MEmoR: A Multimodal Emotion Recognition using affective biomarkers for smart prediction of emotional health for people analytics in smart industries

Highlights：

• A multimodal approach for emotion classification based on visual-physiological signal.

• Application of transfer learning for visual signals for more accurate and cost-effective computation.

• Identifying discrete emotions as well as valence -arousal affective dimensions of emotion.

• Weighted Late Fusion of visual and physiological signals for accurate prediction of the emotion.

• Extending predictive capabilities for turning digital biomarker and visual signals data into actions.