MASc Seminar Announcement – Audio-Visual Feature Fusion through Transformers for Automated Depression Screening in Social Media Content

MASc Seminar Announcement

Audio-Visual Feature Fusion through Transformers for Automated Depression Screening in Social Media Content

Department of Electrical and Computer Engineering, University of Waterloo

I am pleased to share an important milestone in my graduate journey: my MASc seminar at the Department of Electrical and Computer Engineering, University of Waterloo.

This seminar presents my MASc thesis research on multimodal depression screening using social media videos, with a particular focus on transformer-based audio-visual feature fusion. The presentation highlights two core contributions: MDD-Net, which uses a mutual transformer to fuse acoustic and visual representations, and MMFformer, which explores multiple transformer-based fusion strategies for depression detection from audiovisual content.

The work is evaluated on the D-Vlog and LMVD datasets and demonstrates strong improvements over prior approaches, together with encouraging cross-corpus generalization results. All are welcome to attend.