Team for Research in
Ubiquitous Secure Technology

Citation
Li Zhuang, Feng Zhou. "Keyboard Acoustic Emanations Revisted". Talk or presentation, 27, April, 2006; Poster given at Trust NSF Site Visit.

Abstract
I propose a new approach for analyzing keyboard acoustic emanations. Emanations produced by typing on a keyboard can be used as a source of attacking. I present a novel attack taking as input a couple minutes sound recording of a user typing English text using a keyboard and recover typed characters. There is no need for a labeled training recording. A recognizer bootstrapped this way can even recognize random text such as passwords. My attack uses the statistical constraints of the underlying content, English language, to reconstruct text from sound recordings without any labeled training data. The attack uses a combination of standard machine learning and speech recognition techniques, including cepstrum features, Hidden Markov Models, linear classification, and feedback-based incremental learning. Based experiences of recovering text from keyboard acoustic emanations, I propose investigating a systematic way to compare information leaking from typing on keyboards and different attacking techniques. Further more, I propose studying the defenses in different practical scenarios.

Electronic downloads

• zhuang_KeyboardAcousticEmanations_TrustApril06Poster.ppt · application/vnd.ms-powerpoint · 318 kbytes

• HTML

  Li Zhuang, Feng Zhou. <a
  href="http://www.truststc.org/pubs/92.html"
  ><i>Keyboard Acoustic Emanations
  Revisted</i></a>, Talk or presentation,  27,
  April, 2006; Poster given at Trust NSF Site Visit.

• Plain text

  Li Zhuang, Feng Zhou. "Keyboard Acoustic Emanations
  Revisted". Talk or presentation,  27, April, 2006;
  Poster given at Trust NSF Site Visit.

• BibTeX

  @presentation{ZhuangZhou06_KeyboardAcousticEmanationsRevisted,
      author = {Li Zhuang, Feng Zhou},
      title = {Keyboard Acoustic Emanations Revisted},
      day = {27},
      month = {April},
      year = {2006},
      note = {Poster given at Trust NSF Site Visit.},
      abstract = {I propose a new approach for analyzing keyboard
                acoustic emanations. Emanations produced by typing
                on a keyboard can be used as a source of
                attacking. I present a novel attack taking as
                input a couple minutes sound recording of a user
                typing English text using a keyboard and recover
                typed characters. There is no need for a labeled
                training recording. A recognizer bootstrapped this
                way can even recognize random text such as
                passwords. My attack uses the statistical
                constraints of the underlying content, English
                language, to reconstruct text from sound
                recordings without any labeled training data. The
                attack uses a combination of standard machine
                learning and speech recognition techniques,
                including cepstrum features, Hidden Markov Models,
                linear classification, and feedback-based
                incremental learning. Based experiences of
                recovering text from keyboard acoustic emanations,
                I propose investigating a systematic way to
                compare information leaking from typing on
                keyboards and different attacking techniques.
                Further more, I propose studying the defenses in
                different practical scenarios.},
      URL = {http://www.truststc.org/pubs/92.html}
  }
  

Posted by Christopher Brooks on 4 May 2006.
Groups: trust
For additional information, see the Publications FAQ or contact webmaster at www truststc org.