Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313010918> ?p ?o ?g. }
Showing items 1 to 80 of
80
with 100 items per page.
- W4313010918 abstract "In generative adversarial networks, improving discriminators is one of the key components for generation performance. As image classifiers are biased toward texture and debiasing improves accuracy, we investigate 1) if the discriminators are biased, and 2) if debiasing the discriminators will improve generation performance. Indeed, we find empirical evidence that the discriminators are sensitive to the style (e.g., texture and color) of images. As a remedy, we propose feature statistics mixing regularization (FSMR) that encourages the discriminator's prediction to be invariant to the styles of input images. Specifically, we generate a mixed feature of an original and a reference image in the discriminator's feature space and we apply regularization so that the prediction for the mixed feature is consistent with the prediction for the original image. We conduct extensive experiments to demonstrate that our regularization leads to reduced sensitivity to style and consistently improves the performance of various GAN architectures on nine datasets. In addition, adding FSMR to recently-proposed augmentation-based GAN methods further improves image quality. Our code is available at https://github.com/naver-ai/FSMR." @default.
- W4313010918 created "2023-01-05" @default.
- W4313010918 creator A5008789470 @default.
- W4313010918 creator A5029605633 @default.
- W4313010918 creator A5073514062 @default.
- W4313010918 date "2022-06-01" @default.
- W4313010918 modified "2023-10-14" @default.
- W4313010918 title "Feature Statistics Mixing Regularization for Generative Adversarial Networks" @default.
- W4313010918 cites W2603777577 @default.
- W4313010918 cites W2962770929 @default.
- W4313010918 cites W2962793481 @default.
- W4313010918 cites W2962974533 @default.
- W4313010918 cites W2963767194 @default.
- W4313010918 cites W2965141076 @default.
- W4313010918 cites W2987285147 @default.
- W4313010918 cites W3034600949 @default.
- W4313010918 cites W3035574324 @default.
- W4313010918 cites W3096831136 @default.
- W4313010918 cites W3166541011 @default.
- W4313010918 cites W3175491752 @default.
- W4313010918 doi "https://doi.org/10.1109/cvpr52688.2022.01101" @default.
- W4313010918 hasPublicationYear "2022" @default.
- W4313010918 type Work @default.
- W4313010918 citedByCount "6" @default.
- W4313010918 countsByYear W43130109182022 @default.
- W4313010918 countsByYear W43130109182023 @default.
- W4313010918 crossrefType "proceedings-article" @default.
- W4313010918 hasAuthorship W4313010918A5008789470 @default.
- W4313010918 hasAuthorship W4313010918A5029605633 @default.
- W4313010918 hasAuthorship W4313010918A5073514062 @default.
- W4313010918 hasBestOaLocation W43130109182 @default.
- W4313010918 hasConcept C138885662 @default.
- W4313010918 hasConcept C153180895 @default.
- W4313010918 hasConcept C154945302 @default.
- W4313010918 hasConcept C15744967 @default.
- W4313010918 hasConcept C188147891 @default.
- W4313010918 hasConcept C2776135515 @default.
- W4313010918 hasConcept C2776401178 @default.
- W4313010918 hasConcept C2779458634 @default.
- W4313010918 hasConcept C2779803651 @default.
- W4313010918 hasConcept C37736160 @default.
- W4313010918 hasConcept C39890363 @default.
- W4313010918 hasConcept C41008148 @default.
- W4313010918 hasConcept C41895202 @default.
- W4313010918 hasConcept C76155785 @default.
- W4313010918 hasConcept C94915269 @default.
- W4313010918 hasConceptScore W4313010918C138885662 @default.
- W4313010918 hasConceptScore W4313010918C153180895 @default.
- W4313010918 hasConceptScore W4313010918C154945302 @default.
- W4313010918 hasConceptScore W4313010918C15744967 @default.
- W4313010918 hasConceptScore W4313010918C188147891 @default.
- W4313010918 hasConceptScore W4313010918C2776135515 @default.
- W4313010918 hasConceptScore W4313010918C2776401178 @default.
- W4313010918 hasConceptScore W4313010918C2779458634 @default.
- W4313010918 hasConceptScore W4313010918C2779803651 @default.
- W4313010918 hasConceptScore W4313010918C37736160 @default.
- W4313010918 hasConceptScore W4313010918C39890363 @default.
- W4313010918 hasConceptScore W4313010918C41008148 @default.
- W4313010918 hasConceptScore W4313010918C41895202 @default.
- W4313010918 hasConceptScore W4313010918C76155785 @default.
- W4313010918 hasConceptScore W4313010918C94915269 @default.
- W4313010918 hasFunder F4320320451 @default.
- W4313010918 hasFunder F4320335489 @default.
- W4313010918 hasLocation W43130109181 @default.
- W4313010918 hasLocation W43130109182 @default.
- W4313010918 hasOpenAccess W4313010918 @default.
- W4313010918 hasPrimaryLocation W43130109181 @default.
- W4313010918 hasRelatedWork W2412510955 @default.
- W4313010918 hasRelatedWork W2564929713 @default.
- W4313010918 hasRelatedWork W3005996785 @default.
- W4313010918 hasRelatedWork W3007383607 @default.
- W4313010918 hasRelatedWork W3119931323 @default.
- W4313010918 hasRelatedWork W4280544492 @default.
- W4313010918 hasRelatedWork W4280601492 @default.
- W4313010918 hasRelatedWork W4293320219 @default.
- W4313010918 hasRelatedWork W4221165959 @default.
- W4313010918 hasRelatedWork W4225810998 @default.
- W4313010918 isParatext "false" @default.
- W4313010918 isRetracted "false" @default.
- W4313010918 workType "article" @default.