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

Annotating medical images for disease detection is often tedious andexpensive. Moreover, the available training samples for a given task aregenerally scarce and imbalanced. These conditions are not conducive forlearning effective deep neural models. Hence, it is common to 'transfer' neuralnetworks trained on natural images to the medical image domain. However, thisparadigm lacks in performance due to the large domain gap between the naturaland medical image data. To address that, we propose a novel concept of Pre-textRepresentation Transfer (PRT). In contrast to the conventional transferlearning, which fine-tunes a source model after replacing its classificationlayers, PRT retains the original classification layers and updates therepresentation layers through an unsupervised pre-text task. The task isperformed with (original, not synthetic) medical images, without utilizing anyannotations. This enables representation transfer with a large amount oftraining data. This high-fidelity representation transfer allows us to use theresulting model as a more effective feature extractor. Moreover, we can alsosubsequently perform the traditional transfer learning with this model. Wedevise a collaborative representation based classification layer for the casewhen we leverage the model as a feature extractor. We fuse the output of thislayer with the predictions of a model induced with the traditional transferlearning performed over our pre-text transferred model. The utility of ourtechnique for limited and imbalanced data classification problem isdemonstrated with an extensive five-fold evaluation for three large-scalemodels, tested for five different class-imbalance ratios for CT based COVID-19detection. Our results show a consistent gain over the conventional transferlearning with the proposed method.

Pre-text Representation Transfer for Deep Learning with Limited Imbalanced Data : Application to CT-based COVID-19 Detection