Asymmetric Residual Neural Network for Accurate Human Activity Recognition

Human activity recognition (HAR) using deep neural networks has become a hot topic in human−computer interaction. Machines can effectively identify human naturalistic activities by learning from a large collection of sensor data. Activity recognition is not only an interesting research pro...

Full description

Bibliographic Details
Main Authors: Jun Long, Wuqing Sun, Zhan Yang, Osolo Ian Raymond
Format: Article
Language:English
Published: MDPI AG 2019-06-01
Series:Information
Subjects:
Online Access:https://www.mdpi.com/2078-2489/10/6/203
id doaj-517b60997d0544f7be0c47b8419de021
record_format Article
spelling doaj-517b60997d0544f7be0c47b8419de0212020-11-25T00:20:31ZengMDPI AGInformation2078-24892019-06-0110620310.3390/info10060203info10060203Asymmetric Residual Neural Network for Accurate Human Activity RecognitionJun Long0Wuqing Sun1Zhan Yang2Osolo Ian Raymond3School of Computer Science and Engineering, Central South University, Changsha 410083, ChinaSchool of Computer Science and Engineering, Central South University, Changsha 410083, ChinaSchool of Computer Science and Engineering, Central South University, Changsha 410083, ChinaSchool of Computer Science and Engineering, Central South University, Changsha 410083, ChinaHuman activity recognition (HAR) using deep neural networks has become a hot topic in human−computer interaction. Machines can effectively identify human naturalistic activities by learning from a large collection of sensor data. Activity recognition is not only an interesting research problem but also has many real-world practical applications. Based on the success of residual networks in achieving a high level of aesthetic representation of automatic learning, we propose a novel asymmetric residual network, named ARN. ARN is implemented using two identical path frameworks consisting of (1) a short time window, which is used to capture spatial features, and (2) a long time window, which is used to capture fine temporal features. The long time window path can be made very lightweight by reducing its channel capacity, while still being able to learn useful temporal representations for activity recognition. In this paper, we mainly focus on proposing a new model to improve the accuracy of HAR. In order to demonstrate the effectiveness of the ARN model, we carried out extensive experiments on benchmark datasets (i.e., OPPORTUNITY, UniMiB-SHAR) and compared the results with some conventional and state-of-the-art learning-based methods. We discuss the influence of networks parameters on performance to provide insights about its optimization. Results from our experiments show that ARN is effective in recognizing human activities via wearable datasets.https://www.mdpi.com/2078-2489/10/6/203human activity recognitiondeep neural networkresidual networksspatial featurestemporal features
collection DOAJ
language English
format Article
sources DOAJ
author Jun Long
Wuqing Sun
Zhan Yang
Osolo Ian Raymond
spellingShingle Jun Long
Wuqing Sun
Zhan Yang
Osolo Ian Raymond
Asymmetric Residual Neural Network for Accurate Human Activity Recognition
Information
human activity recognition
deep neural network
residual networks
spatial features
temporal features
author_facet Jun Long
Wuqing Sun
Zhan Yang
Osolo Ian Raymond
author_sort Jun Long
title Asymmetric Residual Neural Network for Accurate Human Activity Recognition
title_short Asymmetric Residual Neural Network for Accurate Human Activity Recognition
title_full Asymmetric Residual Neural Network for Accurate Human Activity Recognition
title_fullStr Asymmetric Residual Neural Network for Accurate Human Activity Recognition
title_full_unstemmed Asymmetric Residual Neural Network for Accurate Human Activity Recognition
title_sort asymmetric residual neural network for accurate human activity recognition
publisher MDPI AG
series Information
issn 2078-2489
publishDate 2019-06-01
description Human activity recognition (HAR) using deep neural networks has become a hot topic in human−computer interaction. Machines can effectively identify human naturalistic activities by learning from a large collection of sensor data. Activity recognition is not only an interesting research problem but also has many real-world practical applications. Based on the success of residual networks in achieving a high level of aesthetic representation of automatic learning, we propose a novel asymmetric residual network, named ARN. ARN is implemented using two identical path frameworks consisting of (1) a short time window, which is used to capture spatial features, and (2) a long time window, which is used to capture fine temporal features. The long time window path can be made very lightweight by reducing its channel capacity, while still being able to learn useful temporal representations for activity recognition. In this paper, we mainly focus on proposing a new model to improve the accuracy of HAR. In order to demonstrate the effectiveness of the ARN model, we carried out extensive experiments on benchmark datasets (i.e., OPPORTUNITY, UniMiB-SHAR) and compared the results with some conventional and state-of-the-art learning-based methods. We discuss the influence of networks parameters on performance to provide insights about its optimization. Results from our experiments show that ARN is effective in recognizing human activities via wearable datasets.
topic human activity recognition
deep neural network
residual networks
spatial features
temporal features
url https://www.mdpi.com/2078-2489/10/6/203
work_keys_str_mv AT junlong asymmetricresidualneuralnetworkforaccuratehumanactivityrecognition
AT wuqingsun asymmetricresidualneuralnetworkforaccuratehumanactivityrecognition
AT zhanyang asymmetricresidualneuralnetworkforaccuratehumanactivityrecognition
AT osoloianraymond asymmetricresidualneuralnetworkforaccuratehumanactivityrecognition
_version_ 1725367058480234496