Adversarial Activity Detection Using Keystroke Acoustics

Amin Fallahi, Vir V. Phoha

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Using keystroke acoustics to predict typed text has significant advantages, such as being recorded covertly from a distance and requiring no physical access to the computer system. Recently, some studies have been done on keystroke acoustics, however, to the best of our knowledge none have used them to predict adversarial activities, such as password dictionary attacks, data exfiltration, etc. We show that keystrokes in an adversarial environment have unique characteristics that distinguish it from benign environments and these differences can be used to predict adversarial activities and threat levels against a computer system. On a dataset of two million keystrokes consisting of seven adversarial and one benign activity, we use a signal processing approach to extract keystrokes from the audio and a clustering method to recover the typed letters followed by a text recovery module to regenerate the typed words. Furthermore, we use a neural network model to classify the benign and adversarial activities and achieve significant results: (1) we extract individual keystroke sounds from the raw audio with 91% accuracy and recover words from audio recordings in a noisy environment with 71% average top-10 accuracy. (2) We classify adversarial activities with 93.11% to 98.07% average accuracy under different operating scenarios.

Original languageEnglish (US)
Title of host publicationComputer Security – ESORICS 2021 - 26th European Symposium on Research in Computer Security, Proceedings
EditorsElisa Bertino, Haya Shulman, Michael Waidner
PublisherSpringer Science and Business Media Deutschland GmbH
Pages626-648
Number of pages23
ISBN (Print)9783030884178
DOIs
StatePublished - 2021
Event26th European Symposium on Research in Computer Security, ESORICS 2021 - Virtual, Online
Duration: Oct 4 2021Oct 8 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12972 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference26th European Symposium on Research in Computer Security, ESORICS 2021
CityVirtual, Online
Period10/4/2110/8/21

Keywords

  • Adversarial activity classification
  • Attack detection
  • Keystroke acoustics

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Adversarial Activity Detection Using Keystroke Acoustics'. Together they form a unique fingerprint.

Cite this