Accession Number:

ADA450917

Title:

Designing Computer Agents With Facial Personality To Improve Human-Machine Collaboration

Descriptive Note:

Master's thesis

Corporate Author:

AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OH

Personal Author(s):

Report Date:

2006-05-25

Pagination or Media Count:

115.0

Abstract:

The development of computer agents to enhance human-computer interfaces is an evolving field of study. This study examined whether people perceive personality in static digital faces that portray expressions of emotion, and if the digital faces would influence human performance on a simple human-machine collaborative task. The first experiment measured user-perception of personality based on the emotional expression in two sets of five static digital faces, one face from each set represented the five primary emotions Anger, Fear, Joy, Sad and Neutral. The independent variables were emotional Expression, Personality Trait, Face Set, Gender and Nationality. Using an internet survey, subjects were asked to rate the faces on the 30 personality sub-traits of the Big-Five Factor personality model on a scale of 1 to 5 1 not characteristic, 3 partially characteristic, 5 fully characteristic. These ratings were used to compare the perceived personality of each face. The results from this first phase revealed that participants provided different ratings of the personality sub-traits based on the emotional expression of a static digital face indicating perception of personality based on expression. There were also some isolated differences related to Gender and Face Set. Based on the results from experiment one, three faces were chosen for experiment two. The second experiment measured how faces with identified personality traits influence decision making in a simple collaborative task. Subjects were asked to read a survival scenario and then rank a list of items according to their value for survival.

Subject Categories:

  • Human Factors Engineering and Man Machine Systems

Distribution Statement:

APPROVED FOR PUBLIC RELEASE