Skeleton-based human action recognition with global context-aware attention LSTM networks

Human action recognition in 3D skeleton sequences has attracted a lot of research attention. Recently, long short-term memory (LSTM) networks have shown promising performance in this task due to their strengths in modeling the dependencies and dynamics in sequential data. As not all skeletal joints...

Full description

Saved in:
Bibliographic Details
Main Authors: Liu, Jun, Wang, Gang, Duan, Ling-Yu, Abdiyeva, Kamila, Kot, Alex Chichung
Other Authors: School of Electrical and Electronic Engineering
Format: Article
Language:English
Published: 2020
Subjects:
Online Access:https://hdl.handle.net/10356/142304
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-142304
record_format dspace
spelling sg-ntu-dr.10356-1423042020-06-18T08:46:12Z Skeleton-based human action recognition with global context-aware attention LSTM networks Liu, Jun Wang, Gang Duan, Ling-Yu Abdiyeva, Kamila Kot, Alex Chichung School of Electrical and Electronic Engineering Engineering::Computer science and engineering Action Recognition Long Short-term Memory Human action recognition in 3D skeleton sequences has attracted a lot of research attention. Recently, long short-term memory (LSTM) networks have shown promising performance in this task due to their strengths in modeling the dependencies and dynamics in sequential data. As not all skeletal joints are informative for action recognition, and the irrelevant joints often bring noise which can degrade the performance, we need to pay more attention to the informative ones. However, the original LSTM network does not have explicit attention ability. In this paper, we propose a new class of LSTM network, global context-aware attention LSTM, for skeleton-based action recognition, which is capable of selectively focusing on the informative joints in each frame by using a global context memory cell. To further improve the attention capability, we also introduce a recurrent attention mechanism, with which the attention performance of our network can be enhanced progressively. Besides, a two-stream framework, which leverages coarse-grained attention and fine-grained attention, is also introduced. The proposed method achieves state-of-the-art performance on five challenging datasets for skeleton-based action recognition. NRF (Natl Research Foundation, S’pore) 2020-06-18T08:46:12Z 2020-06-18T08:46:12Z 2017 Journal Article Liu, J., Wang, G., Duan, L.-Y., Abdiyeva, K., & Kot, A. C. (2018). Skeleton-based human action recognition with global context-aware attention LSTM networks. IEEE Transactions on Image Processing, 27(4), 1586-1599. doi:10.1109/TIP.2017.2785279 1057-7149 https://hdl.handle.net/10356/142304 10.1109/TIP.2017.2785279 29324413 2-s2.0-85039779403 4 27 1586 1599 en IEEE Transactions on Image Processing © 2017 IEEE. All rights reserved.
institution Nanyang Technological University
building NTU Library
country Singapore
collection DR-NTU
language English
topic Engineering::Computer science and engineering
Action Recognition
Long Short-term Memory
spellingShingle Engineering::Computer science and engineering
Action Recognition
Long Short-term Memory
Liu, Jun
Wang, Gang
Duan, Ling-Yu
Abdiyeva, Kamila
Kot, Alex Chichung
Skeleton-based human action recognition with global context-aware attention LSTM networks
description Human action recognition in 3D skeleton sequences has attracted a lot of research attention. Recently, long short-term memory (LSTM) networks have shown promising performance in this task due to their strengths in modeling the dependencies and dynamics in sequential data. As not all skeletal joints are informative for action recognition, and the irrelevant joints often bring noise which can degrade the performance, we need to pay more attention to the informative ones. However, the original LSTM network does not have explicit attention ability. In this paper, we propose a new class of LSTM network, global context-aware attention LSTM, for skeleton-based action recognition, which is capable of selectively focusing on the informative joints in each frame by using a global context memory cell. To further improve the attention capability, we also introduce a recurrent attention mechanism, with which the attention performance of our network can be enhanced progressively. Besides, a two-stream framework, which leverages coarse-grained attention and fine-grained attention, is also introduced. The proposed method achieves state-of-the-art performance on five challenging datasets for skeleton-based action recognition.
author2 School of Electrical and Electronic Engineering
author_facet School of Electrical and Electronic Engineering
Liu, Jun
Wang, Gang
Duan, Ling-Yu
Abdiyeva, Kamila
Kot, Alex Chichung
format Article
author Liu, Jun
Wang, Gang
Duan, Ling-Yu
Abdiyeva, Kamila
Kot, Alex Chichung
author_sort Liu, Jun
title Skeleton-based human action recognition with global context-aware attention LSTM networks
title_short Skeleton-based human action recognition with global context-aware attention LSTM networks
title_full Skeleton-based human action recognition with global context-aware attention LSTM networks
title_fullStr Skeleton-based human action recognition with global context-aware attention LSTM networks
title_full_unstemmed Skeleton-based human action recognition with global context-aware attention LSTM networks
title_sort skeleton-based human action recognition with global context-aware attention lstm networks
publishDate 2020
url https://hdl.handle.net/10356/142304
_version_ 1681058758366593024