Human emotion recognition based on brain signal analysis using fuzzy neural network

An important method in the “human–computer interface” (HCI) is emotion recognition. Conventional emotion detection algorithms rely on exterior behaviors like facial expression, which may not be able to accurately capture authentic human emotion because facial expression signals can sometimes be hidden. Recently, it has become increasingly important for the BCI system to become more intelligent to recognize emotions using electroencephalograms (EEG). EEG signal is very similar to human emotion and can therefore accurately represent human emotion. Since EEG signals are often noisy, nonlinear, and contain non-stationary characteristics, it is challenging to develop an intelligent framework that can distinguish emotions with high accuracy. To accurately analyze emotions from brain data, we offer a fuzzy method-based emotion identification model in this study. Initially, the DEAP dataset is collected and preprocessed using Butterworth “bandpass” filter as well as stationary “wavelet transform” filters. Signal features are extracted using “Hilbert Huang transform” and “differential entropy” methods. The proposed “fractional feedback fuzzy neural network” with the “chaotic drifted Red Deer optimization algorithm” effectively classifies the emotions from the extracted features. Our suggested method increases the recognition accuracy rate when compared to conventional emotion recognition algorithms, according to experiments on the DEAP dataset. Applications for real-time automatic emotion recognition can leverage the fuzzy model we built.