Date of Award

8-2016

Degree Type

Thesis

Degree Name

Master of Science (MS)

Department

Technology Leadership and Innovation

First Advisor

Stephen J. Elliott

Committee Chair

Stephen J. Elliott

Committee Member 1

Joel Ebarb

Committee Member 2

Richard M. Guest

Abstract

The focus of this research is the effects of scarring on face recognition. Face recognition is a common biometric modality implemented for access control operations such as customs and borders. The recent report from the Special Group on Issues Affecting Facial Recognition and Best Practices for their Mitigation highlighted scarring as one of the emerging challenges. The significance of this problem extends to the ISO/IEC and national agencies are researching to enhance their intelligence capabilities. Data was collected on face images with and without scars, using theatrical special effects to simulate scarring on the face and also from subjects that have developed scarring within their lifetime. A total of 60 subjects participated in this data collection, 30 without scarring of any kind and 30 with preexisting scars. Controlled data on scarring is problematic for face recognition research as scarring has various manifestations among individuals, yet is universal in that all individuals will manifest some degree of scarring. Effect analysis was done with controlled scarring to observe the factor alone, and wild scarring that is encountered during operations for realistic contextualization. Two environments were included in this study, a controlled studio that represented an ideal face capture setting and a mock border control booth simulating an operational use case.

Share

COinS