Attacking Motion Estimation with Adversarial Snow
This addresses security vulnerabilities in motion estimation systems for applications like autonomous driving, though it is incremental as it builds on existing adversarial attack concepts.
The paper tackles the problem of adversarial attacks on motion estimation by introducing a novel attack using adversarially optimized snow, which significantly impacts optical flow while appearing indistinguishable from real snow, with the largest effect on methods previously robust to small perturbations.
Current adversarial attacks for motion estimation (optical flow) optimize small per-pixel perturbations, which are unlikely to appear in the real world. In contrast, we exploit a real-world weather phenomenon for a novel attack with adversarially optimized snow. At the core of our attack is a differentiable renderer that consistently integrates photorealistic snowflakes with realistic motion into the 3D scene. Through optimization we obtain adversarial snow that significantly impacts the optical flow while being indistinguishable from ordinary snow. Surprisingly, the impact of our novel attack is largest on methods that previously showed a high robustness to small L_p perturbations.