MMPosE: Movie-induced Multi-label Positive Emotion Classification Through EEG Signals

Emotional information plays an important role in various multimedia applications. Movies, as a widely available form of content, can induce multiple positive emotions and stimulate people's pursuit better life. Different from negative emotions, are highly correlated difficult to distinguish the emotional space. Since different often induced simultaneously by movies, traditional single-target or multi-class methods not suitable for classification movie-induced emotions. In this paper, we propose TransEEG, model multi-label emotion viewer's brain activities when watching movies. The key features TransEEG include (1) explicitly modeling spatial correlation temporal dependencies multi-channel EEG signals using Transformer structure based model, which effectively addresses long-distance dependencies, (2) exploiting label-label correlations guide discriminative representation learning, that design Inter-Emotion Mask guiding Multi-Head Attention learn inter-emotion correlations, (3) constructing attention score vector representation-label matrix refine emotion-relevant features. To evaluate ability our classification, demonstrate on state-of-the-art database CPED. Extensive experimental results show proposed method achieves superior performance over competitive approaches.