WristSnoop: Smartphone PINs prediction using smartwatch motion sensors
Authors:
- 1 November 2015
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Abstract
Smartwatches, with motion sensors, are becoming a common utility for users. With the increasing popularity of practical wearable computers, and in particular smartwatches, the security risks linked with sensors on board these devices have yet to be fully explored. Recent research literature has demonstrated the capability of using a smartphone's own accelerometer and gyroscope to infer tap locations; this paper expands on this work to demonstrate a method for inferring smartphone PINs through the analysis of smartwatch motion sensors. This study determines the feasibility and accuracy of inferring user keystrokes on a smartphone through a smartwatch worn by the user. Specifically, we show that with malware accessing only the smartwatch's motion sensors, it is possible to recognize user activity and specific numeric keypad entries. In a controlled scenario, we achieve results no less than 41% and up to 92% accurate for PIN prediction within 5 guesses.Keywords
This publication has 8 references indexed in Scilit:
- MoLe2015
- RisQ2014
- TaintDroidACM Transactions on Computer Systems, 2014
- Practicality of accelerometer side channels on smartphones2012
- TapLogger2012
- ACCessory2012
- (sp)iPhone2011
- Compromising Reflections-or-How to Read LCD Monitors around the Corner2008 IEEE Symposium on Security and Privacy (SP 2008), 2008