The Security Analysis and Enhancement of Photographic Authentication

Abstract

The aims of this project were to analyze the security of photographic authentication (PA) systematically, show that PA is vulnerable under the polling attack, and give some suggestions to enhance its security. To achieve the above goals, an automatic attack tool is designed to analysis the security of PA systematically. The tool captures the displayed photos, matches with historical ones to accumulate their counts. It selects the photo with highest count and repeats the process until successful login. In order to interfere with the photo match of the attack tool, a noise displacement method is also used to add noises into the original photos. Correspondingly, two noise reduction techniques are implemented in the attack tool for security analysis of PA with noise displacement methods. Furthermore, a simulation tool is designed to analysis the security of PA under a large number of photo sets. The security of PA is analyzed clearly from the experimental and simulation studies and enhancement ways of PA are also summarized in this study.