Towards Attack and Defense Views to K-Anonymous Using Information Theory Approach

Ensuring multimedia content security is one of the main targets to solve multimedia information security problems and k-anonymous model is a classic and effective model to protect personal information when we publish multimedia data. The k-anonymous security with different k values has always been a hot topic of research, and there is no specific quantitative method for k-anonymous model security without considering any application scenarios or assumptions. In this paper, we give a quantitative method for the capability of attacker and defender under k-anonymous model. More specifically, we present a new general notion of limitation to provide a precise bound of attack and defense capability in a k-anonymous model. Firstly, we establish the attack channel and the defense channel from the point of the attacker's and the defender's view respectively. Moreover, we describe the capability of both attacking and defending sides by introducing the average mutual information, as well as the maximum value of the the average mutual information is the limitation of the capability of both sides, i.e. the channel capacity. Finally, according to the variation curve of the average mutual information amount, we compare the capability of the attacker and defender and analyze the effect of k value on it and get the best interval of k, that is, the value of k should be greater than 8 and less than 40.