Complex emotion recognition system using basic emotions via facial expression, electroencephalogram, and...

Area of Science:

• Affective computing
• Artificial intelligence
• Cognitive science

Background:

• Complex emotion recognition systems (CERS) analyze combined basic emotions, interconnections, and dynamics for nuanced understanding.
• Machine-based emotion recognition faces challenges in data acquisition and comprehending human-like cognitive processes.

Purpose of the Study:

• To review machine learning, deep learning, and meta-learning approaches for basic and complex emotion recognition.
• To assess the integration of physiological signals (ECG, EEG) with facial expressions for enhanced emotion detection.
• To explore applications, clinical implications, and research gaps in emotion recognition systems.

Main Methods:

• Comprehensive literature review of machine learning, deep learning, and meta-learning techniques.
• Analysis of studies utilizing facial expressions, electrocardiograms (ECG), and electroencephalograms (EEG) for emotion recognition.
• Evaluation of knowledge distillation and cognitive-inspired methods for CERS.

Main Results:

• Physiological signals (ECG, EEG) substantially improve CERS accuracy, dataset quality, and system reliability.
• Machine learning, deep learning, and meta-learning show efficacy in recognizing both basic and complex emotions.
• Meta-learning approaches demonstrate significant potential for enhancing system performance and guiding future research.

Conclusions:

• Integrating diverse data sources, including physiological signals, is crucial for robust CERS.
• Further research is needed to address existing gaps and challenges in emotion recognition systems.
• Meta-learning offers promising avenues for advancing affective computing and its applications in education and healthcare.