Imperceptible Adversarial Examples by Spatial Chroma-Shift

Date

2021-10-20

Author

Aydın, Ayberk
Sen, Deniz
Karli, Berat Tuna
Hanoglu, Oguz
Temizel, Alptekin

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

162
views

4
downloads

Deep Neural Networks have been shown to be vulnerable to various kinds of adversarial perturbations. In addition to widely studied additive noise based perturbations, adversarial examples can also be created by applying a per pixel spatial drift on input images. While spatial transformation based adversarial examples look more natural to human observers due to absence of additive noise, they still possess visible distortions caused by spatial transformations. Since the human vision is more sensitive to the distortions in the luminance compared to those in chrominance channels, which is one of the main ideas behind the lossy visual multimedia compression standards, we propose a spatial transformation based perturbation method to create adversarial examples by only modifying the color components of an input image. While having competitive fooling rates on CIFAR-10 and NIPS2017 Adversarial Learning Challenge datasets, examples created with the proposed method have better scores with regards to various perceptual quality metrics. Human visual perception studies validate that the examples are more natural looking and often indistinguishable from their original counterparts.

Subject Keywords

adversarial examples, computer vision, neural networks

URI

https://hdl.handle.net/11511/99685

DOI

https://doi.org/10.1145/3475724.3483604

Conference Name

1st International Workshop on Adversarial Learning for Multimedia, AdvM 2021, co-located with ACM MM 2021

Collections

Graduate School of Informatics, Conference / Seminar

Suggestions

OpenMETU
Core

Improving Perceptual Quality of Spatially Transformed Adversarial Examples Aydın, Ayberk; Temizel, Alptekin; Department of Modeling and Simulation (2022-8) Deep neural networks are known to be vulnerable to additive adversarial perturbations. The amount of these additive perturbations are generally quantified using Lp metrics over the difference between adversarial and benign examples. However, even when the measured perturbations are small, they tend to be noticeable by human observers since Lp distance metrics are not representative of human perception. Spatially transformed examples work by distorting pixel locations instead of applying an additive perturba...
Neural networks with piecewise constant argument and impact activation Yılmaz, Enes; Akhmet, Marat; Department of Scientific Computing (2011) This dissertation addresses the new models in mathematical neuroscience: artificial neural networks, which have many similarities with the structure of human brain and the functions of cells by electronic circuits. The networks have been investigated due to their extensive applications in classification of patterns, associative memories, image processing, artificial intelligence, signal processing and optimization problems. These applications depend crucially on the dynamical behaviors of the networks. In t...
IMPULSIVE SICNNS WITH CHAOTIC POSTSYNAPTIC CURRENTS Fen, Mehmet Onur; Akhmet, Marat (2016-06-01) In the present study, we investigate the dynamics of shunting inhibitory cellular neural networks (SICNNs) with impulsive effects. We give a mathematical description of the chaos for the multidimensional dynamics of impulsive SICNNs, and prove its existence rigorously by taking advantage of the external inputs. The Li-Yorke definition of chaos is used in our theoretical discussions. In the considered model, the impacts satisfy the cell and shunting principles. This enriches the applications of SICNNs and ma...
Attack Independent Perceptual Improvement of Adversarial Examples Karlı, Berat Tuna; Temizel, Alptekin; Department of Information Systems (2022-12-23) Deep Neural networks (DNNs) are used in a variety of domains with great success, however, it has been proven that these networks are vulnerable to additive non-arbitrary perturbations. Regarding this fact, several attack and defense mechanisms have been developed; nevertheless, adding crafted perturbations has a negative effect on the perceptual quality of images. This study aims to improve the perceptual quality of adversarial examples independent of attack type and the integration of two attack agnostic t...
Attraction of Li-Yorke chaos by retarded SICNNs Akhmet, Marat (Elsevier BV, 2015-01-05) In the present study, dynamics of retarded shunting inhibitory cellular neural networks (SICNNs) is investigated with Li-Yorke chaotic external inputs and outputs. Within the scope of our results, we prove the presence of generalized synchronization in coupled retarded SICNNs, and confirm it by means of the auxiliary system approach. We have obtained more than just synchronization, as it is proved that the Li-yorke chaos is extended with its ingredients, proximality and frequent separation, which have not b...

Citation Formats

A. Aydın, D. Sen, B. T. Karli, O. Hanoglu, and A. Temizel, “Imperceptible Adversarial Examples by Spatial Chroma-Shift,” presented at the 1st International Workshop on Adversarial Learning for Multimedia, AdvM 2021, co-located with ACM MM 2021, Virtual, Online, Çin, 2021, Accessed: 00, 2022. [Online]. Available: https://hdl.handle.net/11511/99685.