Leveraging Probabilistic Segmentation Models for Improved Glaucoma Diagnosis: A Clinical Pipeline Approach
Abstract
The accurate segmentation of the optic cup and disc in fundus images is essential for diagnostic processes such as glaucoma detection. The inherent ambiguity in locating these structures often poses a significant challenge, leading to potential misdiagnosis. To model such ambiguities, numerous probabilistic segmentation models have been proposed. In this paper, we investigate the integration of these probabilistic segmentation models into a multistage pipeline closely resembling clinical practice. Our findings indicate that leveraging the uncertainties provided by these models substantially enhances the quality of glaucoma diagnosis compared to relying on a single segmentation only.
Cite
Text
Wundram et al. "Leveraging Probabilistic Segmentation Models for Improved Glaucoma Diagnosis: A Clinical Pipeline Approach." Proceedings of MIDL 2024, 2024.Markdown
[Wundram et al. "Leveraging Probabilistic Segmentation Models for Improved Glaucoma Diagnosis: A Clinical Pipeline Approach." Proceedings of MIDL 2024, 2024.](https://mlanthology.org/midl/2024/wundram2024midl-leveraging/)BibTeX
@inproceedings{wundram2024midl-leveraging,
title = {{Leveraging Probabilistic Segmentation Models for Improved Glaucoma Diagnosis: A Clinical Pipeline Approach}},
author = {Wundram, Anna M. and Fischer, Paul and Wunderlich, Stephan and Faber, Hanna and Koch, Lisa M. and Berens, Philipp and Baumgartner, Christian F.},
booktitle = {Proceedings of MIDL 2024},
year = {2024},
pages = {1725-1740},
volume = {250},
url = {https://mlanthology.org/midl/2024/wundram2024midl-leveraging/}
}