OpenMIBOOD

Robustness and trustworthiness of medical AI

The growing reliance on Artificial Intelligence (AI) in critical domains such as healthcare demands robust mechanisms to ensure the trustworthiness of these systems, especially when faced with unexpected or anomalous inputs. This project investigates methods to evaluate and improve robustness in medical imaging contexts.

The Open Medical Imaging Benchmarks for Out-Of-Distribution Detection (OpenMIBOOD) framework developed in this project offers a comprehensive approach to assessing out-of-distribution (OOD) detection methods in medical imaging. OpenMIBOOD includes three benchmarks from diverse medical domains, encompassing 14 datasets divided into covariate-shifted in-distribution, near-OOD, and far-OOD categories. Results of our evaluation of post-hoc methods across these benchmarks reveal that findings from broad-scale OOD benchmarks in natural image domains do not translate to medical applications, underscoring the critical need for such benchmarks in the medical field. By mitigating the risk of exposing AI models to inputs outside their training distribution, OpenMIBOOD aims to support the advancement of reliable and trustworthy AI systems in healthcare.

OpenMIBOOD is available on GitHub .