Date of Award

5-2018

Degree Type

Thesis

Degree Name

Master of Science (MS)

Department

Technology Leadership and Innovation

Committee Chair

Stephen J. Elliott

Committee Member 1

Matthias J. Sutton

Committee Member 2

Matthew P. Stephens

Committee Member 3

Matthew R. Young

Abstract

In this thesis, the categorization of Human-Biometric Sensor Interaction (HBSI) metrics, which include errors resulting from human interaction with a biometric device, using a wearable fitness tracker in a variety of biometric use cases is illustrated with a proposed methodology. This answers the question “Is it possible for the stream of data from the sensors in a wearable wristband be classified using the HBSI model?”. The use cases examined are gait recognition, heart rate RR recognition, and liveness detection using heart rate (bpm). The study was conducted on a convenience sample of three subjects. The proposed methodology could classify some subject behaviors with HBSI errors but not all of them. HBSI errors included were successfully processed sample, failure to detect, false interaction, and defective interaction. HBSI errors are used to classify subject behaviors with biometric systems and can be used to help improve the performance of recognition systems by ensuring correct presentations are used. This contribution can give researchers an idea of how to implement best practices when collecting data with wrist-based wearable devices.

Download

Share

COinS