multimodal emotion recognition (MER)

This collection contains all research outputs for DF2025-069, a comparative study of three multimodal fusion architectures for six-class discrete emotion recognition on CMU-MOSEI, conducted as part of the DataraFlow Internship Programme.Items included: ∙ Preprint paper (PDF) ∙ Notebook 1: LSTM experiments (MER-Base and MER-Attn) ∙ Notebook 2: Cross-Modal Transformer (MER-CMT)Supervisor: Winner Emeto (DataraFlow) | dataraflow@gmail.comORCID: 0009-0008-7279-9663Repository: https://github.com/visionbyangelic/Multimodal-Emotion-Recognition​​​​​​​​​​​​​​​​