Authors:
Matthias Trojahn
1
;
Florian Arndt
1
;
Markus Weinmann
2
and
Frank Ortmeier
3
Affiliations:
1
Volkswagen AG, Germany
;
2
TU Braunschweig, Germany
;
3
Otto-von-Guericke University of Magdeburg, Germany
Keyword(s):
Human-Computer Interaction, Emotion Recognition, Keystroke Dynamics, Touchscreen, Smartphone.
Related
Ontology
Subjects/Areas/Topics:
Emotional and Affective Computing
;
Enterprise Information Systems
;
Guidelines, Principles, Patterns and Standards
;
Human-Computer Interaction
Abstract:
Automatic emotion recognition through computers offers a lot of advantages, as the interaction between human
and computers can be improved. For example, it is possible to be responsive to anger or frustration of
customers automatically while working with a webpage. Mouse cursor movements and keystroke dynamics
were already used and examined for such a recognition on conventional keyboards.
The aim of this work is to investigate keystroke dynamics on touchscreen keyboards which gets a cumulative
relevance through the increasingly further circulation of smartphones and tablets. Furthermore, it is possible
to record additional information like pressure and size of keystrokes. This could increase the recognition rate
for emotions. In order to record the keystroke dynamics, an application and keyboard layout for Android
OS were developed. In addition, hypotheses were established on the basis of Yerkes-Dodson-Law and Flow
theory and besides, a study with 152 test persons for the data coll
ection was implemented. Subsequently, a
data evaluation with the SPSS software was accomplished. Most of the hypotheses were confirmed and the
results of the study show that emotions can be explained by the keystroke dynamics and recognized in this
way.
(More)