Spatio-Temporal Joint Density Driven Learning for Skeleton-Based Action Recognition

TL;DR

This paper introduces a novel spatio-temporal joint density measure to enhance self-supervised skeleton-based action recognition by identifying discriminative joints and aligning their representations, leading to improved performance on multiple datasets.

Contribution

It proposes the spatial-temporal joint density (STJD) and two new learning strategies, STJD-CL and STJD-MP, to better utilize joint interactions for action classification.

Findings

Achieved 3.5-3.6% performance improvements over state-of-the-art methods.

Demonstrated effectiveness across NTU RGB+D 60, NTU RGB+D 120, and PKUMMD datasets.

Enhanced recognition accuracy by focusing on prime joints identified through STJD.

Abstract

Traditional approaches in unsupervised or self supervised learning for skeleton-based action classification have concentrated predominantly on the dynamic aspects of skeletal sequences. Yet, the intricate interaction between the moving and static elements of the skeleton presents a rarely tapped discriminative potential for action classification. This paper introduces a novel measurement, referred to as spatial-temporal joint density (STJD), to quantify such interaction. Tracking the evolution of this density throughout an action can effectively identify a subset of discriminative moving and/or static joints termed "prime joints" to steer self-supervised learning. A new contrastive learning strategy named STJD-CL is proposed to align the representation of a skeleton sequence with that of its prime joints while simultaneously contrasting the representations of prime and nonprime joints. In addition, a method called STJD-MP is developed by integrating it with a reconstruction-based framework for more effective learning. Experimental evaluations on the NTU RGB+D 60, NTU RGB+D 120, and PKUMMD datasets in various downstream tasks demonstrate that the proposed STJD-CL and STJD-MP improved performance, particularly by 3.5 and 3.6 percentage points over the state-of-the-art contrastive methods on the NTU RGB+D 120 dataset using X-sub and X-set evaluations, respectively.