Adversarial Framing for Image and Video Classification

December 11, 2018 · Entered Twilight · 🏛 AAAI Conference on Artificial Intelligence

"Last commit was 6.0 years ago (≥5 year threshold)"

Evidence collected by the PWNC Scanner

Repo contents: .gitignore, LICENSE, README.md, assets, deps, draw_examples_imagenet.py, framing.py, imagenet.py, logger.py, main.py, pretrained, requirements.txt, resnet.py, trainer.py, ucf101.py, utils.py

Authors Konrad Zolna, Michal Zajac, Negar Rostamzadeh, Pedro O. Pinheiro arXiv ID 1812.04599 Category cs.CV: Computer Vision Cross-listed cs.AI, cs.LG, stat.ML Citations 62 Venue AAAI Conference on Artificial Intelligence Repository https://github.com/zajaczajac/adv_framing ⭐ 22 Last Checked 1 month ago

Abstract

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. Source code can be found at https://github.com/zajaczajac/adv_framing .