Generative Data Augmentation for Skeleton Action Recognition
For researchers in skeleton-based action recognition, this method provides a practical solution to data scarcity, enabling better model performance with limited data.
The paper tackles the problem of expensive and labor-intensive collection of large-scale 3D skeleton datasets for action recognition. It proposes a conditional generative pipeline for data augmentation that improves recognition accuracy by up to 5% in few-shot settings and maintains competitive performance in full-data settings on HumanAct12 and NTU-RGBD datasets.
Skeleton-based human action recognition is a powerful approach for understanding human behaviour from pose data, but collecting large-scale, diverse, and well-annotated 3D skeleton datasets is both expensive and labor-intensive. To address this challenge, we propose a conditional generative pipeline for data augmentation in skeleton action recognition. Our method learns the distribution of real skeleton sequences under the constraint of action labels, enabling the synthesis of diverse and high-fidelity data. Even with limited training samples, it can effectively generate skeleton sequences and achieve competitive recognition performance in low-data scenarios, demonstrating strong generalisation in downstream tasks. Specifically, we introduce a Transformer-based encoder-decoder architecture, combined with a generative refinement module and a dropout mechanism, to balance fidelity and diversity during sampling. Experiments on HumanAct12 and the refined NTU-RGBD (NTU-VIBE) dataset show that our approach consistently improves the accuracy of multiple skeleton-based action recognition models, validating its effectiveness in both few-shot and full-data settings. The source code can be found at here.