Page 44 - 2024S
P. 44

UEC Int’l Mini-Conference No.52                                                               37
































                    Figure 6: Time plot by models


                                                              Figure 7: Setup of the user study experiment.
            turbo model also excelled in terms of response
            time, significantly outperforming other models.   Participants are shown watching a video with
                                                              the olfactory display device positioned nearby
            It had the lowest average response time, at just  for synchronized odor release.
            33% of the time taken by the gemini-1.0-pro
            model. This efficiency is crucial for real-time
            applications where prompt response speed is es-       Figure 7, participants watched two pre-
            sential.                                              pared videos under two conditions:

                                                                    • Without olfactory feedback: Partici-
            5 User Study                                               pants viewed the video and predicted

                                                                       potential odors based on visual and
            To evaluate the real-world effectiveness and user          auditory cues alone.
            perception of our olfactory-enhanced multime-
            dia system, we conducted a user study with 6            • With olfactory feedback: Participants
            participants. This study aimed to assess the im-           viewed the same video with synchro-
            pact of synchronized olfactory feedback on the             nized odor release from our olfactory
            viewing experience and to compare users’ odor              display system, positioned nearby as
            expectations with the system’s predictions. Fig-           depicted in the figure.
            ure 7 illustrates the setup of our user study ex-
            periment.                                          3. Odor Identification: After each viewing,
                                                                  participants recorded the odors they either
                                                                  expected (in the no-feedback condition) or
            5.1 Experiment Design                                 actually perceived (in the feedback condi-

            The user study consisted of the following com-        tion).
            ponents:
                                                               4. User Experience Survey: Participants
                                                                  completed a questionnaire assessing various
              1. Participant Information: We collected
                basic demographic information (age, sex)          aspects of their experience, including:
                and relevant health data (smoking status,           • The match between released odors and
                current nasal condition) to account for po-            video content
                tential variables affecting odor perception.
                                                                    • The level of immersion experienced
              2. Video Viewing Sessions: As shown in                   with olfactory feedback
   39   40   41   42   43   44   45   46   47   48   49