AI Paper Atlas

Research map / Detection and segmentation / Tracking

Towards Universal Physical Attacks on Single Object Tracking

Li Ding, Yongwei Wang, Kaiwen Yuan, Minyang Jiang, Ping Wang, Hua Huang, Z. Jane Wang

AAAI 2021 · Detection and segmentation › Tracking · approach: Foundation-model and correlation-filter tracking · 33 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
Use a universal adversarial patch with MTD and dilation shape lossSiamese visual trackers · LaSOT person, car, and bottletracking performance drop: Person Success: 8.4, 42.5; Car Success: 46.9, 57.5; Bottle Success: 49.5, 38.2Improved
Use a universal adversarial patch with MTD and shrinking shape lossSiamese visual trackers · LaSOT person, car, and bottletracking performance drop: Person Success: 38.8, 65.1; Car Success: 32.2, 44.5; Bottle Success: 45.7, 25.4Mixed
Add MTD loss to the dilation attack (ablation)SiamMask · LaSOT personsuccess performance drop: 28.5 -> 37.0Improved
Add MTD loss to the shrinking attack (ablation)SiamMask · LaSOT personsuccess performance drop: 53.7 -> 65.1Improved
Add MTD loss to the dilation attack (ablation)SiamMask · LaSOT personprecision performance drop: 26.5 -> 35.8Improved
Add MTD loss to the shrinking attack (ablation)SiamMask · LaSOT personprecision performance drop: 55.5 -> 74.2Improved
Add MTD loss to the dilation attack (ablation)SiamRPN++ · LaSOT personsuccess performance drop: 31.2 -> 42.5Improved
Add MTD loss to the shrinking attack (ablation)SiamRPN++ · LaSOT personsuccess performance drop: 14.5 -> 38.8Improved
Increase patch size ratio from 15% to 35% (ablation)Universal physical attack · LaSOTtracking performanceImproved
Remove the TV smoothness loss (ablation)Universal physical attack · Physical experimentsattack effectivenessHurt
Remove the TV smoothness loss (ablation)Universal physical attack · Digital simulationsattack performanceNo clear effect
Use an untrained random patch instead of an optimized adversarial patchSiamese visual trackers · LaSOT person, car, and bottletracking performance: Person Success: -24.7, -11.9, -2.7; Car Success: -11.9, -2.7; Bottle Success: -9.9, -16.5Hurt

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