Built-in Elastic Transformations for Improved Robustness

We focus on building robustness in the convolutions of neural visual classifiers, especially against natural perturbations like elastic deformations, occlusions and Gaussian noise. Existing CNNs show outstanding performance on clean images, but fail to tackle naturally occurring perturbations. In this paper, we start from elastic perturbations, which approximate (local) view-point changes of the object. We present elastically-augmented convolutions (EAConv) by parameterizing filters as a combination of fixed elasticallyperturbed bases functions and trainable weights for the purpose of integrating unseen viewpoints in the CNN. We show on CIFAR-10 and STL-10 datasets that the general robustness of our method on unseen occlusion and Gaussian perturbations improves, while even improving the performance on clean images slightly without performing any data augmentation.

[1]  J. Zico Kolter,et al.  Provable defenses against adversarial examples via the convex outer adversarial polytope , 2017, ICML.

[2]  Honglak Lee,et al.  An Analysis of Single-Layer Networks in Unsupervised Feature Learning , 2011, AISTATS.

[3]  J. Zico Kolter,et al.  Learning perturbation sets for robust machine learning , 2020, ICLR.

[4]  Benjamin Recht,et al.  Do CIFAR-10 Classifiers Generalize to CIFAR-10? , 2018, ArXiv.

[5]  Ivan Sosnovik,et al.  Scale Equivariance Improves Siamese Tracking , 2020, 2021 IEEE Winter Conference on Applications of Computer Vision (WACV).

[6]  Matthias Bethge,et al.  A Simple Way to Make Neural Networks Robust Against Diverse Image Corruptions , 2020, ECCV.

[7]  Nikos Komodakis,et al.  Wide Residual Networks , 2016, BMVC.

[8]  Matthias Bethge,et al.  Improving robustness against common corruptions by covariate shift adaptation , 2020, NeurIPS.

[9]  Yi Li,et al.  Deformable Convolutional Networks , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[10]  Ivan Sosnovik,et al.  Scale-Equivariant Steerable Networks , 2020, ICLR.

[11]  Michael I. Jordan,et al.  Theoretically Principled Trade-off between Robustness and Accuracy , 2019, ICML.

[12]  Andrew Zisserman,et al.  Spatial Transformer Networks , 2015, NIPS.

[13]  Zhanxing Zhu,et al.  Interpreting Adversarially Trained Convolutional Neural Networks , 2019, ICML.

[14]  Dimitris N. Metaxas,et al.  SelfNorm and CrossNorm for Out-of-Distribution Robustness , 2021, ArXiv.

[15]  Arnold W. M. Smeulders,et al.  Structured Receptive Fields in CNNs , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[16]  Alex Krizhevsky,et al.  Learning Multiple Layers of Features from Tiny Images , 2009 .

[17]  Arnold Smeulders,et al.  Natural Perturbed Training for General Robustness of Neural Network Classifiers , 2021, ArXiv.

[18]  Sung Ju Hwang,et al.  Adversarial Neural Pruning , 2019, ArXiv.

[19]  Aleksander Madry,et al.  Towards Deep Learning Models Resistant to Adversarial Attacks , 2017, ICLR.

[20]  Aleksander Madry,et al.  Exploring the Landscape of Spatial Robustness , 2017, ICML.

[21]  Aleksander Madry,et al.  Robustness May Be at Odds with Accuracy , 2018, ICLR.

[22]  Yair Weiss,et al.  Why do deep convolutional networks generalize so poorly to small image transformations? , 2018, J. Mach. Learn. Res..

[23]  In So Kweon,et al.  Revisiting Batch Normalization for Improving Corruption Robustness , 2021, 2021 IEEE Winter Conference on Applications of Computer Vision (WACV).

[24]  Alexandr A. Kalinin,et al.  Albumentations: fast and flexible image augmentations , 2018, Inf..

[25]  David A. McAllester,et al.  Object Detection with Discriminatively Trained Part Based Models , 2010, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[26]  Jonathon Shlens,et al.  Explaining and Harnessing Adversarial Examples , 2014, ICLR.

[27]  David Dagan Feng,et al.  Elastic Medical Image Registration Based on Image Intensity , 2001, VIP.

[28]  George J. Pappas,et al.  Model-Based Robust Deep Learning , 2020, ArXiv.

[29]  Geoffrey E. Hinton.  A Parallel Computation that Assigns Canonical Object-Based Frames of Reference , 1981, IJCAI.

[30]  Lina J. Karam,et al.  A Study and Comparison of Human and Deep Learning Recognition Performance under Visual Distortions , 2017, 2017 26th International Conference on Computer Communication and Networks (ICCCN).