AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

APRICOT: A Dataset of Physical Adversarial Attacks on Object Detection

A. Braunegg, Amartya Chakraborty, Michael Krumdick, Nicole Lape, Sara Leary, Keith Manville, Elizabeth Merkhofer, Laura Strickhart, Matthew Walmer

ECCV 2020 · Adversarial robustness and security › Attacks on text and graphs · approach: Physical adversarial patches and camouflage · 57 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
move patches from digital insertion to physical photographsFRCNN white-box patch · targeted fooling ratefooling event rate: 0.50 -> 0.17Hurt
move from digital to physicalall patches · untargeted fooling rate on FRCNNfooling event rate: 0.77 -> 0.38Hurt
increase patch size from small to largeFRCNN patch · FRCNN white-box digital targetedfooling event rate: 0.07 -> 0.83Improved
use square instead of circle patchesFRCNN patch · FRCNN white-box physical targetedfooling event rate: 0.23 -> 0.12Hurt
use square instead of circle patchesFRCNN patch · FRCNN white-box digital targetedfooling event rate: 0.47 -> 0.53No clear effect
increase viewing angle from head-on to severeFRCNN patch · FRCNN white-box physical targetedfooling event rate: 0.21 -> 0.0Hurt
transfer patch to a different detector (black-box)SSD patch on FRCNN vs SSD · digital targetedfooling event rate: 0.23 -> 0.08Hurt
jointly train to detect patches and COCO objects (ablation)patch detector trained on high confidence patches · APRICOT AP IoU 0.5:0.95AP: 0.540 -> 0.686Improved
jointly train to detect patches and COCO objects (ablation)patch detector trained on Gaussian patches · APRICOT AP IoU 0.5AP: 0.478 -> 0.936Improved
train on high confidence adversarial patches instead of standard adversarial patches (ablation)joint patch detector · APRICOT AP IoU 0.5:0.95AP: 0.670 -> 0.686Improved
train on adversarial patches instead of Gaussian noise patches (ablation)joint patch detector · APRICOT AP IoU 0.5AP: 0.936 -> 0.937No clear effect
combine KDE and Bayesian uncertainty by logistic regression instead of uncertainty alone (ablation)fooling-event flagging · APRICOT detectionsROCNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy