Please use this identifier to cite or link to this item: https://scidar.kg.ac.rs/handle/123456789/13640
Full metadata record
DC FieldValueLanguage
dc.rights.licenseBY-NC-ND-
dc.contributor.authorVasiljević, Jelica-
dc.contributor.authorFeuerhake F.-
dc.contributor.authorWemmert, Cedric-
dc.contributor.authorLampert, Thomas-
dc.date.accessioned2021-09-24T23:09:43Z-
dc.date.available2021-09-24T23:09:43Z-
dc.date.issued2021-
dc.identifier.issn1945-7928-
dc.identifier.urihttps://scidar.kg.ac.rs/handle/123456789/13640-
dc.description.abstractIt has been shown that unpaired image-to-image translation methods constrained by cycle-consistency hide the information necessary for accurate input reconstruction as imperceptible noise. We demonstrate that, when applied to histopathology data, this hidden noise appears to be related to stain specific features and show that this is the case with two immune-histochemical stainings during translation to Periodic acid-Schiff(PAS), a histochemical staining method commonly applied in renal pathology. Moreover, by perturbing this hidden information, the translation models produce different, plausible outputs. We demonstrate that this property can be used as an augmentation method which, in a case of supervised glomeruli segmentation, leads to improved performance.-
dc.rightsopenAccess-
dc.rights.urihttps://creativecommons.org/licenses/by-nc-nd/4.0/-
dc.sourceProceedings - International Symposium on Biomedical Imaging-
dc.titleSelf adversarial attack as an augmentation method for immunohistochemical stainings-
dc.typeconferenceObject-
dc.identifier.doi10.1109/ISBI48211.2021.9433838-
dc.identifier.scopus2-s2.0-85107214473-
Appears in Collections:Faculty of Science, Kragujevac

Page views(s)

124

Downloads(s)

12

Files in This Item:
File Description SizeFormat 
10.1109-ISBI48211.2021.9433838.pdf6.23 MBAdobe PDFThumbnail
View/Open


This item is licensed under a Creative Commons License Creative Commons