Research map / Adversarial robustness and security / Attacks on text and graphs
Physical Adversarial Textures That Fool Visual Object Tracking
Rey Reza Wiyatno, Anqi Xu
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use guided loss L_ga- (shrink predicted area) instead of non-targeted loss | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | Improved |
| use targeted or other guided losses instead of non-targeted loss | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | Hurt |
| combine L_nt with most targeted or guided losses (hybrid) | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | No clear effect |
| combine L_nt and L_t= at 1:1 weights | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | Improved |
| reduce variety of background, target or lighting (-bg, -target, -light) in EOT (ablation) | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | No clear effect |
| increase diversity of target and background (+target, +bg) in EOT (ablation) | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | No clear effect |
| add randomized lighting (+light) in EOT (ablation) | PAT attack · Simulated PAT attack, GOTURN Lg,sim | convergence speed | Hurt |
| halve the poster size (ablation) | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | Hurt |
| lock target pose to poster centre (-target pose) (ablation) | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | Improved |
| change camera pose ranges (+cam pose / -cam pose) (ablation) | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | Mixed |
| increase perceptual similarity weight w_ps (ablation) | PAT imitation attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | Hurt |
| initialize texture with different patterns (white, yellow, gray, checkerboard, random) (ablation) | PAT attack · Simulated PAT attack, GOTURN Lg,sim | adversarial strength (muIOUd) | No clear effect |
Closest papers by meaning
- Towards Universal Physical Attacks on Single Object Tracking (AAAI 2021 · similarity 0.90)
- Learning To Adversarially Blur Visual Object Tracking (ICCV 2021 · similarity 0.88)
- Efficient Adversarial Attacks for Visual Object Tracking (ECCV 2020 · similarity 0.88)
- Adversarial Attacks on Event-Based Pedestrian Detectors: A Physical Approach (AAAI 2025 · similarity 0.87)
- IoU Attack: Towards Temporally Coherent Black-Box Adversarial Attack for Visual Object Tracking (CVPR 2021 · similarity 0.87)
- AdvPaint: Protecting Images from Inpainting Manipulation via Adversarial Attention Disruption (ICLR 2025 · similarity 0.86)
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.
