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

To measure the volume of specific image structures, a typical approach is tofirst segment those structures using a neural network trained on voxel-wise(strong) labels and subsequently compute the volume from the segmentation. Amore straightforward approach would be to predict the volume directly using aneural network based regression approach, trained on image-level (weak) labelsindicating volume.  In this article, we compared networks optimized with weak and strong labels,and study their ability to generalize to other datasets. We experimented withwhite matter hyperintensity (WMH) volume prediction in brain MRI scans. Neuralnetworks were trained on a large local dataset and their performance wasevaluated on four independent public datasets. We showed that networksoptimized using only weak labels reflecting WMH volume generalized better forWMH volume prediction than networks optimized with voxel-wise segmentations ofWMH. The attention maps of networks trained with weak labels did not seem todelineate WMHs, but highlighted instead areas with smooth contours around ornear WMHs. By correcting for possible confounders we showed that networkstrained on weak labels may have learnt other meaningful features that are moresuited to generalization to unseen data. Our results suggest that for imagingbiomarkers that can be derived from segmentations, training networks to predictthe biomarker directly may provide more robust results than solving anintermediate segmentation step.

When Weak Becomes Strong: Robust Quantification of White Matter Hyperintensities in Brain MRI scans