File Download
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.3389/fnagi.2019.00150
- Scopus: eid_2-s2.0-85069154565
- PMID: 31316369
- WOS: WOS:000473096100001
Supplementary
- Citations:
- Appears in Collections:
Article: Dilated Saliency U-Net for white matter hyperintensities segmentation using Irregularity age map
Title | Dilated Saliency U-Net for white matter hyperintensities segmentation using Irregularity age map |
---|---|
Authors | |
Keywords | White matter hyperintensities Irregularity age map Deep learning Dilated convolution Segmentation Saliency U-Net MRI |
Issue Date | 2019 |
Citation | Frontiers in Aging Neuroscience, 2019, v. 11, article no. 150 How to Cite? |
Abstract | © 2019 Frontiers Media S.A. All Rights Reserved. White matter hyperintensities (WMH) appear as regions of abnormally high signal intensity on T2-weighted magnetic resonance image (MRI) sequences. In particular, WMH have been noteworthy in age-related neuroscience for being a crucial biomarker for all types of dementia and brain aging processes. The automatic WMH segmentation is challenging because of their variable intensity range, size and shape. U-Net tackles this problem through the dense prediction and has shown competitive performances not only on WMH segmentation/detection but also on varied image segmentation tasks. However, its network architecture is high complex. In this study, we propose the use of Saliency U-Net and Irregularity map (IAM) to decrease the U-Net architectural complexity without performance loss. We trained Saliency U-Net using both: a T2-FLAIR MRI sequence and its correspondent IAM. Since IAM guides locating image intensity irregularities, in which WMH are possibly included, in the MRI slice, Saliency U-Net performs better than the original U-Net trained only using T2-FLAIR. The best performance was achieved with fewer parameters and shorter training time. Moreover, the application of dilated convolution enhanced Saliency U-Net by recognizing the shape of large WMH more accurately through multi-context learning. This network named Dilated Saliency U-Net improved Dice coefficient score to 0.5588 which was the best score among our experimental models, and recorded a relatively good sensitivity of 0.4747 with the shortest training time and the least number of parameters. In conclusion, based on our experimental results, incorporating IAM through Dilated Saliency U-Net resulted an appropriate approach for WMH segmentation. |
Persistent Identifier | http://hdl.handle.net/10722/288950 |
PubMed Central ID | |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Jeong, Yunhee | - |
dc.contributor.author | Rachmadi, Muhammad Febrian | - |
dc.contributor.author | Valdés-Hernández, Maria Del C. | - |
dc.contributor.author | Komura, Taku | - |
dc.date.accessioned | 2020-10-12T08:06:17Z | - |
dc.date.available | 2020-10-12T08:06:17Z | - |
dc.date.issued | 2019 | - |
dc.identifier.citation | Frontiers in Aging Neuroscience, 2019, v. 11, article no. 150 | - |
dc.identifier.uri | http://hdl.handle.net/10722/288950 | - |
dc.description.abstract | © 2019 Frontiers Media S.A. All Rights Reserved. White matter hyperintensities (WMH) appear as regions of abnormally high signal intensity on T2-weighted magnetic resonance image (MRI) sequences. In particular, WMH have been noteworthy in age-related neuroscience for being a crucial biomarker for all types of dementia and brain aging processes. The automatic WMH segmentation is challenging because of their variable intensity range, size and shape. U-Net tackles this problem through the dense prediction and has shown competitive performances not only on WMH segmentation/detection but also on varied image segmentation tasks. However, its network architecture is high complex. In this study, we propose the use of Saliency U-Net and Irregularity map (IAM) to decrease the U-Net architectural complexity without performance loss. We trained Saliency U-Net using both: a T2-FLAIR MRI sequence and its correspondent IAM. Since IAM guides locating image intensity irregularities, in which WMH are possibly included, in the MRI slice, Saliency U-Net performs better than the original U-Net trained only using T2-FLAIR. The best performance was achieved with fewer parameters and shorter training time. Moreover, the application of dilated convolution enhanced Saliency U-Net by recognizing the shape of large WMH more accurately through multi-context learning. This network named Dilated Saliency U-Net improved Dice coefficient score to 0.5588 which was the best score among our experimental models, and recorded a relatively good sensitivity of 0.4747 with the shortest training time and the least number of parameters. In conclusion, based on our experimental results, incorporating IAM through Dilated Saliency U-Net resulted an appropriate approach for WMH segmentation. | - |
dc.language | eng | - |
dc.relation.ispartof | Frontiers in Aging Neuroscience | - |
dc.rights | This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. | - |
dc.subject | White matter hyperintensities | - |
dc.subject | Irregularity age map | - |
dc.subject | Deep learning | - |
dc.subject | Dilated convolution | - |
dc.subject | Segmentation | - |
dc.subject | Saliency U-Net | - |
dc.subject | MRI | - |
dc.title | Dilated Saliency U-Net for white matter hyperintensities segmentation using Irregularity age map | - |
dc.type | Article | - |
dc.description.nature | published_or_final_version | - |
dc.identifier.doi | 10.3389/fnagi.2019.00150 | - |
dc.identifier.pmid | 31316369 | - |
dc.identifier.pmcid | PMC6610522 | - |
dc.identifier.scopus | eid_2-s2.0-85069154565 | - |
dc.identifier.volume | 11 | - |
dc.identifier.spage | article no. 150 | - |
dc.identifier.epage | article no. 150 | - |
dc.identifier.eissn | 1663-4365 | - |
dc.identifier.isi | WOS:000473096100001 | - |
dc.identifier.issnl | 1663-4365 | - |