[2603.20777] OmniPatch: A Universal Adversarial Patch for ViT-CNN Cross-Architecture Transfer in Semantic Segmentation

Computer Science > Machine Learning arXiv:2603.20777 (cs) [Submitted on 21 Mar 2026] Title:OmniPatch: A Universal Adversarial Patch for ViT-CNN Cross-Architecture Transfer in Semantic Segmentation Authors:Aarush Aggarwal, Akshat Tomar, Amritanshu Tiwari, Sargam Goyal View a PDF of the paper titled OmniPatch: A Universal Adversarial Patch for ViT-CNN Cross-Architecture Transfer in Semantic Segmentation, by Aarush Aggarwal and 3 other authors View PDF HTML (experimental) Abstract:Robust semantic segmentation is crucial for safe autonomous driving, yet deployed models remain vulnerable to black-box adversarial attacks when target weights are unknown. Most existing approaches either craft image-wide perturbations or optimize patches for a single architecture, which limits their practicality and transferability. We introduce OmniPatch, a training framework for learning a universal adversarial patch that generalizes across images and both ViT and CNN architectures without requiring access to target model parameters. Comments: Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2603.20777 [cs.LG]   (or arXiv:2603.20777v1 [cs.LG] for this version)   https://doi.org/10.48550/arXiv.2603.20777 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Akshat Tomar [view email] [v1] Sat, 21 Mar 2026 12:07:42 UTC (5,031 KB) Full-text links: Access Paper: View a PDF of ...