SA-DVAE: Improving Zero-Shot Skeleton-Based Action Recognition by Disentangled Variational Autoencoders

TL;DR

SA-DVAE introduces a novel disentangled variational autoencoder approach to improve zero-shot skeleton-based action recognition by better aligning semantic and skeleton features, addressing dataset imbalance issues.

Contribution

The paper proposes a disentangled VAE framework that separates semantic and irrelevant features to enhance feature alignment in zero-shot action recognition.

Findings

SA-DVAE outperforms existing methods on benchmark datasets.

Feature disentanglement improves semantic-skeleton feature alignment.

Experimental results demonstrate significant accuracy gains.

Abstract

Existing zero-shot skeleton-based action recognition methods utilize projection networks to learn a shared latent space of skeleton features and semantic embeddings. The inherent imbalance in action recognition datasets, characterized by variable skeleton sequences yet constant class labels, presents significant challenges for alignment. To address the imbalance, we propose SA-DVAE -- Semantic Alignment via Disentangled Variational Autoencoders, a method that first adopts feature disentanglement to separate skeleton features into two independent parts -- one is semantic-related and another is irrelevant -- to better align skeleton and semantic features. We implement this idea via a pair of modality-specific variational autoencoders coupled with a total correction penalty. We conduct experiments on three benchmark datasets: NTU RGB+D, NTU RGB+D 120 and PKU-MMD, and our experimental results show that SA-DAVE produces improved performance over existing methods. The code is available at https://github.com/pha123661/SA-DVAE.