File Download
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.3390/jimaging3040066
- Scopus: eid_2-s2.0-85050406708
- WOS: WOS:000424411800027
Supplementary
- Citations:
- Appears in Collections:
Article: Deep learning vs. conventional machine learning: Pilot study of WMH segmentation in brain MRI with absence or mild vascular pathology
Title | Deep learning vs. conventional machine learning: Pilot study of WMH segmentation in brain MRI with absence or mild vascular pathology |
---|---|
Authors | |
Keywords | Conventional machine learning Dementia Deep learning Brain MRI Medical image analysis Segmentation Machine learning Alzheimer's Disease White matter hyperintensities |
Issue Date | 2017 |
Citation | Journal of Imaging, 2017, v. 3, n. 4, article no. 66 How to Cite? |
Abstract | © 2017 by the authors. In the wake of the use of deep learning algorithms in medical image analysis, we compared performance of deep learning algorithms, namely the deep Boltzmann machine (DBM), convolutional encoder network (CEN) and patch-wise convolutional neural network (patch-CNN), with two conventional machine learning schemes: Support vector machine (SVM) and random forest (RF), for white matter hyperintensities (WMH) segmentation on brain MRI with mild or no vascular pathology. We also compared all these approaches with a method in the Lesion Segmentation Tool public toolbox named lesion growth algorithm (LGA). We used a dataset comprised of 60 MRI data from 20 subjects in the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, each scanned once every year during three consecutive years. Spatial agreement score, receiver operating characteristic and precision-recall performance curves, volume disagreement score, agreement with intra-/inter-observer reliability measurements and visual evaluation were used to find the best configuration of each learning algorithm for WMH segmentation. By using optimum threshold values for the probabilistic output from each algorithm to produce binary masks of WMH, we found that SVM and RF produced good results for medium to very large WMH burden but deep learning algorithms performed generally better than conventional ones in most evaluations. |
Persistent Identifier | http://hdl.handle.net/10722/288581 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Rachmadi, Muhammad Febrian | - |
dc.contributor.author | Del C. Valdés-Hernández, Maria | - |
dc.contributor.author | Agan, Maria Leonora Fatimah | - |
dc.contributor.author | Komura, Taku | - |
dc.date.accessioned | 2020-10-12T08:05:20Z | - |
dc.date.available | 2020-10-12T08:05:20Z | - |
dc.date.issued | 2017 | - |
dc.identifier.citation | Journal of Imaging, 2017, v. 3, n. 4, article no. 66 | - |
dc.identifier.uri | http://hdl.handle.net/10722/288581 | - |
dc.description.abstract | © 2017 by the authors. In the wake of the use of deep learning algorithms in medical image analysis, we compared performance of deep learning algorithms, namely the deep Boltzmann machine (DBM), convolutional encoder network (CEN) and patch-wise convolutional neural network (patch-CNN), with two conventional machine learning schemes: Support vector machine (SVM) and random forest (RF), for white matter hyperintensities (WMH) segmentation on brain MRI with mild or no vascular pathology. We also compared all these approaches with a method in the Lesion Segmentation Tool public toolbox named lesion growth algorithm (LGA). We used a dataset comprised of 60 MRI data from 20 subjects in the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, each scanned once every year during three consecutive years. Spatial agreement score, receiver operating characteristic and precision-recall performance curves, volume disagreement score, agreement with intra-/inter-observer reliability measurements and visual evaluation were used to find the best configuration of each learning algorithm for WMH segmentation. By using optimum threshold values for the probabilistic output from each algorithm to produce binary masks of WMH, we found that SVM and RF produced good results for medium to very large WMH burden but deep learning algorithms performed generally better than conventional ones in most evaluations. | - |
dc.language | eng | - |
dc.relation.ispartof | Journal of Imaging | - |
dc.rights | This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. | - |
dc.subject | Conventional machine learning | - |
dc.subject | Dementia | - |
dc.subject | Deep learning | - |
dc.subject | Brain MRI | - |
dc.subject | Medical image analysis | - |
dc.subject | Segmentation | - |
dc.subject | Machine learning | - |
dc.subject | Alzheimer's Disease | - |
dc.subject | White matter hyperintensities | - |
dc.title | Deep learning vs. conventional machine learning: Pilot study of WMH segmentation in brain MRI with absence or mild vascular pathology | - |
dc.type | Article | - |
dc.description.nature | published_or_final_version | - |
dc.identifier.doi | 10.3390/jimaging3040066 | - |
dc.identifier.scopus | eid_2-s2.0-85050406708 | - |
dc.identifier.volume | 3 | - |
dc.identifier.issue | 4 | - |
dc.identifier.spage | article no. 66 | - |
dc.identifier.epage | article no. 66 | - |
dc.identifier.eissn | 2313-433X | - |
dc.identifier.isi | WOS:000424411800027 | - |
dc.identifier.issnl | 2313-433X | - |