English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In the domain of medical imaging, many supervised learning based methods forsegmentation face several challenges such as high variability in annotationsfrom multiple experts, paucity of labelled data and class imbalanced datasets.These issues may result in segmentations that lack the requisite precision forclinical analysis and can be misleadingly overconfident without associateduncertainty quantification. We propose the PULASki for biomedical imagesegmentation that accurately captures variability in expert annotations, evenin small datasets. Our approach makes use of an improved loss function based onstatistical distances in a conditional variational autoencoder structure(Probabilistic UNet), which improves learning of the conditional decodercompared to the standard cross-entropy particularly in class imbalancedproblems. We analyse our method for two structurally different segmentationtasks (intracranial vessel and multiple sclerosis (MS) lesion) and compare ourresults to four well-established baselines in terms of quantitative metrics andqualitative output. Empirical results demonstrate the PULASKi methodoutperforms all baselines at the 5\% significance level. The generatedsegmentations are shown to be much more anatomically plausible than in the 2Dcase, particularly for the vessel task. Our method can also be applied to awide range of multi-label segmentation tasks and and is useful for downstreamtasks such as hemodynamic modelling (computational fluid dynamics and dataassimilation), clinical decision making, and treatment planning.

PULASki: Learning inter-rater variability using statistical distances to improve probabilistic segmentation