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

The ADMANI datasets (annotated digital mammograms and associated non-imagedatasets) from the Transforming Breast Cancer Screening with AI programme(BRAIx) run by BreastScreen Victoria in Australia are multi-centre, largescale, clinically curated, real-world databases. The datasets are expected toaid in the development of clinically relevant Artificial Intelligence (AI)algorithms for breast cancer detection, early diagnosis, and otherapplications. To ensure high data quality, technical outliers must be removedbefore any downstream algorithm development. As a first step, we randomlyselect 30,000 individual mammograms and use Convolutional VariationalAutoencoder (CVAE), a deep generative neural network, to detect outliers. CVAEis expected to detect all sorts of outliers, although its detection performancediffers among different types of outliers. Traditional image processingtechniques such as erosion and pectoral muscle analysis can compensate for thepoor performance of CVAE in certain outlier types. We identify seven types oftechnical outliers: implant, pacemaker, cardiac loop recorder, improperradiography, atypical lesion/calcification, incorrect exposure parameter andimproper placement. The outlier recall rate for the test set is 61% if CVAE,erosion and pectoral muscle analysis each select the top 1% images ranked inascending or descending order according to image outlier score under eachdetection method, and 83% if each selects the top 5% images. This study offersan overview of technical outliers in the ADMANI dataset and suggests futuredirections to improve outlier detection effectiveness.

Technical outlier detection via convolutional variational autoencoder for the ADMANI breast mammogram dataset