Lightweight Semantic-Guided Neural Networks Based on Single Head Attention for Action Recognition
Skeleton-based action recognition can achieve a relatively high performance by transforming the human skeleton structure in an image into a graph and applying action recognition based on structural changes in the body. Among the many graph convolutional network (GCN) approaches used in skeleton-base...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-11-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/22/23/9249 |
_version_ | 1797462139713093632 |
---|---|
author | Seon-Bin Kim Chanhyuk Jung Byeong-Il Kim Byoung Chul Ko |
author_facet | Seon-Bin Kim Chanhyuk Jung Byeong-Il Kim Byoung Chul Ko |
author_sort | Seon-Bin Kim |
collection | DOAJ |
description | Skeleton-based action recognition can achieve a relatively high performance by transforming the human skeleton structure in an image into a graph and applying action recognition based on structural changes in the body. Among the many graph convolutional network (GCN) approaches used in skeleton-based action recognition, semantic-guided neural networks (SGNs) are fast action recognition algorithms that hierarchically learn spatial and temporal features by applying a GCN. However, because an SGN focuses on global feature learning rather than local feature learning owing to the structural characteristics, there is a limit to an action recognition in which the dependency between neighbouring nodes is important. To solve these problems and simultaneously achieve a real-time action recognition in low-end devices, in this study, a single head attention (SHA) that can overcome the limitations of an SGN is proposed, and a new SGN-SHA model that combines SHA with an SGN is presented. In experiments on various action recognition benchmark datasets, the proposed SGN-SHA model significantly reduced the computational complexity while exhibiting a performance similar to that of an existing SGN and other state-of-the-art methods. |
first_indexed | 2024-03-09T17:33:14Z |
format | Article |
id | doaj.art-c97e026bff9e4087af70575b690e5dfd |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-09T17:33:14Z |
publishDate | 2022-11-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-c97e026bff9e4087af70575b690e5dfd2023-11-24T12:11:03ZengMDPI AGSensors1424-82202022-11-012223924910.3390/s22239249Lightweight Semantic-Guided Neural Networks Based on Single Head Attention for Action RecognitionSeon-Bin Kim0Chanhyuk Jung1Byeong-Il Kim2Byoung Chul Ko3Department of Computer Engineering, Keimyung University, Daegu 42601, Republic of KoreaDepartment of Computer Engineering, Keimyung University, Daegu 42601, Republic of KoreaDepartment of Computer Engineering, Keimyung University, Daegu 42601, Republic of KoreaDepartment of Computer Engineering, Keimyung University, Daegu 42601, Republic of KoreaSkeleton-based action recognition can achieve a relatively high performance by transforming the human skeleton structure in an image into a graph and applying action recognition based on structural changes in the body. Among the many graph convolutional network (GCN) approaches used in skeleton-based action recognition, semantic-guided neural networks (SGNs) are fast action recognition algorithms that hierarchically learn spatial and temporal features by applying a GCN. However, because an SGN focuses on global feature learning rather than local feature learning owing to the structural characteristics, there is a limit to an action recognition in which the dependency between neighbouring nodes is important. To solve these problems and simultaneously achieve a real-time action recognition in low-end devices, in this study, a single head attention (SHA) that can overcome the limitations of an SGN is proposed, and a new SGN-SHA model that combines SHA with an SGN is presented. In experiments on various action recognition benchmark datasets, the proposed SGN-SHA model significantly reduced the computational complexity while exhibiting a performance similar to that of an existing SGN and other state-of-the-art methods.https://www.mdpi.com/1424-8220/22/23/9249action recognitionsemantic-guided neural networkssingle head attentiongraph convolutional networksskeletal structure |
spellingShingle | Seon-Bin Kim Chanhyuk Jung Byeong-Il Kim Byoung Chul Ko Lightweight Semantic-Guided Neural Networks Based on Single Head Attention for Action Recognition Sensors action recognition semantic-guided neural networks single head attention graph convolutional networks skeletal structure |
title | Lightweight Semantic-Guided Neural Networks Based on Single Head Attention for Action Recognition |
title_full | Lightweight Semantic-Guided Neural Networks Based on Single Head Attention for Action Recognition |
title_fullStr | Lightweight Semantic-Guided Neural Networks Based on Single Head Attention for Action Recognition |
title_full_unstemmed | Lightweight Semantic-Guided Neural Networks Based on Single Head Attention for Action Recognition |
title_short | Lightweight Semantic-Guided Neural Networks Based on Single Head Attention for Action Recognition |
title_sort | lightweight semantic guided neural networks based on single head attention for action recognition |
topic | action recognition semantic-guided neural networks single head attention graph convolutional networks skeletal structure |
url | https://www.mdpi.com/1424-8220/22/23/9249 |
work_keys_str_mv | AT seonbinkim lightweightsemanticguidedneuralnetworksbasedonsingleheadattentionforactionrecognition AT chanhyukjung lightweightsemanticguidedneuralnetworksbasedonsingleheadattentionforactionrecognition AT byeongilkim lightweightsemanticguidedneuralnetworksbasedonsingleheadattentionforactionrecognition AT byoungchulko lightweightsemanticguidedneuralnetworksbasedonsingleheadattentionforactionrecognition |