Abstract
Detecting stroke is important to reduce the likelihood of permanent disability and increase the chance of recovery. Brain stroke lesion segmentation is an important procedure, especially when a specific brain portion needs to be analyzed. In this project, a brain stroke lesion segmentation algorithm using a modified U-Net (MUN) architecture will be developed. The MUN has a dimension-fusion capability, in which the images are analyzed separately using 2D U-Net and 3D image downsampling processes, before being fused at two points during the downsampling processes. The MUN accuracy is then compared with a regular 3D U-Net (UN). Three training options are further developed and compared. It is found that the MUN architecture produces higher training accuracy, but slower training duration compared to UN. Despite the capabilities of MUN, it cannot be further validated due to software limitations. Further improvement on the algorithm using other libraries is essential to enhance the capability of the MUN.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Vilela P, Rowley HA (2017) Brain ischemia: CT and MRI techniques in acute ischemic stroke. Eur J Radiol 96:162–172
Mohamed Mokhtarudin MJ (2016) Mathematical modelling of cerebral ischaemia-reperfusion injury. University of Oxford, Oxford
Nadzri AN, Mokhtarudin MJM et al (2020) Simulation of decompressive craniectomy for ischaemic stroke treatment: a conceptual modeling study. In: 2020 IEEE-EMBS Conference on Biomedical Engineering and Sciences (IECBES)
Nadzri AN, Mohamed Mokhtarudin MJ et al (2021) Simulation of craniectomy size in decompressive craniectomy for ischaemic stroke. In: Innovative Manufacturing, Mechatronics and Materials Forum 2021 (iM3F 2021). Springer Nature, Singapore
Shin H, Agyeman R, Rafiq M, Chang MC, Choi GS (2022) Automated segmentation of chronic stroke lesion using efficient U-Net architecture. Biocybern Biomed Eng 42(1):285–294
Liew S-L, Anglin JM et al (2018) A large, open source dataset of stroke anatomical brain images and manual lesion segmentations. Sci Data 5(1):180011
Ronneberger O, Fischer P et al (2015) U-Net: convolutional networks for biomedical image segmentation. In: Medical image computing and computer-assisted intervention—MICCAI 2015. Springer International Publishing, Cham
Zhou Y, Huang W et al (2021) D-UNet: a dimension-fusion U shape network for chronic stroke lesion segmentation. IEEE/ACM Trans Comput Biol Bioinf 18(3):940–950
Wilson, D.R. and T.R. Martinez. (2001) The need for small learning rates on large problems. in International Joint Conference on Neural Networks
Raschka S, Patterson J et al (2020) Machine learning in python: main developments and technology trends in data science, machine learning, and artificial intelligence. Information 11(4):193Â
Acknowledgements
This research is funded by the Fundamental Research Grant Scheme from Malaysia Ministry of Higher Education grant (grant no: FRGS/1/2018/TK03/UMP/02/15).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Lee, X.Y., Mohamed Mokhtarudin, M.J., Junid, R. (2024). Brain Lesion Image Segmentation Using Modified U-NET Architecture. In: Mohd. Isa, W.H., Khairuddin, I.M., Mohd. Razman, M.A., Saruchi, S.'., Teh, SH., Liu, P. (eds) Intelligent Manufacturing and Mechatronics. iM3F 2023. Lecture Notes in Networks and Systems, vol 850. Springer, Singapore. https://doi.org/10.1007/978-981-99-8819-8_46
Download citation
DOI: https://doi.org/10.1007/978-981-99-8819-8_46
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-99-8818-1
Online ISBN: 978-981-99-8819-8
eBook Packages: EngineeringEngineering (R0)