This paper presents a study that utilizes artificial intelligence techniques to perform emotion analysis, investigating the impact of visual augmentations on the olfactory system. The visual stimuli were presented through HoloLens, depicting augmented particles, while the olfactory stimuli were derived from various aroma scents. The use of augmented particles enhanced the perception of the scent compared to a scent experience without visual stimulation.
We experimented by measuring the EEG (Fp1, Fp2, F3, and F4) of 30 participants when they were exposed to visual stimuli through augmented particles concurrent with smelling a scent. We used six scents in the experiment: lavender, lemon blossom, rose, gunpowder, smoke, and tar.
We presented visual stimuli in five different ways by controlling the color and movement of the particles. Two types of movements were examined in the experiment: one is spreading toward the users' faces from an outlet of an aroma shooter, and the other is a backward movement toward the outlet. All participants took part in experiments using all five methods of visual stimulus presentation, and each method used all six scents. After each trial with a particular scent, participants provided subjective evaluations of their experienced emotions (valence and arousal) through a questionnaire.
To analyze emotions, we utilized a support vector machine (SVM) for binary classification and used a 10-fold cross-validation to calculate the average accuracy. Moreover, we used repeated measures analysis of variance (RM ANOVA) to analyze the influence of olfactory perception and emotional stimulation based on the participants' questionnaire results.
We performed emotion classification using SVM to examine whether there are differences in emotions depending on the method of visual stimulus presentation. When we classified emotions using SVM with all four measured EEG channels, we found differences in emotion accuracy depending on the method of visual stimuli presentation, with an average accuracy of 0.93 for valence and 0.9 for arousal.
We used RM ANOVA to investigate the effects of changes in the color and direction of augmented particle movement on the olfactory system. Our study confirmed that visual stimuli can influence scent perception and induce emotional changes. We also discovered that the use of augmented particles of a specific color can intensify or alleviate emotions, and using green-colored augmented particles while setting the particle movement opposite to the direction of the scent emission can alleviate unpleasant emotions.
Through this, we anticipate a more comprehensive understanding of the interaction between visual stimuli and olfaction. Commercially, the appropriate use of visual stimuli and aromas can stimulate consumer emotions, potentially leading to marketing benefits. Additionally, this research can aid in the development of human-computer interaction (HCI) devices to be used in specific environments, such as restrooms and garbage disposal sites, where they can help alleviate the psychological discomfort caused by malodors.