may expose Privacy information; 2. Ignoring context information: when encoding data, the specific downstream data mining tasks are not considered. This scheme may obtain a more general privacy protection encoding, but it is not optimal in specific scenarios; 3. High complexity : In order to obtain high-quality privacy-preserving coding, too much complexity is introduced, such as the use of more complex network structures and training methods

Images