Deep Machine Learning Model Trade-Offs for Malaria Elimination in Resource-Constrained Locations
The success of deep machine learning (DML) models in gaming and robotics has increased its trial in clinical and public healthcare solutions. In applying DML to healthcare problems, a special challenge of inadequate electrical energy and computing resources exists in regional and developing areas of...
Published in: | Bioengineering |
---|---|
Main Authors: | , |
Format: | Text |
Language: | English |
Published: |
Multidisciplinary Digital Publishing Institute
2021
|
Subjects: | |
Online Access: | https://doi.org/10.3390/bioengineering8110150 |
id |
ftmdpi:oai:mdpi.com:/2306-5354/8/11/150/ |
---|---|
record_format |
openpolar |
spelling |
ftmdpi:oai:mdpi.com:/2306-5354/8/11/150/ 2023-08-20T04:06:09+02:00 Deep Machine Learning Model Trade-Offs for Malaria Elimination in Resource-Constrained Locations Peter U. Eze Clement O. Asogwa agris 2021-10-21 application/pdf https://doi.org/10.3390/bioengineering8110150 EN eng Multidisciplinary Digital Publishing Institute https://dx.doi.org/10.3390/bioengineering8110150 https://creativecommons.org/licenses/by/4.0/ Bioengineering; Volume 8; Issue 11; Pages: 150 deep learning resource optimisation model quantisation malaria digital health edge devices Text 2021 ftmdpi https://doi.org/10.3390/bioengineering8110150 2023-08-01T03:01:50Z The success of deep machine learning (DML) models in gaming and robotics has increased its trial in clinical and public healthcare solutions. In applying DML to healthcare problems, a special challenge of inadequate electrical energy and computing resources exists in regional and developing areas of the world. In this paper, we evaluate and report the computational and predictive performance design trade-offs for four candidate deep learning models that can be deployed for rapid malaria case finding. The goal is to maximise malaria detection accuracy while reducing computing resource and energy consumption. Based on our experimental results using a blood smear malaria test data set, the quantised versions of Basic Convolutional Neural Network (B-CNN) and MobileNetV2 have better malaria detection performance (up to 99% recall), lower memory usage (2MB 8-bit quantised model) and shorter inference time (33–95 microseconds on mobile phones) than VGG-19 fine-tuned and quantised models. Hence, we have implemented MobileNetV2 in our mobile application as it has even a lower memory requirement than B-CNN. This work will help to counter the negative effects of COVID-19 on the previous successes towards global malaria elimination. Text DML MDPI Open Access Publishing Bioengineering 8 11 150 |
institution |
Open Polar |
collection |
MDPI Open Access Publishing |
op_collection_id |
ftmdpi |
language |
English |
topic |
deep learning resource optimisation model quantisation malaria digital health edge devices |
spellingShingle |
deep learning resource optimisation model quantisation malaria digital health edge devices Peter U. Eze Clement O. Asogwa Deep Machine Learning Model Trade-Offs for Malaria Elimination in Resource-Constrained Locations |
topic_facet |
deep learning resource optimisation model quantisation malaria digital health edge devices |
description |
The success of deep machine learning (DML) models in gaming and robotics has increased its trial in clinical and public healthcare solutions. In applying DML to healthcare problems, a special challenge of inadequate electrical energy and computing resources exists in regional and developing areas of the world. In this paper, we evaluate and report the computational and predictive performance design trade-offs for four candidate deep learning models that can be deployed for rapid malaria case finding. The goal is to maximise malaria detection accuracy while reducing computing resource and energy consumption. Based on our experimental results using a blood smear malaria test data set, the quantised versions of Basic Convolutional Neural Network (B-CNN) and MobileNetV2 have better malaria detection performance (up to 99% recall), lower memory usage (2MB 8-bit quantised model) and shorter inference time (33–95 microseconds on mobile phones) than VGG-19 fine-tuned and quantised models. Hence, we have implemented MobileNetV2 in our mobile application as it has even a lower memory requirement than B-CNN. This work will help to counter the negative effects of COVID-19 on the previous successes towards global malaria elimination. |
format |
Text |
author |
Peter U. Eze Clement O. Asogwa |
author_facet |
Peter U. Eze Clement O. Asogwa |
author_sort |
Peter U. Eze |
title |
Deep Machine Learning Model Trade-Offs for Malaria Elimination in Resource-Constrained Locations |
title_short |
Deep Machine Learning Model Trade-Offs for Malaria Elimination in Resource-Constrained Locations |
title_full |
Deep Machine Learning Model Trade-Offs for Malaria Elimination in Resource-Constrained Locations |
title_fullStr |
Deep Machine Learning Model Trade-Offs for Malaria Elimination in Resource-Constrained Locations |
title_full_unstemmed |
Deep Machine Learning Model Trade-Offs for Malaria Elimination in Resource-Constrained Locations |
title_sort |
deep machine learning model trade-offs for malaria elimination in resource-constrained locations |
publisher |
Multidisciplinary Digital Publishing Institute |
publishDate |
2021 |
url |
https://doi.org/10.3390/bioengineering8110150 |
op_coverage |
agris |
genre |
DML |
genre_facet |
DML |
op_source |
Bioengineering; Volume 8; Issue 11; Pages: 150 |
op_relation |
https://dx.doi.org/10.3390/bioengineering8110150 |
op_rights |
https://creativecommons.org/licenses/by/4.0/ |
op_doi |
https://doi.org/10.3390/bioengineering8110150 |
container_title |
Bioengineering |
container_volume |
8 |
container_issue |
11 |
container_start_page |
150 |
_version_ |
1774717088106020864 |