Cargando…

Fall Detection with the Spatial-Temporal Correlation Encoded by a Sequence-to-Sequence Denoised GAN

Falling is a major cause of personal injury and accidental death worldwide, in particular for the elderly. For aged care, a falling alarm system is highly demanded so that medical aid can be obtained immediately when the fall accidents happen. Previous studies on fall detection lacked practical cons...

Descripción completa

Detalles Bibliográficos
Autores principales: Hsu, Wei-Wen, Guo, Jing-Ming, Chen, Chien-Yu, Chang, Yao-Chung
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9185321/
https://www.ncbi.nlm.nih.gov/pubmed/35684812
http://dx.doi.org/10.3390/s22114194
_version_ 1784724695710433280
author Hsu, Wei-Wen
Guo, Jing-Ming
Chen, Chien-Yu
Chang, Yao-Chung
author_facet Hsu, Wei-Wen
Guo, Jing-Ming
Chen, Chien-Yu
Chang, Yao-Chung
author_sort Hsu, Wei-Wen
collection PubMed
description Falling is a major cause of personal injury and accidental death worldwide, in particular for the elderly. For aged care, a falling alarm system is highly demanded so that medical aid can be obtained immediately when the fall accidents happen. Previous studies on fall detection lacked practical considerations to deal with real-world situations, including the camera’s mounting angle, lighting differences between day and night, and the privacy protection for users. In our experiments, IR-depth images and thermal images were used as the input source for fall detection; as a result, detailed facial information is not captured by the system for privacy reasons, and it is invariant to the lighting conditions. Due to the different occurrence rates between fall accidents and other normal activities, supervised learning approaches may suffer from the problem of data imbalance in the training phase. Accordingly, in this study, anomaly detection is performed using unsupervised learning approaches so that the models were trained only with the normal cases while the fall accident was defined as an anomaly event. The proposed system takes sequential frames as the inputs to predict future frames based on a GAN structure, and it provides (1) multi-subject detection, (2) real-time fall detection triggered by motion, (3) a solution to the situation that subjects were occluded after falling, and (4) a denoising scheme for depth images. The experimental results show that the proposed system achieves the state-of-the-art performance and copes with the real-world cases successfully.
format Online
Article
Text
id pubmed-9185321
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-91853212022-06-11 Fall Detection with the Spatial-Temporal Correlation Encoded by a Sequence-to-Sequence Denoised GAN Hsu, Wei-Wen Guo, Jing-Ming Chen, Chien-Yu Chang, Yao-Chung Sensors (Basel) Article Falling is a major cause of personal injury and accidental death worldwide, in particular for the elderly. For aged care, a falling alarm system is highly demanded so that medical aid can be obtained immediately when the fall accidents happen. Previous studies on fall detection lacked practical considerations to deal with real-world situations, including the camera’s mounting angle, lighting differences between day and night, and the privacy protection for users. In our experiments, IR-depth images and thermal images were used as the input source for fall detection; as a result, detailed facial information is not captured by the system for privacy reasons, and it is invariant to the lighting conditions. Due to the different occurrence rates between fall accidents and other normal activities, supervised learning approaches may suffer from the problem of data imbalance in the training phase. Accordingly, in this study, anomaly detection is performed using unsupervised learning approaches so that the models were trained only with the normal cases while the fall accident was defined as an anomaly event. The proposed system takes sequential frames as the inputs to predict future frames based on a GAN structure, and it provides (1) multi-subject detection, (2) real-time fall detection triggered by motion, (3) a solution to the situation that subjects were occluded after falling, and (4) a denoising scheme for depth images. The experimental results show that the proposed system achieves the state-of-the-art performance and copes with the real-world cases successfully. MDPI 2022-05-31 /pmc/articles/PMC9185321/ /pubmed/35684812 http://dx.doi.org/10.3390/s22114194 Text en © 2022 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Hsu, Wei-Wen
Guo, Jing-Ming
Chen, Chien-Yu
Chang, Yao-Chung
Fall Detection with the Spatial-Temporal Correlation Encoded by a Sequence-to-Sequence Denoised GAN
title Fall Detection with the Spatial-Temporal Correlation Encoded by a Sequence-to-Sequence Denoised GAN
title_full Fall Detection with the Spatial-Temporal Correlation Encoded by a Sequence-to-Sequence Denoised GAN
title_fullStr Fall Detection with the Spatial-Temporal Correlation Encoded by a Sequence-to-Sequence Denoised GAN
title_full_unstemmed Fall Detection with the Spatial-Temporal Correlation Encoded by a Sequence-to-Sequence Denoised GAN
title_short Fall Detection with the Spatial-Temporal Correlation Encoded by a Sequence-to-Sequence Denoised GAN
title_sort fall detection with the spatial-temporal correlation encoded by a sequence-to-sequence denoised gan
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9185321/
https://www.ncbi.nlm.nih.gov/pubmed/35684812
http://dx.doi.org/10.3390/s22114194
work_keys_str_mv AT hsuweiwen falldetectionwiththespatialtemporalcorrelationencodedbyasequencetosequencedenoisedgan
AT guojingming falldetectionwiththespatialtemporalcorrelationencodedbyasequencetosequencedenoisedgan
AT chenchienyu falldetectionwiththespatialtemporalcorrelationencodedbyasequencetosequencedenoisedgan
AT changyaochung falldetectionwiththespatialtemporalcorrelationencodedbyasequencetosequencedenoisedgan