Abstract:With the prevalence of accessible depth sensors, dynamic human body skeletons have attracted much attention as a robust modality for action recognition. Previous methods model skeletons based on RNN or CNN, which has limited expressive power for irregular skeleton joints. While graph convolutional networks (GCN) have been proposed to address irregular graph-structured data, the fundamental graph construction remains challenging. In this paper, we represent skeletons naturally on graphs, and propose a graph regression based GCN (GR-GCN) for skeleton-based action recognition, aiming to capture the spatio-temporal variation in the data. As the graph representation is crucial to graph convolution, we first propose graph regression to statistically learn the underlying graph from multiple observations. In particular, we provide spatio-temporal modeling of skeletons and pose an optimization problem on the graph structure over consecutive frames, which enforces the sparsity of the underlying graph for efficient representation. The optimized graph not only connects each joint to its neighboring joints in the same frame strongly or weakly, but also links with relevant joints in the previous and subsequent frames. We then feed the optimized graph into the GCN along with the coordinates of the skeleton sequence for feature learning, where we deploy high-order and fast Chebyshev approximation of spectral graph convolution. Further, we provide analysis of the variation characterization by the Chebyshev approximation. Experimental results validate the effectiveness of the proposed graph regression and show that the proposed GR-GCN achieves the state-of-the-art performance on the widely used NTU RGB+D, UT-Kinect and SYSU 3D datasets.

Joints Relation Inference Network for Skeleton-Based Action Recognition.

Convolutional Relation Network for Skeleton-Based Action Recognition.

Skeleton-based Action Recognition with JRR-GCN

Skeleton-Based Action Recognition With Directed Graph Neural Networks

Actional-Structural Graph Convolutional Networks for Skeleton-based Action Recognition

Effective skeleton topology and semantics-guided adaptive graph convolution network for action recognition

Generalized Graph Convolutional Networks for Skeleton-based Action Recognition

Spatio-Temporal Inception Graph Convolutional Networks for Skeleton-Based Action Recognition.

An improved spatial temporal graph convolutional network for robust skeleton-based action recognition

Multi-Stage Attention-Enhanced Sparse Graph Convolutional Network for Skeleton-Based Action Recognition

Optimized Skeleton-based Action Recognition via Sparsified Graph Regression

Skeleton-based action recognition by part-aware graph convolutional networks

Richly Activated Graph Convolutional Network for Robust Skeleton-Based Action Recognition

Combining channel-wise joint attention and temporal attention in graph convolutional networks for skeleton-based action recognition

Graph Edge Convolutional Neural Networks for Skeleton Based Action Recognition

High-Performance Inference Graph Convolutional Networks for Skeleton-Based Action Recognition

Temporal Enhanced Multi-Stream Graph Convolutional Nerual Networks For Skeleton-Based Action Recognition

Graph Edge Convolutional Neural Networks for Skeleton-Based Action Recognition

MFGCN: an efficient graph convolutional network based on multi-order feature information for human skeleton action recognition

Spatiotemporal Progressive Inward-Outward Aggregation Network for skeleton-based action recognition

Part-Wise Adaptive Topology Graph Convolutional Network for Skeleton-Based Action Recognition