Facial Emotional State Recognition Inspired by the Human Brain

Facial emotional expressions play a fundamental role in conveying intentions and enhancing the quality of human communication. With the rapid advancement of facial recognition technologies and the increasing importance of emotion recognition, this study investigates facial emotional expression recognition using a Brain Emotional Learning (BEL) model. The Brain Emotional Learning model is inspired by the human brain’s limbic system, which is responsible for processing emotional stimuli. The primary objective of this study is to improve the recognition rate of facial emotional expressions. The input to the proposed model is the standard Cohn–Kanade (CK) dataset, which contains six emotional states: happiness, sadness, anger, surprise, fear, and disgust. Image features are extracted using Principal Component Analysis (PCA) and subsequently fed into the classification stage of the Brain Emotional Learning model to determine the recognition rate of facial emotional expressions. In addition, a correlation matrix comprising the eyebrow, eye, and mouth features is constructed, and facial emotional states are identified through an averaging process. The correlation matrix is capable of recognizing facial emotional states when only one of these features is provided as input to the system. The analysis of the dataset demonstrates a facial emotion recognition rate of ۹۵.۸۱%. The findings indicate that the proposed Brain Emotional Learning model achieves a high level of accuracy in recognizing facial emotional expressions.

کلیدواژه ها:

Brain Emotional Learning (BEL) Model ، Principal component analysis (PCA) ، Correlation Matrix ، Facial expression recognition