SiaCon-DetNet with HySHO: a cutting-edge transformer-based deep learning framework for emotion-aware facial...

Area of Science:

• Artificial Intelligence
• Computer Vision
• Affective Computing

Background:

• Facial emotion recognition (FER) is crucial for human-computer interaction and psychological analysis.
• Existing FER methods struggle with feature representation, expression variation, and model generalization.
• There is a need for more robust and accurate FER systems.

Purpose of the Study:

• To present a novel FER model combining SiaCon-DetNet and HySHO algorithms.
• To improve fine-grained facial feature detection and model generalization.
• To develop an adaptive and efficient emotion detection framework.

Main Methods:

• The proposed model integrates SiaCon-DetNet for feature learning and transformer attention mechanisms.
• It combines bio-inspired optimization with deep learning (HySHO algorithm) for adaptive parameter adjustment.
• The process involves face region detection using a Siamese network and feature enhancement via multi-head self-attention.

Main Results:

• The new FER model achieved up to 99.20% accuracy on the JAFFE database.
• The model demonstrated superior performance compared to existing FER methods.
• Precision, recall, and F1-scores were consistently between 98-99%, indicating high reliability.

Conclusions:

• The proposed FER model effectively addresses limitations in current approaches.
• The integration of SiaCon-DetNet and HySHO results in a highly accurate and efficient emotion recognition system.
• The framework shows significant potential for advancing affective computing applications.