A Sound for a Sound: Mitigating Acoustic Side Channel Attacks on Password Keystrokes with Active Sounds

摘要

Keyboard acoustic side channel attacks have been shown to utilize the audio leakage from typing on the keyboard to infer the typed words up to a certain degree of accuracy. Researchers have continued to improve upon the accuracy of such attacks by employing different techniques and attack vectors such as feature extraction and classification, keyboard geometry and triangulation. While research is still ongoing towards further improving acoustic side channel attacks, much work has been lacking in building a working defense mechanism against such class of attacks. In this paper, we set out to propose a practical defense mechanism against keyboard acoustic attacks specifically on password typing and test its performance against several attack vectors. Our defense involves the use of various background sounds to mask the audio leakage from the keyboard thereby preventing the side channel attacks from gaining usable information about the typed password. The background sounds are generated by the device that is used to input the passwords. We also evaluate the usability of our approach and show that the addition of background sounds does not hamper users' capability to input passwords.