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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Deep learning is revolutionising pathology, offering novel opportunities indisease prognosis and personalised treatment. Historically, stain normalisationhas been a crucial preprocessing step in computational pathology pipelines, andpersists into the deep learning era. Yet, with the emergence of featureextractors trained using self-supervised learning (SSL) on diverse pathologydatasets, we call this practice into question. In an empirical evaluation ofpublicly available feature extractors, we find that omitting stainnormalisation and image augmentations does not compromise downstreamperformance, while incurring substantial savings in memory and compute.Further, we show that the top-performing feature extractors are remarkablyrobust to variations in stain and augmentations like rotation in their latentspace. Contrary to previous patch-level benchmarking studies, our approachemphasises clinical relevance by focusing on slide-level prediction tasks in aweakly supervised setting with external validation cohorts. This workrepresents the most comprehensive robustness evaluation of public pathology SSLfeature extractors to date, involving more than 6,000 training runs across ninetasks, five datasets, three downstream architectures, and various preprocessingsetups. Our findings stand to streamline digital pathology workflows byminimising preprocessing needs and informing the selection of featureextractors.

A Good Feature Extractor Is All You Need for Weakly Supervised Learning in Histopathology