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

Mitosis detection is one of the fundamental tasks in computational pathology,which is extremely challenging due to the heterogeneity of mitotic cell. Mostof the current studies solve the heterogeneity in the technical aspect byincreasing the model complexity. However, lacking consideration of thebiological knowledge and the complex model design may lead to the overfittingproblem while limited the generalizability of the detection model. In thispaper, we systematically study the morphological appearances in differentmitotic phases as well as the ambiguous non-mitotic cells and identify thatbalancing the data and feature diversity can achieve better generalizability.Based on this observation, we propose a novel generalizable framework (MitDet)for mitosis detection. The data diversity is considered by the proposeddiversity-guided sample balancing (DGSB). And the feature diversity ispreserved by inter- and intra- class feature diversity-preserved module(InCDP). Stain enhancement (SE) module is introduced to enhance thedomain-relevant diversity of both data and features simultaneously. Extensiveexperiments have demonstrated that our proposed model outperforms all the SOTAapproaches in several popular mitosis detection datasets in both internal andexternal test sets using minimal annotation efforts with point annotationsonly. Comprehensive ablation studies have also proven the effectiveness of therethinking of data and feature diversity balancing. By analyzing the resultsquantitatively and qualitatively, we believe that our proposed model not onlyachieves SOTA performance but also might inspire the future studies in newperspectives. Source code is at https://github.com/Onehour0108/MitDet.

Rethinking Mitosis Detection: Towards Diverse Data and Feature Representation