Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Predicting Human Performance for Face Recognition1 Alice J. O'Toole Fang Jiang
 

Summary: Predicting Human Performance for Face Recognition1
Alice J. O'Toole Fang Jiang
The University of Texas at Dallas The University of Texas at Dallas
Email: otoole@utdallas.edu Email: fxj018100@utdallas.edu
Dana Roark Hervé Abdi
The University of Texas at Dallas The University of Texas at Dallas
Email: danar@utdallas.edu Email: herve@utdallas.edu
Abstract
The ability of humans to recognize faces provides an implicit benchmark for gauging
the performance of automatic face recognition algorithms. In this chapter we review the
factors that affect human accuracy. These factors can be classified into facial stucture
constraints and viewing parameters. The former include factors such as face typicality,
gender, and ethnicity. The latter include changes in illumination and viewpoint, as well
as the perceptual complications introduced when we see faces and people in motion. The
common thread of the chapter is that human experience and familiarity with faces can
overcome many, if not all, of these challenges to face recognition. A goal of computional
algorithms should be to emulate the ways in which humans acquire familiarity with
faces. It may then be possible to apply these principles to the design of algorithms to
meet the pressing challenges of face recognition in naturalistic vieiwng conditions.
9.1 Introduction

  

Source: Abdi, Hervé - School of Behavioral and Brain Sciences, University of Texas at Dallas

 

Collections: Biology and Medicine; Computer Technologies and Information Sciences