PSTNet: Point Spatio-Temporal Convolution on Point Cloud Sequences

TL;DR

This paper introduces PSTNet, a novel deep learning architecture that employs point spatio-temporal convolution to effectively model and analyze irregular point cloud sequences for tasks like 3D action recognition and semantic segmentation.

Contribution

The paper proposes a new PST convolution that disentangles spatial and temporal features in point cloud sequences, enabling hierarchical feature extraction with improved modeling of 3D dynamics.

Findings

PSTNet outperforms existing methods on 3D action recognition datasets.

PSTNet achieves superior results in 4D semantic segmentation tasks.

The proposed method effectively captures local spatial structures and temporal dynamics.

Abstract

Point cloud sequences are irregular and unordered in the spatial dimension while exhibiting regularities and order in the temporal dimension. Therefore, existing grid based convolutions for conventional video processing cannot be directly applied to spatio-temporal modeling of raw point cloud sequences. In this paper, we propose a point spatio-temporal (PST) convolution to achieve informative representations of point cloud sequences. The proposed PST convolution first disentangles space and time in point cloud sequences. Then, a spatial convolution is employed to capture the local structure of points in the 3D space, and a temporal convolution is used to model the dynamics of the spatial regions along the time dimension. Furthermore, we incorporate the proposed PST convolution into a deep network, namely PSTNet, to extract features of point cloud sequences in a hierarchical manner. Extensive experiments on widely-used 3D action recognition and 4D semantic segmentation datasets demonstrate the effectiveness of PSTNet to model point cloud sequences.