Memory-Based Pruning of Deep Neural Networks for IoT Devices Applied to Flood Detection

Automatic flood detection may be an important component for triggering damage control systems and minimizing the risk of social or economic impacts caused by flooding. Riverside images from regular cameras are a widely available resource that can be used for tackling this problem. Nevertheless, stat...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Francisco Erivaldo Fernandes Junior, Luis Gustavo Nonato, Caetano Mazzoni Ranieri, Jó Ueyama
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	deep neural networks semantic segmentation random pruning Internet of Things flood detection user preference Chemical technology TP1-1185
Acceso en línea:	https://doaj.org/article/d1c0e36827e74069b0d998d7eeaacc52
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:d1c0e36827e74069b0d998d7eeaacc52
record_format	dspace
spelling	oai:doaj.org-article:d1c0e36827e74069b0d998d7eeaacc522021-11-25T18:57:00ZMemory-Based Pruning of Deep Neural Networks for IoT Devices Applied to Flood Detection10.3390/s212275061424-8220https://doaj.org/article/d1c0e36827e74069b0d998d7eeaacc522021-11-01T00:00:00Zhttps://www.mdpi.com/1424-8220/21/22/7506https://doaj.org/toc/1424-8220Automatic flood detection may be an important component for triggering damage control systems and minimizing the risk of social or economic impacts caused by flooding. Riverside images from regular cameras are a widely available resource that can be used for tackling this problem. Nevertheless, state-of-the-art neural networks, the most suitable approach for this type of computer vision task, are usually resource-consuming, which poses a challenge for deploying these models within low-capability Internet of Things (IoT) devices with unstable internet connections. In this work, we propose a deep neural network (DNN) architecture pruning algorithm capable of finding a pruned version of a given DNN within a user-specified memory footprint. Our results demonstrate that our proposed algorithm can find a pruned DNN model with the specified memory footprint with little to no degradation of its segmentation performance. Finally, we show that our algorithm can be used in a memory-constraint wireless sensor network (WSN) employed to detect flooding events of urban rivers, and the resulting pruned models have competitive results compared with the original models.Francisco Erivaldo Fernandes JuniorLuis Gustavo NonatoCaetano Mazzoni RanieriJó UeyamaMDPI AGarticledeep neural networkssemantic segmentationrandom pruningInternet of Thingsflood detectionuser preferenceChemical technologyTP1-1185ENSensors, Vol 21, Iss 7506, p 7506 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	deep neural networks semantic segmentation random pruning Internet of Things flood detection user preference Chemical technology TP1-1185
spellingShingle	deep neural networks semantic segmentation random pruning Internet of Things flood detection user preference Chemical technology TP1-1185 Francisco Erivaldo Fernandes Junior Luis Gustavo Nonato Caetano Mazzoni Ranieri Jó Ueyama Memory-Based Pruning of Deep Neural Networks for IoT Devices Applied to Flood Detection
description	Automatic flood detection may be an important component for triggering damage control systems and minimizing the risk of social or economic impacts caused by flooding. Riverside images from regular cameras are a widely available resource that can be used for tackling this problem. Nevertheless, state-of-the-art neural networks, the most suitable approach for this type of computer vision task, are usually resource-consuming, which poses a challenge for deploying these models within low-capability Internet of Things (IoT) devices with unstable internet connections. In this work, we propose a deep neural network (DNN) architecture pruning algorithm capable of finding a pruned version of a given DNN within a user-specified memory footprint. Our results demonstrate that our proposed algorithm can find a pruned DNN model with the specified memory footprint with little to no degradation of its segmentation performance. Finally, we show that our algorithm can be used in a memory-constraint wireless sensor network (WSN) employed to detect flooding events of urban rivers, and the resulting pruned models have competitive results compared with the original models.
format	article
author	Francisco Erivaldo Fernandes Junior Luis Gustavo Nonato Caetano Mazzoni Ranieri Jó Ueyama
author_facet	Francisco Erivaldo Fernandes Junior Luis Gustavo Nonato Caetano Mazzoni Ranieri Jó Ueyama
author_sort	Francisco Erivaldo Fernandes Junior
title	Memory-Based Pruning of Deep Neural Networks for IoT Devices Applied to Flood Detection
title_short	Memory-Based Pruning of Deep Neural Networks for IoT Devices Applied to Flood Detection
title_full	Memory-Based Pruning of Deep Neural Networks for IoT Devices Applied to Flood Detection
title_fullStr	Memory-Based Pruning of Deep Neural Networks for IoT Devices Applied to Flood Detection
title_full_unstemmed	Memory-Based Pruning of Deep Neural Networks for IoT Devices Applied to Flood Detection
title_sort	memory-based pruning of deep neural networks for iot devices applied to flood detection
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/d1c0e36827e74069b0d998d7eeaacc52
work_keys_str_mv	AT franciscoerivaldofernandesjunior memorybasedpruningofdeepneuralnetworksforiotdevicesappliedtoflooddetection AT luisgustavononato memorybasedpruningofdeepneuralnetworksforiotdevicesappliedtoflooddetection AT caetanomazzoniranieri memorybasedpruningofdeepneuralnetworksforiotdevicesappliedtoflooddetection AT joueyama memorybasedpruningofdeepneuralnetworksforiotdevicesappliedtoflooddetection
_version_	1718410555364999168

Memory-Based Pruning of Deep Neural Networks for IoT Devices Applied to Flood Detection

Ejemplares similares