Fumin Shen

Heterogeneous Graph Embedding for Multimodal Multi-Label Emotion Recognition

Cross-modal attention preservation with self-contrastive learning for composed query-based image retrieval