Filters








4,372 Hits in 3.4 sec

Are Facial Attributes Adversarially Robust? [article]

Andras Rozsa, Manuel Günther, Ethan M. Rudd, Terrance E. Boult
2016 arXiv   pre-print
We show that FFA generates more adversarial examples than other related algorithms, and that DCNNs for certain attributes are generally robust to adversarial inputs, while DCNNs for other attributes are  ...  Facial attributes are emerging soft biometrics that have the potential to reject non-matches, for example, based on mismatching gender.  ...  adversarial images for each of the attribute networks and find that our facial attribute networks attain no additional robustness to adversarial images with longer training. • We introduce the notion  ... 
arXiv:1605.05411v3 fatcat:bvdhrmxuyraq7mq6odjrvbbpme

Facial attributes: Accuracy and adversarial robustness

Andras Rozsa, Manuel Günther, Ethan M. Rudd, Terrance E. Boult
2017 Pattern Recognition Letters  
We show that our fast flipping attribute (FFA) technique generates more adversarial examples than traditional algorithms, and that the adversarial robustness of DNNs varies highly between facial attributes  ...  While recent methods extract facial attributes using deep neural networks (DNNs) trained on labeled facial attribute data, the robustness of deep attribute representations has not been evaluated.  ...  Table 1 : 1 FACIAL ATTRIBUTE FLIPPABILITY BY Table 2 : 2 PROGRESS OF ADVERSARIAL ROBUSTNESS.  ... 
doi:10.1016/j.patrec.2017.10.024 fatcat:yp2arf6qyfcsdl5lstagjrp7ce

FakeTagger: Robust Safeguards against DeepFake Dissemination via Provenance Tracking [article]

Run Wang, Felix Juefei-Xu, Meng Luo, Yang Liu, Lina Wang
2021 arXiv   pre-print
In recent years, DeepFake is becoming a common threat to our society, due to the remarkable progress of generative adversarial networks (GAN) in image synthesis.  ...  The embedded message could be employed to represent the identity of facial images, which further contributed to DeepFake detection and provenance.  ...  In attribute editing, the manipulated facial attribute is changing the color of skin which is the most drastic facial attribute manipulation.  ... 
arXiv:2009.09869v3 fatcat:eemjt2fnxrgr7eg4dursbec5hu

Interpretable and Robust Face Verification

Preetam Prabhu Srikar Dammu, Srinivasa Rao Chalamala, Ajeet Kumar Singh, Bayya Yegnanarayana
2021 International Conference on Information and Knowledge Management  
In this, representations for each individual facial parts such as nose, mouth, eyes etc. are learned separately.  ...  Additionally, most of the existing face recognition models are highly susceptible to adversarial attacks.  ...  The most important attribute of the proposed method is that it is both robust to adversarial attacks and inherently interpretable.  ... 
dblp:conf/cikm/DammuCSY21 fatcat:aqaiqxxwyngbbabyjlwxobiwmu

Robust SleepNets [article]

Yigit Alparslan, Edward Kim
2021 arXiv   pre-print
It is crucial that machine critical systems, where machine learning models are deployed, utilize robust models to handle a wide range of variability in the real world and malicious actors that may use  ...  State-of-the-art convolutional neural networks excel in machine learning tasks such as face recognition, and object classification but suffer significantly when adversarial attacks are present.  ...  We attack the models with adversarial attacks to study the robustness of the networks under adversarial conditions.  ... 
arXiv:2102.12555v1 fatcat:l3ai7omexnagzlotyfmnmxz5ze

Fairness Through Robustness: Investigating Robustness Disparity in Deep Learning [article]

Vedant Nanda and Samuel Dooley and Sahil Singla and Soheil Feizi and John P. Dickerson
2021 arXiv   pre-print
subgroups (in some cases based on sensitive attributes like race, gender, etc) which are less robust and are thus at a disadvantage.  ...  In this paper, we argue that traditional notions of fairness that are only based on models' outputs are not sufficient when the model is vulnerable to adversarial attacks.  ...  We observe that the Adience dataset, which exhibited some adversarial robustness bias in the partition on C only exhibits minor adversarial robustness bias in the partition on S for the attribute 'Female  ... 
arXiv:2006.12621v4 fatcat:ryitjqczrnf6fpe4d4nzeivedu

Robust Deepfake On Unrestricted Media: Generation And Detection [article]

Trung-Nghia Le and Huy H Nguyen and Junichi Yamagishi and Isao Echizen
2022 arXiv   pre-print
It also discusses possible ways to improve the robustness of deepfake detection for a wide variety of media (e.g., in-the-wild images and videos).  ...  Although deepfake media have potential application in a wide range of areas and are drawing much attention from both the academic and industrial communities, it also leads to serious social and criminal  ...  Thus, GAN techniques for face synthesis are usually used for facial attribute manipulation. Choi et al.  ... 
arXiv:2202.06228v1 fatcat:a37q2lf7w5bcbekk5esmbx2goe

Towards Evaluating Driver Fatigue with Robust Deep Learning Models [article]

Ken Alparslan, Yigit Alparslan, Matthew Burlick
2021 arXiv   pre-print
We also see that we can improve our accuracy on the face model by 6% via adversarial training and data augmentation.  ...  We also explore different techniques to make the models more robust by adding noise. We achieve 95.84% accuracy on our eye model and 80.01% accuracy on our face model.  ...  We also studied facial fatigue detection under an adversarial detection by applying data augmentation in order to increase robustness.  ... 
arXiv:2007.08453v4 fatcat:wf2f6dihj5fapjyunae3j2vyae

Robust Conditional GAN from Uncertainty-Aware Pairwise Comparisons

Ligong Han, Ruijiang Gao, Mun Kim, Xin Tao, Bo Liu, Dimitris Metaxas
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
In the light of Bayesian uncertainty estimation and noise-tolerant adversarial training, PC-GAN can estimate attribute rating efficiently and demonstrate robust performance in noise resistance.  ...  To address this problem, we propose a novel generative adversarial network utilizing weak supervision in the form of pairwise comparisons (PC-GAN) for image attribute editing.  ...  However, most facial attributes like attractiveness and age are not localized features thus cannot be exploited by local regions.  ... 
doi:10.1609/aaai.v34i07.6723 fatcat:bxl7iyx2xjg7xfbyclhym7ollu

Robust Conditional GAN from Uncertainty-Aware Pairwise Comparisons [article]

Ligong Han, Ruijiang Gao, Mun Kim, Xin Tao, Bo Liu, Dimitris Metaxas
2020 arXiv   pre-print
In the light of Bayesian uncertainty estimation and noise-tolerant adversarial training, PC-GAN can estimate attribute rating efficiently and demonstrate robust performance in noise resistance.  ...  To address this problem, we propose a novel generative adversarial network utilizing weak supervision in the form of pairwise comparisons (PC-GAN) for image attribute editing.  ...  Facial attribute classifiers are finetuned ResNet-18 (He et al. 2016) .  ... 
arXiv:1911.09298v2 fatcat:o4veqikuvfa45ajefdrknxb65m

Anti-Forgery: Towards a Stealthy and Robust DeepFake Disruption Attack via Adversarial Perceptual-aware Perturbations [article]

Run Wang, Ziheng Huang, Zhikai Chen, Li Liu, Jing Chen, Lina Wang
2022 arXiv   pre-print
However, the existing studies on proactive DeepFake defense via injecting adversarial noises are not robust, which could be easily bypassed by employing simple image reconstruction revealed in a recent  ...  and robust manner.  ...  CelebA contains more than 200K facial images with 40 attributes annotation for each face. All the facial images are cropped to 256×256. Model Architectures.  ... 
arXiv:2206.00477v1 fatcat:w6lrbyg6ajahvahvdz2ppffr5u

WGAN-based Robust Occluded Facial Expression Recognition

Yang Lu, Shigang Wang, Wenting Zhao, Yan Zhao
2019 IEEE Access  
Currently, most of the related works on this technology are focused on un-occluded FER.  ...  INDEX TERMS Facial expression recognition, partial occlusion, image complementation, Wasserstein generative adversarial network.  ...  FIGURE 1 . 1 Framework of the robust occluded facial expression recognition based on the proposed method. FIGURE 2 . 2 The architecture of the proposed generator.  ... 
doi:10.1109/access.2019.2928125 fatcat:wo3xkvvh2bbybpfvoe6bzc2r74

RoCGAN: Robust Conditional GAN

Grigorios G. Chrysos, Jean Kossaifi, Stefanos Zafeiriou
2020 International Journal of Computer Vision  
The focus so far has largely been on performance improvement, with little effort in making cGANs more robust to noise.  ...  We also empirically demonstrate the performance of our approach in the face of two types of noise (adversarial and Bernoulli).  ...  Both models are more robust when trained with Gaussian noise; it requires 15 adversarial steps instead of 10 to achieve the same degradation.  ... 
doi:10.1007/s11263-020-01348-5 fatcat:dep7zvp4ene23dlg42qw42vdza

Protecting Facial Privacy: Generating Adversarial Identity Masks via Style-robust Makeup Transfer [article]

Shengshan Hu, Xiaogeng Liu, Yechao Zhang, Minghui Li, Leo Yu Zhang, Hai Jin, Libing Wu
2022 arXiv   pre-print
AMT-GAN leverages generative adversarial networks (GAN) to synthesize adversarial face images with makeup transferred from reference images.  ...  In this paper, we propose adversarial makeup transfer GAN (AMT-GAN), a novel face protection method aiming at constructing adversarial face images that preserve stronger black-box transferability and better  ...  Note that PGD, MI-FGSM, and TI-DIM are very famous for their strong attack ability, TIP-IM is a very recent work which leverages adversarial examples to protect facial privacy, and Adv-makeup is the most  ... 
arXiv:2203.03121v2 fatcat:kogfrvyslvhqli3tdfylfkvqv4

RoPAD: Robust Presentation Attack Detection through Unsupervised Adversarial Invariance [article]

Ayush Jaiswal, Shuai Xia, Iacopo Masi, Wael AbdAlmageed
2019 arXiv   pre-print
We present RoPAD, an end-to-end deep learning model for presentation attack detection that employs unsupervised adversarial invariance to ignore visual distractors in images for increased robustness and  ...  However, current authentication technologies are still vulnerable to presentation attacks.  ...  The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the  ... 
arXiv:1903.03691v2 fatcat:crmwyw7isvfghf6a5q66kl4h6u
« Previous Showing results 1 — 15 out of 4,372 results