You Are Not Acting Like Yourself: A Study on Soft Biometric Classification, Person Identification, and Mobile Device Use

In this paper, we explore the soft biometric classification of 13 demographic and behavioral attributes using phone data collected from 46 subjects. We utilize the results of this analysis to further evaluate reduced search spaces for the primary identification task by implementing a ranking formula...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on biometrics, behavior, and identity science behavior, and identity science, 2019-04, Vol.1 (2), p.109-122
Main Authors: Neal, Tempestt J., Woodard, Damon L.
Format: Article
Language:English
Subjects:
Authentication
Biometrics
Biometrics (access control)
Bluetooth
Classification
Demographics
Feature extraction
identification
Mobile communication systems
Mobile handsets
mobile security
soft biometrics
Task analysis
Wireless fidelity
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we explore the soft biometric classification of 13 demographic and behavioral attributes using phone data collected from 46 subjects. We utilize the results of this analysis to further evaluate reduced search spaces for the primary identification task by implementing a ranking formula which quantifies the performance of each attribute considering the predictability of some attributes compared to others. Results show that soft biometric classification is feasible with up to 90% accuracy; however, because people exhibit high intra-class variance, templates and queries are significantly affected in terms of how well they match, even with reduced search spaces. We analyze these findings using a combination of approaches, including an evaluation of the biometric menagerie, visualizing the distribution of the data, and observing how subjects vary in their soft biometric class across different times of the day.
ISSN:2637-6407
2637-6407
DOI:10.1109/TBIOM.2019.2905868