Affiliation:
1. Technische Hochschule Ingolstadt
Abstract
Robustness is a key requirement for any method in medicine, especially when the method in question is being used as part of a diagnostic process. This is particularly true for artificial intelligence-based decision support systems, which, although being used as a supportive tool, will ultimately influence diagnostic assessments. In pathology, attaining clinical robustness in AI methods poses a particularly challenging task, primarily due to the extensive diversity of digital images, which humans can adapt to far more easily. This paper presents factors that contribute to this challenge, but also identifies and evaluates common solutions to counteract domain shift, which is known to deteriorate the performance of artificial intelligence models.
Publisher
Trillium GmbH Medizinischer Fachverlag
Reference5 articles.
1. Aubreville, M. et al. (2023). Mitosis domain generalization in histopathology images—the MIDOG challenge. Medical Image Analysis, 84, 102699
2. Zhou, K., et al. (2023): Domain generalization: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence. 45(4): pp. 4396–4415
3. Ganz, J., et al. (2024). Assessment of Scanner Domain Shifts in Deep Multiple Instance Learning. In BVM Workshop (pp. 137–142). Wiesbaden: Springer Fachmedien Wiesbaden
4. Weitz, Philippe, et al. "The ACROBAT 2022 Challenge: Automatic Registration Of Breast Cancer Tissue." arXiv preprint arXiv:2305.18033 (2023)
5. Fick, R. et al. (2024). Improving CNN-Based Mitosis Detection through Rescanning Annotated Glass Slides and Atypical Mitosis Subtyping. Medical Imaging with Deep Learning (MIDL) https://openreview.net/forum?id=00gWBAAbMI