MimicGAN: Corruption-Mimicking for Blind Image Recovery &amp; Adversarial Defense

Anirudh, Rushil; Thiagarajan, Jayaraman J.; Kailkhura, Bhavya; Bremer, Timo

doi:10.48550/arxiv.1811.08484

Cited by 2 publications

(5 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We run all our reconstruction experiments on a subset of the 10K validation set. Corruption-mimicking requires choosing 4 main hyperparameters [7]:…”

Section: Methodsmentioning

confidence: 99%

“…In order to avoid this, we propose to use a recently proposed modification of the GAN prior, that performs better even with heavily distorted images. The process called corruption mimicking was proposed in [7], was designed to improve the quality of projection onto the manifold under a variety of corruptions.…”

Section: Proposed Approachmentioning

confidence: 99%

“…In our setting, f contains 2 convolutional layers with ReLU activations, followed by a masking layer (pixel-wise multiplication). Finally, we also include a shortcut connection at the end to encourage it to learn identity [7]. The GAN prior now becomes a special case of the Robust GAN prior, when f = I, the identity function.…”

Section: Proposed Approachmentioning

confidence: 99%

“…Though this has been conventionally carried out using projected gradient descent (PGD) [5,6], as we demonstrate in our results, this approach performs poorly when the initial estimate is too noisy or has too many artifacts, which is common under extremely limited angle scenarios. Instead, our approach utilizes a recently proposed technique referred to as corruption mimicking, used in the design of MimicGAN [7], that achieves robustness to the noisy seed reconstruction through the use of a randomly initialized shallow convolutional neural network (CNN), in addition to PGD. By modeling the initial guess of this network as a random corruption for the unknown clean image, the process of corruption mimicking alternates between estimating the unknown corruption and finding the clean solution, and this alternating optimization is repeated until convergence, in terms of effectively matching the observed noisy data.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Improving Limited Angle CT Reconstruction with a Robust GAN Prior

Kim

Thiagarajan

Mohan

et al. 2019

Self Cite

View full text Add to dashboard Cite

Limited angle CT reconstruction is an under-determined linear inverse problem that requires appropriate regularization techniques to be solved. In this work we study how pre-trained generative adversarial networks (GANs) can be used to clean noisy, highly artifact laden reconstructions from conventional techniques, by effectively projecting onto the inferred image manifold. In particular, we use a robust version of the popularly used GAN prior for inverse problems, based on a recent technique called corruption mimicking, that significantly improves the reconstruction quality. The proposed approach operates in the image space directly, as a result of which it does not need to be trained or require access to the measurement model, is scanner agnostic, and can work over a wide range of sensing scenarios.

show abstract

“…We run all our reconstruction experiments on a subset of the 10K validation set. Corruption-mimicking requires choosing 4 main hyperparameters [7]:…”

Section: Methodsmentioning

confidence: 99%

Section: Proposed Approachmentioning

confidence: 99%

Section: Proposed Approachmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Improving Limited Angle CT Reconstruction with a Robust GAN Prior

Kim

Thiagarajan

Mohan

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, this method require the used GAN to gain some prior knowledge about the attackers. To make the defence independent of the prior knowledge, study [42] proposes MimicGAN, which formulates the adversarial defence as a general inverse problem and address it with a unsupervised technique. However, since this kind of defence requires to train many generative models for different datasets, too many computational resources will be consumed.…”

Section: Related Workmentioning

confidence: 99%

DeSVig: Decentralized Swift Vigilance Against Adversarial Attacks in Industrial Artificial Intelligence Systems

Ota

Dong

et al. 2020

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

Individually reinforcing the robustness of a single deep learning model only gives limited security guarantees especially when facing adversarial examples. In this paper, we propose DeSVig, a Decentralized Swift Vigilance framework to identify adversarial attacks in an industrial artificial intelligence systems (IAISs), which enables IAISs to correct the mistake in a few seconds. DeSVig is highly decentralized, which improves the effectiveness of recognizing abnormal inputs. We try to overcome the challenges on ultra-low latency caused by dynamics in industries using peculiarly-designated mobile edge computing and generative adversarial networks (GANs). The most important advantage of our work is that it can significantly reduce the failure risks of being deceived by adversarial examples, which is critical for safety-prioritized and delay-sensitive environments.In our experiments, adversarial examples of industrial electronic components are generated by several classical attacking models. Experimental results demonstrate the DeSVig is more robust, efficient, and scalable than some state-of-art defences.

show abstract

MimicGAN: Corruption-Mimicking for Blind Image Recovery & Adversarial Defense

Cited by 2 publications

References 20 publications

Improving Limited Angle CT Reconstruction with a Robust GAN Prior

Improving Limited Angle CT Reconstruction with a Robust GAN Prior

DeSVig: Decentralized Swift Vigilance Against Adversarial Attacks in Industrial Artificial Intelligence Systems

Contact Info

Product

Resources

About