Texte intégral / Full text (pdf, 20 MiB) - Infoscience - EPFL
Texte intégral / Full text (pdf, 20 MiB) - Infoscience - EPFL
Texte intégral / Full text (pdf, 20 MiB) - Infoscience - EPFL
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
6.4. Eye-tracking for Interaction<br />
Did the character on screen seem interested or indifferent? (1: very interested, 2: fairly<br />
interested, 3: neither interested nor indifferent, 4: fairly indifferent, 5: very indifferent)<br />
Did the character on screen seem engaged or distracted during the conversation? (1:<br />
very engaged, 2: fairly engaged, 3: neither engaged nor distracted, 4: fairly distracted,<br />
5: very distracted)<br />
Did the character seem friendly or not? (1: very friendly, 2: fairly friendly, 3: neither<br />
friendly nor unfriendly, 4: fairly unfriendly, 5: very unfriendly)<br />
Did the character seem to behave in a normal way with regards to what you were<br />
telling him/her? (1: very normal, 2: fairly normal, 3: neither normal nor abnormal, 4:<br />
fairly abnormal, 5: very abnormal)<br />
The first two questions were asked in order to verify the distinction between attentive<br />
and bored behaviors. The third question does not indicate much, it was only asked in order<br />
to divert attention from the questions specific to our experiment. The fourth question was<br />
asked in order to verify our first hypothesis; that a character which changes behavior is more<br />
realistic than one who does not. It was equally asked in order to verify our second hypothesis;<br />
that a character changing behavior randomly is less realistic than one who changes with<br />
regards to what the subject is looking at. A fifth question was asked in order to obtain<br />
feedback from the subjects on any point which would not have been addressed by their<br />
answers to the previous questions:<br />
On the whole, how would you describe the differences in the characters’ behaviors, if<br />
any, in the four scenes?<br />
6.4.2 Experimental Protocol - Case Study<br />
In addition to the validation conducted on these 12 healthy subjects, we have tested our<br />
application on a young 14 year-old girl suffering from Asperger syndrome. As explained<br />
in the Glossary (see Appendix D), the Asperger syndrome is an Autism Spectrum Disorder<br />
(ASD). One of the diagnostic criteria of the Asperger syndrome is a marked impairment in<br />
social interaction [American Psychiatric Association, 1994].<br />
As a first point, it is important to note that the experiment did not aim at any therapeutic<br />
benefit for the subject. It was aimed to test our architecture with someone having problems<br />
with social situations. This young girl accepted to test our setup and give us her feedback.<br />
She was accompanied by her therapist and her mother for the experiment. Her mother, however,<br />
did not stay in the same room as the experiment went on. We did not tell her what<br />
we were looking for. She was thus naive towards our objectives. As a first step, we seated<br />
her in front of the large back-projection screen on which we displayed the scene depicted in<br />
Figure 6.11. We asked her to wear the coupled eye- and head-tracker and recorded her eye<br />
movements throughout the experiment. We then asked her to do a public speaking exercise,<br />
and talk to the virtual character for a couple of minutes. The complete session lasted for approximately<br />
10 minutes. We first made sure that the character would be attentive even if not<br />
97