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/

Zendy Plus

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

Zendy Tools

Zendy Open

Amidst the ongoing pandemic, several studies have shown that COVID-19classification and grading using computed tomography (CT) images can beautomated with convolutional neural networks (CNNs). Many of these studiesfocused on reporting initial results of algorithms that were assembled fromcommonly used components. The choice of these components was often pragmaticrather than systematic. For instance, several studies used 2D CNNs even thoughthese might not be optimal for handling 3D CT volumes. This paper identifies avariety of components that increase the performance of CNN-based algorithms forCOVID-19 grading from CT images. We investigated the effectiveness of using a3D CNN instead of a 2D CNN, of using transfer learning to initialize thenetwork, of providing automatically computed lesion maps as additional networkinput, and of predicting a continuous instead of a categorical output. A 3D CNNwith these components achieved an area under the ROC curve (AUC) of 0.934 onour test set of 105 CT scans and an AUC of 0.923 on a publicly available set of742 CT scans, a substantial improvement in comparison with a previouslypublished 2D CNN. An ablation study demonstrated that in addition to using a 3DCNN instead of a 2D CNN transfer learning contributed the most and continuousoutput contributed the least to improving the model performance.

Improving Automated COVID-19 Grading with Convolutional Neural Networks in Computed Tomography Scans: An Ablation Study