In the domain of medical imaging, many supervised learning based methods for segmentation face several challenges such as high variability in annotations from multiple experts, paucity of labelled data and class imbalanced datasets. These issues may result in segmentations that lack the requisite precision for clinical analysis and can be misleadingly overconfident without associated uncertainty quantification.

We proposed the PULASki as a computationally efficient generative tool for biomedical image segmentation that accurately captures variability in expert annotations, even in small datasets. Our approach makes use of an improved loss function based on statistical distances in a conditional variational autoencoder structure (Probabilistic UNet), which improves learning of the conditional decoder compared to the standard cross-entropy particularly in class imbalanced problems.

It was introduced in PULASki: Learning inter-rater variability using statistical distances to improve probabilistic segmentation by Soumick Chatterjee, Franziska Gaidzik, Alessandro Sciarra, Hendrik Mattern, Gábor Janiga, Oliver Speck, Andreas Nürnberger, Sahani Pathiraja.

If you use this approach in your research or use codes from this repository or these weights, please cite the following in your publications:

Chatterjee, S., Gaidzik, F., Sciarra, A., Mattern, H., Janiga, G., Speck, O., Nuernberger, A., & Pathiraja, S. (2023). PULASki: Learning inter-rater variability using statistical distances to improve probabilistic segmentation. arXiv preprint arXiv:2312.15686.