Physical Adversarial Textures that Fool Visual Object Tracking
This work addresses a security vulnerability in visual object tracking for applications like robotics, though it is incremental in focusing on a specific regression task.
The paper tackles the problem of fooling visual object tracking systems by generating physical adversarial textures that cause trackers to lock onto them, allowing targets to evade; it compares optimization strategies and demonstrates that textures optimized in simulation can confuse real-world trackers.
We present a system for generating inconspicuous-looking textures that, when displayed in the physical world as digital or printed posters, cause visual object tracking systems to become confused. For instance, as a target being tracked by a robot's camera moves in front of such a poster, our generated texture makes the tracker lock onto it and allows the target to evade. This work aims to fool seldom-targeted regression tasks, and in particular compares diverse optimization strategies: non-targeted, targeted, and a new family of guided adversarial losses. While we use the Expectation Over Transformation (EOT) algorithm to generate physical adversaries that fool tracking models when imaged under diverse conditions, we compare the impacts of different conditioning variables, including viewpoint, lighting, and appearances, to find practical attack setups with high resulting adversarial strength and convergence speed. We further showcase textures optimized solely using simulated scenes can confuse real-world tracking systems.