Matches in SemOpenAlex for { <https://semopenalex.org/work/W3193285461> ?p ?o ?g. }
Showing items 1 to 92 of
92
with 100 items per page.
- W3193285461 abstract "Images synthesized by powerful generative adversarial network (GAN) based methods have drawn moral and privacy concerns. Although image forensic models have reached great performance in detecting fake images from real ones, these models can be easily fooled with a simple adversarial attack. But, the noise adding adversarial samples are also arousing suspicion. In this paper, instead of adding adversarial noise, we optimally search adversarial points on face manifold to generate anti-forensic fake face images. We iteratively do a gradient-descent with each small step in the latent space of a generative model, e.g. Style-GAN, to find an adversarial latent vector, which is similar to norm-based adversarial attack but in latent space. Then, the generated fake images driven by the adversarial latent vectors with the help of GANs can defeat main-stream forensic models. For examples, they make the accuracy of deepfake detection models based on Xception or EfficientNet drop from over 90% to nearly 0%, mean-while maintaining high visual quality. In addition, we find manipulating noise vectors n at different levels have different impacts on attack success rate, and the generated adversarial images mainly have changes on facial texture or face attributes." @default.
- W3193285461 created "2021-08-30" @default.
- W3193285461 creator A5017743261 @default.
- W3193285461 creator A5043976050 @default.
- W3193285461 creator A5077686672 @default.
- W3193285461 creator A5091396758 @default.
- W3193285461 date "2021-06-01" @default.
- W3193285461 modified "2023-09-29" @default.
- W3193285461 title "Exploring Adversarial Fake Images on Face Manifold" @default.
- W3193285461 cites W1834627138 @default.
- W3193285461 cites W2080277992 @default.
- W3193285461 cites W2108598243 @default.
- W3193285461 cites W2148685281 @default.
- W3193285461 cites W2486034530 @default.
- W3193285461 cites W2558151185 @default.
- W3193285461 cites W2774018344 @default.
- W3193285461 cites W2774037436 @default.
- W3193285461 cites W2774644650 @default.
- W3193285461 cites W2962770929 @default.
- W3193285461 cites W2962958939 @default.
- W3193285461 cites W2963626105 @default.
- W3193285461 cites W2963693747 @default.
- W3193285461 cites W2964082701 @default.
- W3193285461 cites W2982058372 @default.
- W3193285461 cites W2984700035 @default.
- W3193285461 cites W3034176567 @default.
- W3193285461 cites W3034196597 @default.
- W3193285461 doi "https://doi.org/10.1109/cvpr46437.2021.00573" @default.
- W3193285461 hasPublicationYear "2021" @default.
- W3193285461 type Work @default.
- W3193285461 sameAs 3193285461 @default.
- W3193285461 citedByCount "15" @default.
- W3193285461 countsByYear W31932854612021 @default.
- W3193285461 countsByYear W31932854612022 @default.
- W3193285461 countsByYear W31932854612023 @default.
- W3193285461 crossrefType "proceedings-article" @default.
- W3193285461 hasAuthorship W3193285461A5017743261 @default.
- W3193285461 hasAuthorship W3193285461A5043976050 @default.
- W3193285461 hasAuthorship W3193285461A5077686672 @default.
- W3193285461 hasAuthorship W3193285461A5091396758 @default.
- W3193285461 hasBestOaLocation W31932854612 @default.
- W3193285461 hasConcept C115961682 @default.
- W3193285461 hasConcept C119857082 @default.
- W3193285461 hasConcept C144024400 @default.
- W3193285461 hasConcept C153180895 @default.
- W3193285461 hasConcept C154945302 @default.
- W3193285461 hasConcept C17744445 @default.
- W3193285461 hasConcept C191795146 @default.
- W3193285461 hasConcept C199539241 @default.
- W3193285461 hasConcept C2779304628 @default.
- W3193285461 hasConcept C2988773926 @default.
- W3193285461 hasConcept C31972630 @default.
- W3193285461 hasConcept C36289849 @default.
- W3193285461 hasConcept C37736160 @default.
- W3193285461 hasConcept C39890363 @default.
- W3193285461 hasConcept C41008148 @default.
- W3193285461 hasConcept C99498987 @default.
- W3193285461 hasConceptScore W3193285461C115961682 @default.
- W3193285461 hasConceptScore W3193285461C119857082 @default.
- W3193285461 hasConceptScore W3193285461C144024400 @default.
- W3193285461 hasConceptScore W3193285461C153180895 @default.
- W3193285461 hasConceptScore W3193285461C154945302 @default.
- W3193285461 hasConceptScore W3193285461C17744445 @default.
- W3193285461 hasConceptScore W3193285461C191795146 @default.
- W3193285461 hasConceptScore W3193285461C199539241 @default.
- W3193285461 hasConceptScore W3193285461C2779304628 @default.
- W3193285461 hasConceptScore W3193285461C2988773926 @default.
- W3193285461 hasConceptScore W3193285461C31972630 @default.
- W3193285461 hasConceptScore W3193285461C36289849 @default.
- W3193285461 hasConceptScore W3193285461C37736160 @default.
- W3193285461 hasConceptScore W3193285461C39890363 @default.
- W3193285461 hasConceptScore W3193285461C41008148 @default.
- W3193285461 hasConceptScore W3193285461C99498987 @default.
- W3193285461 hasFunder F4320321001 @default.
- W3193285461 hasLocation W31932854611 @default.
- W3193285461 hasLocation W31932854612 @default.
- W3193285461 hasOpenAccess W3193285461 @default.
- W3193285461 hasPrimaryLocation W31932854611 @default.
- W3193285461 hasRelatedWork W2901368259 @default.
- W3193285461 hasRelatedWork W3037184111 @default.
- W3193285461 hasRelatedWork W3156291593 @default.
- W3193285461 hasRelatedWork W3156763702 @default.
- W3193285461 hasRelatedWork W3178813832 @default.
- W3193285461 hasRelatedWork W3198184493 @default.
- W3193285461 hasRelatedWork W4206598283 @default.
- W3193285461 hasRelatedWork W4220812973 @default.
- W3193285461 hasRelatedWork W4312525628 @default.
- W3193285461 hasRelatedWork W4386519296 @default.
- W3193285461 isParatext "false" @default.
- W3193285461 isRetracted "false" @default.
- W3193285461 magId "3193285461" @default.
- W3193285461 workType "article" @default.