EmoTVR: A Hybrid Model to Estimate Continuous-Time and Continuous-Level Emotion from Electroencephalography.

Fusing Multi-Level Features from Audio and Contextual Sentence Embedding from Text for Interview-Based Depression Detection.