Automatic Emotion Perception Using Eye Movement Information for E-Healthcare Systems

Area of Science:

• Biomedical Engineering
• Affective Computing
• Human-Computer Interaction

Background:

• Adolescent emotional state detection is crucial for effective E-Healthcare rehabilitation.
• Current E-Healthcare systems require improved methods for real-time emotional monitoring.
• Eye movement patterns offer a potential biomarker for emotional states.

Purpose of the Study:

• To develop and validate an eye movement-based algorithm for adolescent emotion perception.
• To integrate electrooculography (EOG) signals and eye movement video for enhanced emotion recognition.
• To compare feature-level fusion (FLF) and decision-level fusion (DLF) strategies for emotion classification.

Main Methods:

• Synchronous collection and analysis of electrooculography (EOG) signals and eye movement video.
• Extraction of time-frequency eye movement features using Short-Time Fourier Transform (STFT).
• Integration of time-domain features (saccade duration, fixation duration, pupil diameter) using FLF and DLF.

Main Results:

• The proposed algorithm achieved high accuracy in recognizing positive, neutral, and negative emotional states.
• Feature level fusion (FLF) yielded an average accuracy of 88.64%.
• Decision level fusion (DLF) with a maximal rule achieved an average accuracy of 88.35%.

Conclusions:

• Eye movement information effectively reflects adolescent emotional states.
• The developed algorithm provides a promising tool for enhancing E-Healthcare systems.
• Synchronous EOG and video analysis offers a robust approach for emotion perception in E-Healthcare.