001033913 001__ 1033913
001033913 005__ 20250310131247.0
001033913 0247_ $$2doi$$a10.1109/CVPRW63382.2024.00012
001033913 0247_ $$2datacite_doi$$a10.34734/FZJ-2024-06751
001033913 0247_ $$2WOS$$aWOS:001327781700008
001033913 037__ $$aFZJ-2024-06751
001033913 1001_ $$0P:(DE-Juel1)196697$$aKazimi, Bashir$$b0
001033913 1112_ $$a2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)$$cSeattle$$d2024-06-17 - 2024-06-18$$wWA
001033913 245__ $$aSelf-Supervised Learning with Generative Adversarial Networks for Electron Microscopy
001033913 260__ $$bIEEE$$c2024
001033913 300__ $$a71-81
001033913 3367_ $$2ORCID$$aCONFERENCE_PAPER
001033913 3367_ $$033$$2EndNote$$aConference Paper
001033913 3367_ $$2BibTeX$$aINPROCEEDINGS
001033913 3367_ $$2DRIVER$$aconferenceObject
001033913 3367_ $$2DataCite$$aOutput Types/Conference Paper
001033913 3367_ $$0PUB:(DE-HGF)8$$2PUB:(DE-HGF)$$aContribution to a conference proceedings$$bcontrib$$mcontrib$$s1733987837_21022
001033913 520__ $$aIn this work, we explore the potential of self-supervised learning with Generative Adversarial Networks (GANs) for electron microscopy datasets. We show how self-supervised pretraining facilitates efficient fine-tuning for a spectrum of downstream tasks, including semantic segmentation, denoising, noise & background removal, and super-resolution. Experimentation with varying model complexities and receptive field sizes reveals the remarkable phenomenon that fine-tuned models of lower complexity consistently outperform more complex models with random weight initialization. We demonstrate the versatility of self-supervised pretraining across various downstream tasks in the context of electron microscopy, allowing faster convergence and better performance. We conclude that self-supervised pretraining serves as a powerful catalyst, being especially advantageous when limited annotated data are available and efficient scaling of computational cost is important.
001033913 536__ $$0G:(DE-HGF)POF4-5111$$a5111 - Domain-Specific Simulation & Data Life Cycle Labs (SDLs) and Research Groups (POF4-511)$$cPOF4-511$$fPOF IV$$x0
001033913 588__ $$aDataset connected to CrossRef Conference
001033913 7001_ $$0P:(DE-Juel1)180323$$aRuzaeva, Karina$$b1$$ufzj
001033913 7001_ $$0P:(DE-Juel1)186075$$aSandfeld, Stefan$$b2$$eCorresponding author
001033913 773__ $$a10.1109/CVPRW63382.2024.00012
001033913 8564_ $$uhttps://ieeexplore.ieee.org/document/10678518/authors#authors
001033913 8564_ $$uhttps://juser.fz-juelich.de/record/1033913/files/2402.18286v2.pdf$$yOpenAccess
001033913 909CO $$ooai:juser.fz-juelich.de:1033913$$pdnbdelivery$$pdriver$$pVDB$$popen_access$$popenaire
001033913 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)196697$$aForschungszentrum Jülich$$b0$$kFZJ
001033913 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)180323$$aForschungszentrum Jülich$$b1$$kFZJ
001033913 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)186075$$aForschungszentrum Jülich$$b2$$kFZJ
001033913 9131_ $$0G:(DE-HGF)POF4-511$$1G:(DE-HGF)POF4-510$$2G:(DE-HGF)POF4-500$$3G:(DE-HGF)POF4$$4G:(DE-HGF)POF$$9G:(DE-HGF)POF4-5111$$aDE-HGF$$bKey Technologies$$lEngineering Digital Futures – Supercomputing, Data Management and Information Security for Knowledge and Action$$vEnabling Computational- & Data-Intensive Science and Engineering$$x0
001033913 9141_ $$y2024
001033913 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess
001033913 920__ $$lyes
001033913 9201_ $$0I:(DE-Juel1)IAS-9-20201008$$kIAS-9$$lMaterials Data Science and Informatics$$x0
001033913 980__ $$acontrib
001033913 980__ $$aVDB
001033913 980__ $$aUNRESTRICTED
001033913 980__ $$aI:(DE-Juel1)IAS-9-20201008
001033913 9801_ $$aFullTexts