SAR-NAS: Skeleton-based Action Recognition via Neural Architecture Searching

10/29/2020
by   Haoyuan Zhang, et al.
11

This paper presents a study of automatic design of neural network architectures for skeleton-based action recognition. Specifically, we encode a skeleton-based action instance into a tensor and carefully define a set of operations to build two types of network cells: normal cells and reduction cells. The recently developed DARTS (Differentiable Architecture Search) is adopted to search for an effective network architecture that is built upon the two types of cells. All operations are 2D based in order to reduce the overall computation and search space. Experiments on the challenging NTU RGB+D and Kinectics datasets have verified that most of the networks developed to date for skeleton-based action recognition are likely not compact and efficient. The proposed method provides an approach to search for such a compact network that is able to achieve comparative or even better performance than the state-of-the-art methods.

READ FULL TEXT

page 1

page 2

page 3

page 4

page 5

research
03/17/2020

Predictively Encoded Graph Convolutional Network for Noise-Robust Skeleton-based Action Recognition

In skeleton-based action recognition, graph convolutional networks (GCNs...
research
11/11/2019

Learning Graph Convolutional Network for Skeleton-based Human Action Recognition by Neural Searching

Human action recognition from skeleton data, fueled by the Graph Convolu...
research
10/28/2021

Skeleton-Based Mutually Assisted Interacted Object Localization and Human Action Recognition

Skeleton data carries valuable motion information and is widely explored...
research
07/10/2019

Video Action Recognition Via Neural Architecture Searching

Deep neural networks have achieved great success for video analysis and ...
research
06/09/2023

How Object Information Improves Skeleton-based Human Action Recognition in Assembly Tasks

As the use of collaborative robots (cobots) in industrial manufacturing ...
research
11/24/2020

KShapeNet: Riemannian network on Kendall shape space for Skeleton based Action Recognition

Deep Learning architectures, albeit successful in most computer vision t...
research
02/22/2020

Neural Architecture Search for Compressed Sensing Magnetic Resonance Image Reconstruction

Recent works have demonstrated that deep learning (DL) based compressed ...

Please sign up or login with your details

Forgot password? Click here to reset