Project Info
Vibration-based Keystroke Snooping
Dejun Yang | djyang@mines.edu
Keystroke snooping, i.e. inferring a user’s typed inputs, is a serious threat to user privacy. The adversary can learn user’s sensitive information such as usernames, passwords, credit card numbers, SSNs and confidential documents. The goal of this project is to show that one can exploit subtle vibration patterns to infer keystroke accurately and reliably. We also plan to design simple but effective defense techniques. Through this project, we also aim to raise awareness of such attacks.
More Information
K. Jin et al., “ViViSnoop: Someone is snooping your typing without seeing it!,” 2017 IEEE Conference on Communications and Network Security (CNS), Las Vegas, NV, 2017, pp. 1-9. (download at https://www.cs.purdue.edu/homes/chunyi/pubs/cns17-jin.pdf)
S. Pan et al., “SurfaceVibe: Vibration-Based Tap & Swipe Tracking on Ubiquitous Surfaces,” 2017 16th ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN), Pittsburgh, PA, 2017, pp. 197-208. (download at https://users.ece.cmu.edu/~shijiapa/documentations/Pan_IPSN_2017.pdf)
Grand Engineering Challenge: Secure cyberspace
Student Preparation
Qualifications
Self-motivated
Good programming skills
Knowing or be willing to learn signal processing techniques.
Not afraid of playing with simple hardware devices (e.g. https://www.sparkfun.com/products/11744) and mini computer boards, e.g. Arduino and Raspberry Pi.
Time Commitment
20-30 hours/month
Skills/Techniques Gained
Using geophones (e.g. https://www.sparkfun.com/products/11744) to collect vibration data
Basic signal processing techniques
Machine learning techniques
How to conduct research experiments
Mentoring Plan
1. Monthly or weekly meeting
2. Beside me, I will assign a grad student as a mentor as well.