Deceiving Image-to-Image Translation Networks for Autonomous Driving with Adversarial Perturbations

Cited 9 time in webofscience Cited 12 time in scopus
  • Hit : 359
  • Download : 0
Deep neural networks (DNNs) have achieved impressive performance on handling computer vision problems. However, it has been found that DNNs are vulnerable to adversarial examples. For such reason, adversarial perturbations have been recently studied in several respects. However, most previous works have focused on image classification tasks, and it has never been studied regarding adversarial perturbations on Image-to-image (Im2Im) translation tasks, showing great success in handling paired and/or unpaired mapping problems in the field of autonomous driving and robotics. This letter examines different types of adversarial perturbations that can fool Im2Im frameworks for autonomous driving purposes. We propose both quasi-physical and digital adversarial perturbations that can make Im2Im models yield unexpected results. We then empirically analyze these perturbations and show that they generalize well under both paired for image synthesis and unpaired settings for style transfer. We also validate that there exist some perturbation thresholds over which the Im2Im mapping is disrupted or impossible. The existence of these perturbations reveals that there exist crucial weaknesses in Im2Im models. Lastly, we show that our methods illustrate how perturbations affect the quality of outputs, pioneering the improvement of the robustness of current SOTA networks for autonomous driving.
Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
Issue Date
2020-04
Language
English
Article Type
Article
Citation

IEEE ROBOTICS AND AUTOMATION LETTERS, v.5, no.2, pp.1421 - 1428

ISSN
2377-3766
DOI
10.1109/LRA.2020.2967289
URI
http://hdl.handle.net/10203/274106
Appears in Collection
ME-Journal Papers(저널논문)
Files in This Item
There are no files associated with this item.
This item is cited by other documents in WoS
⊙ Detail Information in WoSⓡ Click to see webofscience_button
⊙ Cited 9 items in WoS Click to see citing articles in records_button

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0