EEG data augmentation for emotion recognition with a multiple generator conditional Wasserstein GAN

Abstract EEG-based emotion recognition has attracted substantial attention from researchers due to its extensive application prospects, and progress been made in feature extraction classification modelling EEG data. However, insufficient high-quality training data are available for building models via machine learning or deep methods. The artificial generation of is an effective approach overcoming this problem. In paper, a multi-generator conditional Wasserstein GAN method proposed the that covers more comprehensive distribution real through use various generators. Experimental results demonstrate generated by model can effectively improve performance based on EEG.