TY - GEN
T1 - MeshAdv
T2 - 32nd IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2019
AU - Xiao, Chaowei
AU - Yang, Dawei
AU - Li, Bo
AU - Deng, Jia
AU - Liu, Mingyan
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/6
Y1 - 2019/6
N2 - Highly expressive models such as deep neural networks (DNNs) have been widely applied to various applications. However, recent studies show that DNNs are vulnerable to adversarial examples, which are carefully crafted inputs aiming to mislead the predictions. Currently, the majority of these studies have focused on perturbation added to image pixels, while such manipulation is not physically realistic. Some works have tried to overcome this limitation by attaching printable 2D patches or painting patterns onto surfaces, but can be potentially defended because 3D shape features are intact. In this paper, we propose meshAdv to generate 'adversarial 3D meshes' from objects that have rich shape features but minimal textural variation. To manipulate the shape or texture of the objects, we make use of a differentiable renderer to compute accurate shading on the shape and propagate the gradient. Extensive experiments show that the generated 3D meshes are effective in attacking both classifiers and object detectors. We evaluate the attack under different viewpoints. In addition, we design a pipeline to perform black-box attack on a photorealistic renderer with unknown rendering parameters.
AB - Highly expressive models such as deep neural networks (DNNs) have been widely applied to various applications. However, recent studies show that DNNs are vulnerable to adversarial examples, which are carefully crafted inputs aiming to mislead the predictions. Currently, the majority of these studies have focused on perturbation added to image pixels, while such manipulation is not physically realistic. Some works have tried to overcome this limitation by attaching printable 2D patches or painting patterns onto surfaces, but can be potentially defended because 3D shape features are intact. In this paper, we propose meshAdv to generate 'adversarial 3D meshes' from objects that have rich shape features but minimal textural variation. To manipulate the shape or texture of the objects, we make use of a differentiable renderer to compute accurate shading on the shape and propagate the gradient. Extensive experiments show that the generated 3D meshes are effective in attacking both classifiers and object detectors. We evaluate the attack under different viewpoints. In addition, we design a pipeline to perform black-box attack on a photorealistic renderer with unknown rendering parameters.
KW - Deep Learning
KW - Vision + Graphics
UR - http://www.scopus.com/inward/record.url?scp=85078760755&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85078760755&partnerID=8YFLogxK
U2 - 10.1109/CVPR.2019.00706
DO - 10.1109/CVPR.2019.00706
M3 - Conference contribution
AN - SCOPUS:85078760755
T3 - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
SP - 6891
EP - 6900
BT - Proceedings - 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2019
PB - IEEE Computer Society
Y2 - 16 June 2019 through 20 June 2019
ER -