Touch-Screen Mobile-Device Data Collection for Biometrics Studies

Post on 26-Feb-2016

36 views 0 download

Tags:

description

Touch-Screen Mobile-Device Data Collection for Biometrics Studies. W. Ciaurro, B. Major, D. Martinez, D. Panchal, G. Perez, M. Rana, R. Rana, R. Reyes, S. Rodriguez, R. Valdez, D. Zuluaga. Research Purpose. Assist in Data Collection Understand Biometric Security Improve Authentication System - PowerPoint PPT Presentation

Transcript of Touch-Screen Mobile-Device Data Collection for Biometrics Studies

Touch-Screen Mobile-Device

Data Collection for Biometrics Studies

W. Ciaurro, B. Major, D. Martinez, D. Panchal, G. Perez, M. Rana, R. Rana, R.

Reyes, S. Rodriguez, R. Valdez, D. Zuluaga

Research Purpose

• Assist in Data Collection• Understand Biometric Security• Improve Authentication System

o What Features are best at Authentication?o 2008 US Higher Education Opportunity Acto Mobile Device Authentication

Biometric Technology• Physical versus Behavioral characteristics

• Keystrokes biometrics

• Physiological biometricso Use a person’s physical attributeso Ex. fingerprint, face or iris recognition for

identification of user identity.

• Behavioral biometricso Use a person’s speech, writing, or keystrokes for

such verification.

Literature Review● Long-text-input keystroke studies at Pace

University.● Touchscreen input for continuous

authentication at University of Oxford.● Passive user authentication at Carnegie

Mellon University.

Project Details

• 10 college participants• 3 days of data gathering• Android phone, recorded

features and gestures;o Text inputso Number pad inputso Other inputs (scrolling, pinching, etc.)

Keyboard Recorded Features:

• Duration

• Transitions

• PressureGesture Recorded Features

• Locations of touch

• Major/Minor Axis of touch

• Touch Timing

• Gyroscope Recording

Project Details

HardwareA variety of Android devices were used to capture data, including:

• LG Nexus 5

• LG Nexus 4

• Sony Xperia Z1Pros of Android:

• Ease of Data Capture (lots of sensors)

• Open Source (Java platform)

•Android Development Environments: 1. Eclipse

2. NetBeans • Both Java IDEs

● Data Collection Apps:○ Biometric Soft Keyboard○ Text Data Entry & Numeric Data Entry○ Biometric Gestures Capture

Software

ExperimentsScenario 1 - Numeric Input:

The numeric data samples were collected using the "Numeric Data Entry" app which consisted in entering the numeric sequence (9141937761) followed by a enter key.

Scenario 2 - Text Input: The character data samples were collected using the "Text Data Entry" app which involved entering 2 sets of characters similar to composing a large text message.

Experiments continued

Scenario 3 - Gestures:The touch gestures data samples were collected using the biometric gesture app. The 1st phase was to zoom in and out to locate an object within a picture. The 2nd phase was to zoom in on a question click on the answer them zoom out and go to the next questions.

One of our advantages was that we had no restrictions. Since this is a behavioral study each subject was unique to how they enter their keystroke.

Literature Construction

• Learned the proper means of writing about research.

• Developed real world experience.• Learned the challenges of working in groups

on a schedule

Conclusion• We were able to take part in a graduate study.

• Deeper understanding of the experiment process.

• Chance to develop professional research publication.

• We have hopes of implementing a Biometric System in the future.

Special ThanksWe would like to thank Gonzalo Perez,our

project manager, Andreea Cotoranu and Charles Tappert, our research advisors, and

all PACE Graduate Students involved!.

Thank You! Q&A