Publicação: A Fully Unsupervised Deep Learning Framework for Non-Rigid Fundus Image Registration
dc.contributor.author | Benvenuto, Giovana A. [UNESP] | |
dc.contributor.author | Colnago, Marilaine | |
dc.contributor.author | Dias, Maurício A. [UNESP] | |
dc.contributor.author | Negri, Rogério G. [UNESP] | |
dc.contributor.author | Silva, Erivaldo A. [UNESP] | |
dc.contributor.author | Casaca, Wallace [UNESP] | |
dc.contributor.institution | Universidade Estadual Paulista (UNESP) | |
dc.contributor.institution | Universidade de São Paulo (USP) | |
dc.date.accessioned | 2023-03-01T21:13:09Z | |
dc.date.available | 2023-03-01T21:13:09Z | |
dc.date.issued | 2022-08-01 | |
dc.description.abstract | In ophthalmology, the registration problem consists of finding a geometric transformation that aligns a pair of images, supporting eye-care specialists who need to record and compare images of the same patient. Considering the registration methods for handling eye fundus images, the literature offers only a limited number of proposals based on deep learning (DL), whose implementations use the supervised learning paradigm to train a model. Additionally, ensuring high-quality registrations while still being flexible enough to tackle a broad range of fundus images is another drawback faced by most existing methods in the literature. Therefore, in this paper, we address the above-mentioned issues by introducing a new DL-based framework for eye fundus registration. Our methodology combines a U-shaped fully convolutional neural network with a spatial transformation learning scheme, where a reference-free similarity metric allows the registration without assuming any pre-annotated or artificially created data. Once trained, the model is able to accurately align pairs of images captured under several conditions, which include the presence of anatomical differences and low-quality photographs. Compared to other registration methods, our approach achieves better registration outcomes by just passing as input the desired pair of fundus images. | en |
dc.description.affiliation | Faculty of Science and Technology (FCT) São Paulo State University (UNESP) | |
dc.description.affiliation | Institute of Mathematics and Computer Science (ICMC) São Paulo University (USP) | |
dc.description.affiliation | Science and Technology Institute (ICT) São Paulo State University (UNESP) | |
dc.description.affiliation | Institute of Biosciences Letters and Exact Sciences (IBILCE) São Paulo State University (UNESP) | |
dc.description.affiliationUnesp | Faculty of Science and Technology (FCT) São Paulo State University (UNESP) | |
dc.description.affiliationUnesp | Science and Technology Institute (ICT) São Paulo State University (UNESP) | |
dc.description.affiliationUnesp | Institute of Biosciences Letters and Exact Sciences (IBILCE) São Paulo State University (UNESP) | |
dc.identifier | http://dx.doi.org/10.3390/bioengineering9080369 | |
dc.identifier.citation | Bioengineering, v. 9, n. 8, 2022. | |
dc.identifier.doi | 10.3390/bioengineering9080369 | |
dc.identifier.issn | 2306-5354 | |
dc.identifier.scopus | 2-s2.0-85137360100 | |
dc.identifier.uri | http://hdl.handle.net/11449/241615 | |
dc.language.iso | eng | |
dc.relation.ispartof | Bioengineering | |
dc.source | Scopus | |
dc.subject | computer vision applications | |
dc.subject | deep learning | |
dc.subject | fundus image | |
dc.subject | image registration | |
dc.title | A Fully Unsupervised Deep Learning Framework for Non-Rigid Fundus Image Registration | en |
dc.type | Artigo | |
dspace.entity.type | Publication | |
unesp.author.orcid | 0000-0002-0531-1284[1] | |
unesp.author.orcid | 0000-0003-1599-491X[2] | |
unesp.author.orcid | 0000-0002-1361-6184[3] | |
unesp.author.orcid | 0000-0002-4808-2362[4] | |
unesp.author.orcid | 0000-0002-7069-0479[5] | |
unesp.author.orcid | 0000-0002-1073-9939[6] | |
unesp.department | Estatística - FCT | pt |