Vis enkel innførsel

dc.contributor.authorFossen-Romsaas, Sondre
dc.contributor.authorStorm-Johannessen, Adrian
dc.contributor.authorLundervold, Alexander Selvikvåg
dc.date.accessioned2021-01-13T08:21:07Z
dc.date.available2021-01-13T08:21:07Z
dc.date.created2020-11-26T05:34:08Z
dc.date.issued2020
dc.identifier.citationFossen-Romsaas, S., Storm-Johannessen, A., & Lundervold, A. S. (2020). Synthesizing skin lesion images using CycleGANs – a case study. NIK – Norsk Informatikkonferanse, 12.en_US
dc.identifier.urihttps://hdl.handle.net/11250/2722685
dc.description.abstractGenerative adversarial networks (GANs) have seen some success as a way to synthesize training data for supervised machine learning models. In this work, we design two novel approaches for synthetic image generation based on CycleGANs, aimed at generating realistic-looking, class-specific dermoscopic skin lesion images. We evaluate the images’ usefulness as additional training data for a convolutional neural network trained to perform a difficult lesion classification task. We are able to generate visually striking images, but their value for augmenting the classifier’s training data set is low. This is in-line with other researcher’s investigations into similar GAN models, indicating the need for further research into forcing GAN models to produce samples further from the training data distribution, and to find ways of guiding the image generation using feedback from the ultimate classification objective.en_US
dc.language.isoengen_US
dc.publisherBibsys Open Journal Systemsen_US
dc.titleSynthesizing skin lesion images using CycleGANs – a case studyen_US
dc.typeJournal articleen_US
dc.description.versionpublishedVersionen_US
dc.source.journalNorsk Informatikkonferanseen_US
dc.identifier.cristin1852551
dc.relation.projectBergens forskningsstiftelse: BFS2018TMT07en_US
cristin.ispublishedtrue
cristin.fulltextoriginal


Tilhørende fil(er)

Thumbnail

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel