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

Automatic segmentation of lung lesions associated with COVID-19 in CT imagesrequires large amount of annotated volumes. Annotations mandate expertknowledge and are time-intensive to obtain through fully manual segmentationmethods. Additionally, lung lesions have large inter-patient variations, withsome pathologies having similar visual appearance as healthy lung tissues. Thisposes a challenge when applying existing semi-automatic interactivesegmentation techniques for data labelling. To address these challenges, wepropose an efficient convolutional neural networks (CNNs) that can be learnedonline while the annotator provides scribble-based interaction. To acceleratelearning from only the samples labelled through user-interactions, apatch-based approach is used for training the network. Moreover, we useweighted cross-entropy loss to address the class imbalance that may result fromuser-interactions. During online inference, the learned network is applied tothe whole input volume using a fully convolutional approach. We compare ourproposed method with state-of-the-art using synthetic scribbles and show thatit outperforms existing methods on the task of annotating lung lesionsassociated with COVID-19, achieving 16% higher Dice score while reducingexecution time by 3$\times$ and requiring 9000 lesser scribbles-based labelledvoxels. Due to the online learning aspect, our approach adapts quickly to userinput, resulting in high quality segmentation labels. Source code for ECONet isavailable at: https://github.com/masadcv/ECONet-MONAILabel.

ECONet: Efficient Convolutional Online Likelihood Network for Scribble-based Interactive Segmentation