VSF-Med: A Vulnerability Scoring Framework for Medical Vision-Language Models
Published in IEEE ISBI 2026 (pilot abstract), 2026
| Project Page | arXiv |
VSF-Med is an early pilot that scores vulnerabilities in medical Vision-Language Models (VLMs) using text-prompt attack templates, imperceptible visual perturbations, and an eight-dimensional risk rubric. It established the initial paraphrase-vulnerability and attention-stability setup that matured into PSF-Med. The ISBI contribution was a pilot abstract, not a peer-reviewed full paper.
Recommended citation: Sadanandan, B., & Behzadan, V. (2025). VSF-Med: A vulnerability scoring framework for medical vision-language models. IEEE ISBI 2026 (pilot abstract). arXiv:2507.00052.
Download Paper