Matches in SemOpenAlex for { <https://semopenalex.org/work/W3121736286> ?p ?o ?g. }
- W3121736286 abstract "Despite their recent successes, GAN models for semantic image synthesis still suffer from poor image quality when trained with only adversarial supervision. Historically, additionally employing the VGG-based perceptual loss has helped to overcome this issue, significantly improving the synthesis quality, but at the same time limiting the progress of GAN models for semantic image synthesis. In this work, we propose a novel, simplified GAN model, which needs only adversarial supervision to achieve high quality results. We re-design the discriminator as a semantic segmentation network, directly using the given semantic label maps as the ground truth for training. By providing stronger supervision to the discriminator as well as to the generator through spatially- and semantically-aware discriminator feedback, we are able to synthesize images of higher fidelity with better alignment to their input label maps, making the use of the perceptual loss superfluous. Moreover, we enable high-quality multi-modal image synthesis through global and local sampling of a 3D noise tensor injected into the generator, which allows complete or partial image change. We show that images synthesized by our model are more diverse and follow the color and texture distributions of real images more closely. We achieve an average improvement of 6 FID and 5 mIoU points over the state of the art across different datasets using only adversarial supervision." @default.
- W3121736286 created "2021-02-01" @default.
- W3121736286 creator A5008608555 @default.
- W3121736286 creator A5012240246 @default.
- W3121736286 creator A5017543462 @default.
- W3121736286 creator A5026832311 @default.
- W3121736286 creator A5051534545 @default.
- W3121736286 creator A5086559201 @default.
- W3121736286 date "2021-05-03" @default.
- W3121736286 modified "2023-09-23" @default.
- W3121736286 title "You Only Need Adversarial Supervision for Semantic Image Synthesis" @default.
- W3121736286 cites W1580389772 @default.
- W3121736286 cites W1901129140 @default.
- W3121736286 cites W2039051707 @default.
- W3121736286 cites W2099471712 @default.
- W3121736286 cites W2108598243 @default.
- W3121736286 cites W2125389028 @default.
- W3121736286 cites W2143668817 @default.
- W3121736286 cites W2331128040 @default.
- W3121736286 cites W2340897893 @default.
- W3121736286 cites W2405756170 @default.
- W3121736286 cites W2475287302 @default.
- W3121736286 cites W2561196672 @default.
- W3121736286 cites W2737258237 @default.
- W3121736286 cites W2778764040 @default.
- W3121736286 cites W2804078698 @default.
- W3121736286 cites W2805984778 @default.
- W3121736286 cites W2884805522 @default.
- W3121736286 cites W2884822772 @default.
- W3121736286 cites W2891331468 @default.
- W3121736286 cites W2893749619 @default.
- W3121736286 cites W2908977068 @default.
- W3121736286 cites W2962754210 @default.
- W3121736286 cites W2962770929 @default.
- W3121736286 cites W2962785568 @default.
- W3121736286 cites W2962835968 @default.
- W3121736286 cites W2962850830 @default.
- W3121736286 cites W2962879692 @default.
- W3121736286 cites W2962974533 @default.
- W3121736286 cites W2963073614 @default.
- W3121736286 cites W2963163163 @default.
- W3121736286 cites W2963522749 @default.
- W3121736286 cites W2963539305 @default.
- W3121736286 cites W2963783084 @default.
- W3121736286 cites W2963800363 @default.
- W3121736286 cites W2963836885 @default.
- W3121736286 cites W2963864522 @default.
- W3121736286 cites W2963881378 @default.
- W3121736286 cites W2963981733 @default.
- W3121736286 cites W2964121744 @default.
- W3121736286 cites W2964193438 @default.
- W3121736286 cites W2964288706 @default.
- W3121736286 cites W2970955049 @default.
- W3121736286 cites W2971128425 @default.
- W3121736286 cites W2971337561 @default.
- W3121736286 cites W2984854289 @default.
- W3121736286 cites W2992308087 @default.
- W3121736286 cites W3014134643 @default.
- W3121736286 cites W3014705500 @default.
- W3121736286 cites W3015643106 @default.
- W3121736286 cites W3016719260 @default.
- W3121736286 cites W3034214199 @default.
- W3121736286 cites W3034954643 @default.
- W3121736286 cites W3035687950 @default.
- W3121736286 cites W3035757124 @default.
- W3121736286 cites W3093407252 @default.
- W3121736286 hasPublicationYear "2021" @default.
- W3121736286 type Work @default.
- W3121736286 sameAs 3121736286 @default.
- W3121736286 citedByCount "2" @default.
- W3121736286 countsByYear W31217362862021 @default.
- W3121736286 countsByYear W31217362862022 @default.
- W3121736286 crossrefType "proceedings-article" @default.
- W3121736286 hasAuthorship W3121736286A5008608555 @default.
- W3121736286 hasAuthorship W3121736286A5012240246 @default.
- W3121736286 hasAuthorship W3121736286A5017543462 @default.
- W3121736286 hasAuthorship W3121736286A5026832311 @default.
- W3121736286 hasAuthorship W3121736286A5051534545 @default.
- W3121736286 hasAuthorship W3121736286A5086559201 @default.
- W3121736286 hasConcept C111472728 @default.
- W3121736286 hasConcept C115961682 @default.
- W3121736286 hasConcept C121332964 @default.
- W3121736286 hasConcept C124504099 @default.
- W3121736286 hasConcept C138885662 @default.
- W3121736286 hasConcept C146849305 @default.
- W3121736286 hasConcept C154945302 @default.
- W3121736286 hasConcept C163258240 @default.
- W3121736286 hasConcept C2779530757 @default.
- W3121736286 hasConcept C2779803651 @default.
- W3121736286 hasConcept C2780992000 @default.
- W3121736286 hasConcept C31972630 @default.
- W3121736286 hasConcept C37736160 @default.
- W3121736286 hasConcept C41008148 @default.
- W3121736286 hasConcept C50494287 @default.
- W3121736286 hasConcept C55020928 @default.
- W3121736286 hasConcept C62520636 @default.
- W3121736286 hasConcept C63099799 @default.
- W3121736286 hasConcept C76155785 @default.
- W3121736286 hasConcept C89600930 @default.
- W3121736286 hasConcept C94915269 @default.