Emotion Detection Using Hybrid Deep Sequential Learning Architecture from EEG Signals

Emotions play a vital role in decision-making, planning, reasoning, and other mental states in daily life, which are widely recognized as crucial roles in human interactions. It recognizes through non-physiological signals such as facial expressions, speech, and body language or physiological signals such as changes in mental state, heart rate, breathing, body temperature, and muscle tension. Electroencephalography (EEG) signals have good time resolution, which can provide a direct and comprehensive means for emotion recognition with higher classification accuracy. In this study, a novel emotion detection using hybrid deep sequential learning architecture from EEG signal dataset is proposed. It utilizes the advantages of Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) networks as it concatenation of LSTM and GRU deep sequential architectures. The proposed method outperforms other compared deep sequential classifiers in terms of accuracy, precision, recall, and $F1$-score, with 99.41%, 0.99, 0.99, and 0.99 respectively.