Multiple Sclerosis Lesion Synthesis in MRI Using an Encoder-Decoder U-NET
dc.contributor.author
dc.date.accessioned
2020-06-04T09:44:24Z
dc.date.available
2020-06-04T09:44:24Z
dc.date.issued
2019-02-20
dc.identifier.issn
2169-3536
dc.identifier.uri
dc.description.abstract
In this paper, we propose generating synthetic multiple sclerosis (MS) lesions on MRI images with the final aim to
improve the performance of supervised machine learning algorithms, therefore avoiding the problem of the lack of
available ground truth. We propose a two-input two-output fully convolutional neural network model for MS lesion
synthesis in MRI images. The lesion information is encoded as discrete binary intensity level masks passed to the
model and stacked with the input images. The model is trained end-to-end without the need for manually annotating
the lesions in the training set. We then perform the generation of synthetic lesions on healthy images via registration
of patient images, which are subsequently used for data augmentation to increase the performance for supervised MS
lesion detection algorithms. Our pipeline is evaluated on MS patient data from an in-house clinical dataset and the
public ISBI2015 challenge dataset. The evaluation is based on measuring the similarities between the real and the
synthetic images as well as in terms of lesion detection performance by segmenting both the original and synthetic
images individually using a state-of-the-art segmentation framework. We also demonstrate the usage of synthetic MS
lesions generated on healthy images as data augmentation. We analyze a scenario of limited training data (one-image
training) to demonstrate the effect of the data augmentation on both datasets. Our results significantly show the
effectiveness of the usage of synthetic MS lesion images. For the ISBI2015 challenge, our one-image model trained
using only a single image plus the synthetic data augmentation strategy showed a performance similar to that of
other CNN
dc.description.sponsorship
This work was supported by the Retos de Investigación from the Ministerio de Ciencia y Tecnología under Grant TIN2014- 55710-R and
Grant DPI2017-86696-R
dc.format.mimetype
application/pdf
dc.language.iso
eng
dc.publisher
Institute of Electrical and Electronics Engineers (IEEE)
dc.relation
info:eu-repo/grantAgreement/MINECO//TIN2014-55710-R/ES/HERRAMIENTAS DE NEUROIMAGEN PARA MEJORAR EL DIAGNOSIS Y EL SEGUIMIENTO CLINICO DE LOS PACIENTES CON ESCLEROSIS MULTIPLE/
info:eu-repo/grantAgreement/AEI/Plan Estatal de Investigación Científica y Técnica y de Innovación 2013-2016/DPI2017-86696-R/ES/MODELOS PREDICTIVOS PARA LA ESCLEROSIS MULTIPE USANDO BIOMARCADORES DE RESONANCIA MAGNETICA DEL CEREBRO/
dc.relation.isformatof
Versió preprint del document publicat a: 10.1109/ACCESS.2019.2900198
dc.relation.ispartof
IEEE Access, 2019, vol. 7, p. 25171 - 25184
dc.relation.ispartofseries
Articles publicats (D-ATC)
dc.rights
Tots els drets reservats
dc.subject
dc.title
Multiple Sclerosis Lesion Synthesis in MRI Using an Encoder-Decoder U-NET
dc.type
info:eu-repo/semantics/article
dc.rights.accessRights
info:eu-repo/semantics/openAccess
dc.type.version
info:eu-repo/semantics/submittedVersion
dc.identifier.doi
dc.identifier.idgrec
29685
dc.contributor.funder
dc.relation.FundingProgramme
dc.relation.ProjectAcronym