English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recent work demonstrated the lack of robustness of optical flow networks tophysical patch-based adversarial attacks. The possibility to physically attacka basic component of automotive systems is a reason for serious concerns. Inthis paper, we analyze the cause of the problem and show that the lack ofrobustness is rooted in the classical aperture problem of optical flowestimation in combination with bad choices in the details of the networkarchitecture. We show how these mistakes can be rectified in order to makeoptical flow networks robust to physical patch-based attacks. Additionally, wetake a look at global white-box attacks in the scope of optical flow. We findthat targeted white-box attacks can be crafted to bias flow estimation modelstowards any desired output, but this requires access to the input images andmodel weights. However, in the case of universal attacks, we find that opticalflow networks are robust. Code is available athttps://github.com/lmb-freiburg/understanding_flow_robustness.

Towards Understanding Adversarial Robustness of Optical Flow Networks