Date of Award

7-2021

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Computer Engineering and Sciences

First Advisor

Michael King

Second Advisor

Kevin Bowyer

Third Advisor

Vanessa Edkins

Fourth Advisor

Philip Bernhard

Abstract

The gap of accuracy observed in some commercial face analytic systems based on race and gender raised questions about the equity and fairness of those systems. Since these systems are part of several applications today, some more critical than others, it urges designers to detect and mitigate any sources of bias. In this thesis, we begin by clarifying the confusion between face analytic, face recognition, and face processing systems. Then, we analyze gender classification accuracy using two datasets and three classifiers. The Pilot Parliaments Benchmark dataset is examined with an open-source algorithm to corroborate the gender shade. Secondly, the Morph dataset is employed to investigate the relationship between gender classification and face recognition as it is also suitable for face matching. Finally, we analyze the role of a person’s skin in gender classification accuracy by correlating misclassified with false match pairs resulting from face match comparisons. We contribute to knowledge by providing evidence on the non-effect of gender classification on the face matching outcomes and providing the first investigation work on the skin tone-driven factor on the face processing results using an automated skin tone rating algorithm.

Comments

Copyright held by author

Download

Share

COinS