Food Image-Induced Discrete Emotion Recognition Using a Single-Channel Scalp-EEG Recording

Abstract

We construct a food image data set as visual stimulus cues for the field of affective computing. After 40 volunteers rating each of 500 food images in terms of the levels of arousal, valence, liking, and familiarity, 188 food images with high consistency scores are selected as the experimental stimuli that evoked three emotional tags: positive, neutral, and negative. The electroencephalogram (EEG) signals of 20 participants were recorded as each watched food images, comprising a food emotion dataset which can be an effective stimulus material. The results show that the food data sets can induce different emotions and the features of each EEG sub-band of different emotional labels have significant differences. This study proposes an innovative theme of emotional stimuli selection based on the food visual cues and expect to construct the optimal emotional feature space as input of deep learning model to achieve better classification accuracy on the real-time emotional recognition system.