Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Emotion Recognition in Contemporary Dance Performances Using Laban Movement Analysis

Muhammad Turab, Philippe Colantoni, Damien Muselet, Alain Tremeau

TL;DR

This work tackles emotion recognition in contemporary dance by enhancing Laban Movement Analysis (LMA) descriptors with temporal dynamics and introducing robust, novel features that capture both quantitative and qualitative motion aspects. It harnesses 3D pose data estimated via Neural Localizer Fields, derives a 54-feature LMA descriptor, and evaluates multiclass emotion classification using Random Forests and SVMs, with SHAP-based explainability to interpret predictions. The approach achieves high performance, reaching up to 96.85% accuracy with RF and provides transparent insights into feature importance, including body volume and movement timing. The framework offers practical benefits for performance analysis, dance training, and human–computer interaction, and paves the way for extending emotion and style recognition in dance while prioritizing interpretability.

Abstract

This paper presents a novel framework for emotion recognition in contemporary dance by improving existing Laban Movement Analysis (LMA) feature descriptors and introducing robust, novel descriptors that capture both quantitative and qualitative aspects of the movement. Our approach extracts expressive characteristics from 3D keypoints data of professional dancers performing contemporary dance under various emotional states, and trains multiple classifiers, including Random Forests and Support Vector Machines. Additionally, we provide in-depth explanation of features and their impact on model predictions using explainable machine learning methods. Overall, our study improves emotion recognition in contemporary dance and offers promising applications in performance analysis, dance training, and human--computer interaction, with a highest accuracy of 96.85\%.

Emotion Recognition in Contemporary Dance Performances Using Laban Movement Analysis

TL;DR

This work tackles emotion recognition in contemporary dance by enhancing Laban Movement Analysis (LMA) descriptors with temporal dynamics and introducing robust, novel features that capture both quantitative and qualitative motion aspects. It harnesses 3D pose data estimated via Neural Localizer Fields, derives a 54-feature LMA descriptor, and evaluates multiclass emotion classification using Random Forests and SVMs, with SHAP-based explainability to interpret predictions. The approach achieves high performance, reaching up to 96.85% accuracy with RF and provides transparent insights into feature importance, including body volume and movement timing. The framework offers practical benefits for performance analysis, dance training, and human–computer interaction, and paves the way for extending emotion and style recognition in dance while prioritizing interpretability.

Abstract

This paper presents a novel framework for emotion recognition in contemporary dance by improving existing Laban Movement Analysis (LMA) feature descriptors and introducing robust, novel descriptors that capture both quantitative and qualitative aspects of the movement. Our approach extracts expressive characteristics from 3D keypoints data of professional dancers performing contemporary dance under various emotional states, and trains multiple classifiers, including Random Forests and Support Vector Machines. Additionally, we provide in-depth explanation of features and their impact on model predictions using explainable machine learning methods. Overall, our study improves emotion recognition in contemporary dance and offers promising applications in performance analysis, dance training, and human--computer interaction, with a highest accuracy of 96.85\%.
Paper Structure (11 sections, 5 equations, 6 figures, 2 tables)

This paper contains 11 sections, 5 equations, 6 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Data Pre-processing
  5. 3D Human Body Pose Estimation
  6. LMA Feature Extraction
  7. Multi-Class Emotion Classification
  8. Model Explainability Using SHAP
  9. Results
  10. Conclusion
  11. Acknowledgments

Figures (6)

  • Figure 1: Overview of the proposed method for Emotion Recognition in Contemporary Dance Performances.
  • Figure 2: Ground truth and model predictions for three unseen dance performances. Images from Aristidou:2019:JOCCH.
  • Figure 3: Impact of sliding window size on accuracy.
  • Figure 4: Impact of top 10 features on models predictions.
  • Figure 5: Impact of features on the model's predictions for dance movements expressing anger.
  • ...and 1 more figures