Return
Milmer: a framework for multiple instance learning based multimodal emotion recognition
DOI:10.1016/j.neucom.2026.134368.png)
Abstract
En 中文
• We provide a comprehensive fusion of facial expressions and EEG for emotion recognition, with open-source code. • Transformer-based fusion effectively integrates visual and physiological modalities, outperforming simplistic methods. • Multi-instance learning and a fine-tuned Swin Transformer [1] extract facial features; cross-attention fuses with EEG.
Journal
IF:
6.5
Papers:
2.5W
Citations:
6.5W

