Deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling

Automated segmentation is important for early detection and treatments to reduce disability and death risks among brain stroke patients. The existing segmentation algorithm is limited due to its computationally expensiveness in achieving a small accuracy. This work aims to develop a computationally...

Full description

Bibliographic Details
Main Authors: Muhammad Ismaill, Mat Lizah, Nasrul Hadi, Johari, Mohd Jamil, Mohamed Mokhtarudin
Format: Article
Language:English
Published: Faculty Mechanical Engineering, UMP 2025
Subjects:
Online Access:http://umpir.ump.edu.my/id/eprint/44249/
http://umpir.ump.edu.my/id/eprint/44249/1/7.%2Blizah%2Bet%2Bal._0824.pdf
_version_ 1848827062138175488
author Muhammad Ismaill, Mat Lizah
Nasrul Hadi, Johari
Mohd Jamil, Mohamed Mokhtarudin
author_facet Muhammad Ismaill, Mat Lizah
Nasrul Hadi, Johari
Mohd Jamil, Mohamed Mokhtarudin
author_sort Muhammad Ismaill, Mat Lizah
building UMP Institutional Repository
collection Online Access
description Automated segmentation is important for early detection and treatments to reduce disability and death risks among brain stroke patients. The existing segmentation algorithm is limited due to its computationally expensiveness in achieving a small accuracy. This work aims to develop a computationally economical automated brain infarct segmentation from T1-weighted Magnetic Resonance Imaging (MRI) using convolutional neural network architecture U-Net, but with reasonable accuracy compared to existing algorithm. The data used is taken from the Anatomical Tracing of Lesion After Stroke (ATLAS) open-source dataset, consisting of 304 brain t1-weigthed MRI images. The data is divided into training, test, and validation sets according to the 8:1:1 ratio. The data is then pre-processed so that all of them have similar size for the U-Net input. Then, the U-Net architecture is generated using encoder depth of 7. Certain hyperparameters including the number of epochs, encoder depth, and optimizers are varied. The U-Net with encoder depth 7 and using Adam optimizer gives the highest accuracy and loss, which are 92.33% and 0.9771, respectively. Further comparison with previous works shows that the present U-Net beaten the regular U-Net and also gives relatively similar accuracy and loss. Future improvements on the present U-Net is necessary so that the accuracy can be increased further, computationally economic, and to produce a near accurate semantic segmentation of brain lesion.
first_indexed 2025-11-15T03:54:44Z
format Article
id ump-44249
institution Universiti Malaysia Pahang
institution_category Local University
language English
last_indexed 2025-11-15T03:54:44Z
publishDate 2025
publisher Faculty Mechanical Engineering, UMP
recordtype eprints
repository_type Digital Repository
spelling ump-442492025-04-15T01:55:50Z http://umpir.ump.edu.my/id/eprint/44249/ Deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling Muhammad Ismaill, Mat Lizah Nasrul Hadi, Johari Mohd Jamil, Mohamed Mokhtarudin TJ Mechanical engineering and machinery Automated segmentation is important for early detection and treatments to reduce disability and death risks among brain stroke patients. The existing segmentation algorithm is limited due to its computationally expensiveness in achieving a small accuracy. This work aims to develop a computationally economical automated brain infarct segmentation from T1-weighted Magnetic Resonance Imaging (MRI) using convolutional neural network architecture U-Net, but with reasonable accuracy compared to existing algorithm. The data used is taken from the Anatomical Tracing of Lesion After Stroke (ATLAS) open-source dataset, consisting of 304 brain t1-weigthed MRI images. The data is divided into training, test, and validation sets according to the 8:1:1 ratio. The data is then pre-processed so that all of them have similar size for the U-Net input. Then, the U-Net architecture is generated using encoder depth of 7. Certain hyperparameters including the number of epochs, encoder depth, and optimizers are varied. The U-Net with encoder depth 7 and using Adam optimizer gives the highest accuracy and loss, which are 92.33% and 0.9771, respectively. Further comparison with previous works shows that the present U-Net beaten the regular U-Net and also gives relatively similar accuracy and loss. Future improvements on the present U-Net is necessary so that the accuracy can be increased further, computationally economic, and to produce a near accurate semantic segmentation of brain lesion. Faculty Mechanical Engineering, UMP 2025-03 Article PeerReviewed pdf en cc_by_nc_4 http://umpir.ump.edu.my/id/eprint/44249/1/7.%2Blizah%2Bet%2Bal._0824.pdf Muhammad Ismaill, Mat Lizah and Nasrul Hadi, Johari and Mohd Jamil, Mohamed Mokhtarudin (2025) Deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling. Journal of Mechanical Engineering and Sciences (JMES), 19 (1). pp. 10521-10529. ISSN 2289-4659 (print); 2231-8380 (online). (Published) https://doi.org/10.15282/jmes.19.1.2025.7.0824 10.15282/jmes.19.1.2025.7.0824
spellingShingle TJ Mechanical engineering and machinery
Muhammad Ismaill, Mat Lizah
Nasrul Hadi, Johari
Mohd Jamil, Mohamed Mokhtarudin
Deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling
title Deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling
title_full Deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling
title_fullStr Deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling
title_full_unstemmed Deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling
title_short Deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling
title_sort deep learning segmentation of brain ischemic lesion from magnetic resonance images for three-dimensional modelling
topic TJ Mechanical engineering and machinery
url http://umpir.ump.edu.my/id/eprint/44249/
http://umpir.ump.edu.my/id/eprint/44249/
http://umpir.ump.edu.my/id/eprint/44249/
http://umpir.ump.edu.my/id/eprint/44249/1/7.%2Blizah%2Bet%2Bal._0824.pdf