.png){kind=icon}
.png){kind=icon}
Return
Continual spatio-temporal graph convolutional networks
.png){kind=icon}
OA
AI
DOI：10.1016/j.patcog.2023.109528.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Graph-based reasoning over skeleton data has emerged as a promising approach for human action recognition. However, the application of prior graph-based methods, which predominantly employ whole temporal sequences as their input, to the setting of online inference entails considerable computational redundancy. In this paper, we tackle this issue by reformulating the Spatio-Temporal Graph Convolutional Neural Network as a Continual Inference Network, which can perform step-by-step predictions in time without repeat frame processing. To evaluate our method, we create a continual version of ST-GCN, Co STGCN, alongside two derived methods with different self-attention mechanisms, Co AGCN and Co S-TR. We investigate weight transfer strategies and architectural modifications for inference acceleration, and perform experiments on the NTU RGB+D 60, NTU RGB+D 120, and Kinetics Skeleton 400 datasets. Retaining similar predictive accuracy, we observe up to 109 x reduction in time complexity, on-hardware accelerations of 26 x, and reductions in maximum allocated memory of 52% during online inference. (c) 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ )
Keywords:
Graph convolutional networks
Continual inference
Efficient deep learning
Skeleton-based action recognition
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.