Recognition of facial emotion based on SOAR model

Area of Science:

• Computer Science
• Artificial Intelligence
• Cognitive Science

Background:

• Facial emotion recognition is crucial for natural human-machine interaction.
• Accurate detection of emotional states from facial expressions enhances communication.
• Existing models often struggle with temporal dynamics and visual feature representation.

Purpose of the Study:

• To develop an efficient method for recognizing emotional states from facial images.
• To integrate deep learning with a cognitive model for improved emotion detection.
• To enhance the temporal learning and visual feature display in facial expression analysis.

Main Methods:

• A novel approach combining a 3D convolutional neural network (3DCNN) with learning automata (LA).
• Preprocessing video data into images for analysis, preserving all dimensions.
• Utilizing 3DCNN for temporal information processing and LA for optimizing backpropagation and network training.

Main Results:

• The proposed model achieved a facial emotion recognition accuracy rate of 85.3%.
• Demonstrated superior performance compared to competing facial emotion recognition models.
• Effectively learned the temporal order of frames and improved visual feature representation.

Conclusions:

• The integrated 3DCNN and LA model offers a significant advancement in facial emotion recognition.
• The SOAR model provides a robust framework for human-machine interaction by accurately interpreting emotions.
• This approach holds promise for various applications requiring sophisticated emotion detection capabilities.