Facial Expression Emotion Recognition Model Integrating Philosophy and Machine Learning Theory.

Area of Science:

• Computer Vision
• Machine Learning
• Psychology

Background:

• Facial expression emotion recognition is crucial for interpersonal communication and psychological analysis.
• Traditional methods struggle with feature extraction complexity and environmental influences.
• Ancient Chinese wisdom, like Zeng Guofan's Bing Jian, recognized facial cues for understanding personality.

Purpose of the Study:

• To propose a novel feature fusion dual-channel algorithm for enhanced facial expression emotion recognition.
• To address limitations of traditional methods, including insufficient feature extraction and susceptibility to external factors.
• To integrate machine learning with philosophical insights for a more robust emotion recognition system.

Main Methods:

• A dual-channel approach combining Gabor features from segmented regions of interest (ROI) and an efficient channel attention network.
• The first channel uses Gabor transform on the ROI to capture detailed local expression features.
• The second channel employs a depthwise separable convolution-based attention network to focus on critical features and reduce complexity.

Main Results:

• The proposed algorithm effectively extracts subtle facial expression features.
• The dual-channel fusion approach significantly improves emotion recognition accuracy.
• The method demonstrates superior performance on the FER2013 dataset compared to existing approaches.

Conclusions:

• The novel feature fusion dual-channel algorithm offers a more accurate and robust solution for facial expression emotion recognition.
• Integrating detailed local features with attention-based global feature extraction enhances system performance.
• This approach holds promise for applications in psychology and human-computer interaction.