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

CXRs are a crucial and extraordinarily common diagnostic tool, leading toheavy research for CAD solutions. However, both high classification accuracyand meaningful model predictions that respect and incorporate clinicaltaxonomies are crucial for CAD usability. To this end, we present a deep HMLCapproach for CXR CAD. Different than other hierarchical systems, we show thatfirst training the network to model conditional probability directly and thenrefining it with unconditional probabilities is key in boosting performance. Inaddition, we also formulate a numerically stable cross-entropy loss functionfor unconditional probabilities that provides concrete performanceimprovements. Finally, we demonstrate that HMLC can be an effective means tomanage missing or incomplete labels. To the best of our knowledge, we are thefirst to apply HMLC to medical imaging CAD. We extensively evaluate ourapproach on detecting abnormality labels from the CXR arm of the PLCO dataset,which comprises over $198,000$ manually annotated CXRs. When using completelabels, we report a mean AUC of 0.887, the highest yet reported for thisdataset. These results are supported by ancillary experiments on the PadChestdataset, where we also report significant improvements, 1.2% and 4.1% in AUCand AP, respectively over strong "flat" classifiers. Finally, we demonstratethat our HMLC approach can much better handle incompletely labelled data. Theseperformance improvements, combined with the inherent usefulness of taxonomicpredictions, indicate that our approach represents a useful step forward forCXR CAD.

Deep Hiearchical Multi-Label Classification Applied to Chest X-Ray Abnormality Taxonomies