Publication Details

The constant use of smartphones in all kinds of situations has resulted in many formerly private activities becoming public. In the course of this, smartphone users gradually give up more and more of their privacy - often without realizing it. Photo browsing on the mobile phone is one example. While users scroll through their personal galleries, an attacker or a curious friend can easily view many photographs, which often contain sensitive, private information.

A particularity of human perception could be leveraged to solve this problem: the ability to recognize familiar images, even when they are distorted - for instance by a graphical filter. This technical report investigates filters which allow to distord photographs in a way that prevents attackers from understanding the content of these photographs, while the owner can still recognize them.

In this report, suitable graphical distortion filters were researched and evaluated, within a prototype gallery, during a user study with 24 participants.The results indicate the feasibility of the concept and show promising tendencies for three specific image filter.