Banshee: Target Switch Attacks on Gimbal-Stabilized Visual Tracking Systems via Acoustic Injection
Jiarui Li, Joseph Brewington, Qingzhao Zhang, Z. Morley Mao
cs.CV, cs.CR
Submitted: 2026-07-10
Comments: Accepted by the 47th IEEE Symposium on Security and Privacy (IEEE S&P 2026). 20 pages. Artifact: https://github.com/U1ltra/Banshee
Journal ref: 2026 IEEE Symposium on Security and Privacy (SP), pp. 1539-1557, 2026
DOI: 10.1109/SP63933.2026.00129
Code: https://github.com/U1ltra/Banshee
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- ControlLoc: Physical-World Hijacking Attack on Visual Perception in Autonomous Driving
- Learning Spatio-Temporal Transformer for Visual Tracking
- ATOM: Accurate Tracking by Overlap Maximization
- Learning Discriminative Model Prediction for Tracking
- FlyTrap: Physical Distance-Pulling Attack Towards Camera-based Autonomous Target Tracking Systems
- VIMU: Effective Physics-based Realtime Detection and Recovery against Stealthy Attacks on UAVs
- MARS: Defending Unmanned Aerial Vehicles From Attacks on Inertial Sensors with Model-based Anomaly Detection and Recovery
- Attacking Optical Flow
- A Perturbation-Constrained Adversarial Attack for Evaluating the Robustness of Optical Flow
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models