Lipschitz Optimization for Formal Verification of Homographies
The paper presents a formal verification approach for ensuring the robustness of vision neural networks against 3D motion perturbations. This method addresses a significant challenge in safety-critical applications, such as healthcare and autonomous vehicles, where camera motion can affect performance. The authors demonstrate improvements in speed and accuracy compared to previous methods, highlighting practical vulnerabilities in real-world scenarios.
- ▪The approach targets robustness against 3D motion perturbations of the capturing camera.
- ▪It establishes a closed-form mapping from camera pose to pixel values, allowing for formal verification of projective geometry transforms.
- ▪The implementation shows up to 89% speedup and 7% tighter bounds over prior work.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Computer Vision and Pattern Recognition arXiv:2605.23203 (cs) [Submitted on 22 May 2026] Title:Lipschitz Optimization for Formal Verification of Homographies Authors:Jean-Guillaume Durand, Panagiotis Kouvaros, Maxime Gariel, Alessio Lomuscio View a PDF of the paper titled Lipschitz Optimization for Formal Verification of Homographies, by Jean-Guillaume Durand and 3 other authors View PDF HTML (experimental) Abstract:The adoption of vision neural networks in regulated industries requires formal robustness guarantees, especially in safety-critical domains such as healthcare, autonomous vehicles, and aerospace.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
