Bodily expressed emotion understanding through integrating Laban movement analysis

Area of Science:

• Computer Science
• Psychology
• Affective Computing

Background:

• Bodily expressed emotion understanding (BEEU) aims to recognize emotions from human movement.
• Psychological studies confirm that specific motor elements convey emotions.
• Integrating motor elements can enhance automated emotion recognition systems.

Purpose of the Study:

• To develop a novel approach for BEEU by incorporating human motor elements.
• To introduce a precise dataset of human motor elements for research.
• To enhance BEEU models using a dual-source training strategy.

Main Methods:

• Introduction of BoME (body motor elements), a novel dataset for human motor elements.
• Application of baseline deep learning models to estimate motor elements from movement data.
• Development of a dual-source solution training BEEU models with both motor element and emotion labels.

Main Results:

• Deep learning models effectively learn representations of human movement from the BoME dataset.
• The proposed dual-source solution significantly improves BEEU performance on the BoLD benchmark.
• Experimental results demonstrate the benefit of integrating motor elements for emotion recognition.

Conclusions:

• Human motor elements are crucial for accurate bodily expressed emotion understanding.
• The BoME dataset and dual-source approach offer a promising direction for BEEU research.
• This work has implications for advancing emotion understanding and mental health analysis tools.