Classification of User Emotional Experiences on B2C Websites Utilizing Infrared Thermal Imaging

Area of Science:

• Human-Computer Interaction
• Affective Computing
• Psychophysiology

Background:

• Traditional physiological signal acquisition for emotion analysis is often intrusive and can yield inaccurate results.
• Noninvasive methods are needed to accurately capture emotional experiences during human-computer interaction.
• Infrared thermal imaging (IRTIs) presents a promising noninvasive approach for emotion detection.

Purpose of the Study:

• To investigate the efficacy of IRTIs in classifying user emotional experiences while interacting with business-to-consumer (B2C) websites.
• To explore the relationship between facial thermal patterns and self-reported emotional states.
• To identify key facial regions of interest (ROIs) and machine learning features for emotion classification.

Main Methods:

• Facial thermal images of 24 participants were captured while they interacted with B2C websites with manipulated usability and aesthetics.
• Machine learning techniques, including support vector machine (SVM), were employed for emotion classification.
• Participant self-assessments were used as ground truth for validating the classification results.

Main Results:

• Significant fluctuations in emotional valence were observed, allowing categorization into positive and negative states.
• Participant arousal levels remained consistent, supporting the valence-based classification.
• The SVM model demonstrated strong performance in differentiating baseline and emotional states.
• Key facial ROIs and effective machine learning features for emotion classification were identified.

Conclusions:

• IRTIs provide a viable noninvasive method for analyzing user emotional experiences during website interactions.
• The study establishes a significant connection between user emotions and IRTIs, advancing emotion analysis research.
• This research expands the application scope of IRTIs in understanding human affective states.