Dance Style Recognition Using Laban Movement Analysis

TL;DR

This work tackles dance style recognition by enriching Laban Movement Analysis (LMA) features with temporal context. It presents a pipeline that combines floor-aware 3D body modeling (3D pose estimation, floor reconstruction, SMPL fitting) with a sliding-window approach to capture movement evolution, followed by multi-class classification using RF and SVM and SHAP-based explainability. The approach achieves state-of-the-art performance on the AIST++ dataset, reaching up to 99.68% accuracy with RF and providing interpretable feature contributions via SHAP, with temporal dynamics like Effort Time and Body Volume emerging as key drivers. The study demonstrates the value of temporal context in expressive movement analysis and outlines directions to improve robustness across diverse performances and recording conditions, enabling more reliable, explainable dance style recognition in real-world settings.

Abstract

The growing interest in automated movement analysis has presented new challenges in recognition of complex human activities including dance. This study focuses on dance style recognition using features extracted using Laban Movement Analysis. Previous studies for dance style recognition often focus on cross-frame movement analysis, which limits the ability to capture temporal context and dynamic transitions between movements. This gap highlights the need for a method that can add temporal context to LMA features. For this, we introduce a novel pipeline which combines 3D pose estimation, 3D human mesh reconstruction, and floor aware body modeling to effectively extract LMA features. To address the temporal limitation, we propose a sliding window approach that captures movement evolution across time in features. These features are then used to train various machine learning methods for classification, and their explainability explainable AI methods to evaluate the contribution of each feature to classification performance. Our proposed method achieves a highest classification accuracy of 99.18\% which shows that the addition of temporal context significantly improves dance style recognition performance.

Figures (7)

• Figure 1: Overview of the proposed method for dance style recognition.
• Figure 2: Impact of sliding window size on dance style classification accuracy.
• Figure 3: Impact and contribution of the top 10 features to the model predictions.
• Figure 4: SHAP values illustrating feature contributions to the model’s predictions for the Middle Hip Hop dance style.
• Figure 5: SHAP values illustrating feature contributions to the model’s predictions for the Waack dance style.