Adaptive Optimization Multi-Layer Perceptron for Emotion Recognition from EEG Brain Maps

Emotion recognition is an important issue in Affective Computing and Human Computer Interaction, and plays a crucial role in interpersonal communication and media effectiveness. In recent years, there has been a surge of interest in the field of emotion recognition, driven by advancements in machine learning models and computer vision techniques. These technologies have the potential to revolutionize the way people understand and interpret human emotions. This paper proposes a new electroencephalography (EEG) emotion recognition dataset based on daily media content that serves as emotion-inducing materials, including 16 video clips, 65 valid EEG signals and valence-arousal states self-assessment reports of subjects. And a novel EEG emotion recognition model based on multi-layer perceptron taking brain maps as spatial representations is proposed, using genetic algorithm to adaptively optimize the multi-layer perceptron structure in the algorithm that uses brain maps as input data and aims to predict emotional states of valence and arousal. The experimental results show that the method proposed in this paper has better performance than existing results in the literature.