Liang-Ze Yin

ADMC: Attention-based Diffusion model for Missing modalities Completion

Multimodal emotion and intent recognition is essential for automated human-computer interaction, and it aims to analyze users’ speech, text, and visual information to predict their emotions or intent. One of the significant challenges is modality missingness caused by sensor malfunctions or incomplete data. Traditional...