UAA-GAN generates query-specific adversarial perturbations via unsupervised GAN training that reduce retrieval accuracy in deep feature spaces while keeping changes visually small.
Least squares generative adversarial networks,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2019 2verdicts
UNVERDICTED 2representative citing papers
The paper shows that multiple-identity image attacks succeed due to modest angular separation between matching (~90°) and non-matching (40-60°) face representations, with image morphing and representation inversion realizing effective attacks that transfer across comparators.
citing papers explorer
-
Unsupervised Adversarial Attacks on Deep Feature-based Retrieval with GAN
UAA-GAN generates query-specific adversarial perturbations via unsupervised GAN training that reduce retrieval accuracy in deep feature spaces while keeping changes visually small.
-
Multiple-Identity Image Attacks Against Face-based Identity Verification
The paper shows that multiple-identity image attacks succeed due to modest angular separation between matching (~90°) and non-matching (40-60°) face representations, with image morphing and representation inversion realizing effective attacks that transfer across comparators.