CFBM: Cubic Filtered Brain Map Creation Method and Models for Eeg-Based Emotion Recognition

Emotion recognition based on electroencephalogram (EEG) is one of the most prominent areas within biomedical re-search. With the advancement of deep learning, many studies have focused on developing more complex models to im-prove recognition performance. However, these approaches can not resol ve the challenges posed by the low discrim-inability of EEG data. In this paper, we propose a Cubic Filtered Brain Map (CFBM) creation method for EEG-based emotion recognition, which introduces an innovative use of Gaussian filtering to enhance feature extraction, thereby improving the discriminability of the EEG data. We theo-retically demonstrate the effectiveness of CFBM in reducing noise, enhancing spatial continuity, and improving feature discriminability. Additionally, we propose convolutional neural network (CNN) models that incorporate instance nor-malization (IN) and batch normalization (BN). Extensive experiments conducted on the DEAP dataset demonstrate the effectiveness of CFBM in enhancing EEG data discrim-inability across various model architectures. The recognition accuracies for valence and arousal reach 99.34% and 99.29%, achieving state-of-the-art performance. Our code is available at https://github.com/Gloria-G2021/CFBM.