Adversarial Framing for Image and Video Classification
cytuj
pobierz pliki
RIS BIB ENDNOTEChoose format
RIS BIB ENDNOTEAdversarial Framing for Image and Video Classification
Publication date: 2018
Schedae Informaticae, 2018, Volume 27, pp. 155 - 164
https://doi.org/10.4467/20838476SI.18.012.10417Authors
Adversarial Framing for Image and Video Classification
Neural networks are prone to adversarial attacks. In general, such attacks deteriorate the quality of the input by either slightly modifying most of its pixels, or by occluding it with a patch. In this paper, we propose a method that keeps the image unchanged and only adds an adversarial framing on the border of the image. We show empirically that our method is able to successfully attack state-of-the-art methods on both image and video classification problems. Notably, the proposed method results in a universal attack which is very fast at test time.
Information: Schedae Informaticae, 2018, Volume 27, pp. 155 - 164
Article type: Original article
Titles:
Adversarial Framing for Image and Video Classification
Adversarial Framing for Image and Video Classification
Jagiellonian University in Kraków
Element AI, Montreal, Canada
Element AI, Montreal, Canada
Published at: 2018
Article status: Open
Licence: CC BY-NC-ND
Percentage share of authors:
Article corrections:
-Publication languages:
EnglishView count: 1273
Number of downloads: 0