Attacking Optical Flow (ICCV 2019)

23 October 2019

01:51

Deep neural nets achieve state-of-the-art performance on the problem of optical flow estimation. Since optical flow is used in several safety-critical applications like self-driving cars, it is important to gain insights into the robustness of those techniques. Recently, it has been shown that adversarial attacks easily fool deep neural networks to misclassify objects. The robustness of optical flow networks to adversarial attacks, however, has not been studied so far. In this paper, we extend adversarial patch attacks to optical flow networks and show that such attacks can compromise their performance. We show that corrupting a small patch of less than 1% of the image size can significantly affect optical flow estimates. Our attacks lead to noisy flow estimates that extend significantly beyond the region of the attack, in many cases even completely erasing the motion of objects in the scene. While networks using an encoder-decoder architecture are very sensitive to these attacks, we found that networks using a spatial pyramid architecture are less affected. We analyse the success and failure of attacking both architectures by visualizing their feature maps and comparing them to classical optical flow techniques which are robust to these attacks. We also demonstrate that such attacks are practical by placing a printed pattern into real scenes. Note: Due to the potential public safety issues, we informed major makers of autonomous-car technology in advance of our public disclosure. We provided them with the paper, the patches and code if they wanted it. Project: https://flowattack.is.tue.mpg.de/ Paper: https://arxiv.org/pdf/1910.10053.pdf Code: https://github.com/anuragranj/flowattack/ Citation: @inproceedings{ranjan2019attacking, title = {Attacking Optical Flow}, author = {Ranjan, Anurag and Joel Janai and Andreas Geiger and Michael J. Black}, booktitle = {International Conference on Computer Vision (ICCV)}, year = {2019}, url = {http://flowattack.is.tue.mpg.de/} }