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

The process of annotating histological gigapixel-sized whole slide images(WSIs) at the pixel level for the purpose of training a supervised segmentationmodel is time-consuming. Region-based active learning (AL) involves trainingthe model on a limited number of annotated image regions instead of requestingannotations of the entire images. These annotation regions are iterativelyselected, with the goal of optimizing model performance while minimizing theannotated area. The standard method for region selection evaluates theinformativeness of all square regions of a specified size and then selects aspecific quantity of the most informative regions. We find that the efficiencyof this method highly depends on the choice of AL step size (i.e., thecombination of region size and the number of selected regions per WSI), and asuboptimal AL step size can result in redundant annotation requests or inflatedcomputation costs. This paper introduces a novel technique for selectingannotation regions adaptively, mitigating the reliance on this ALhyperparameter. Specifically, we dynamically determine each region by firstidentifying an informative area and then detecting its optimal bounding box, asopposed to selecting regions of a uniform predefined shape and size as in thestandard method. We evaluate our method using the task of breast cancermetastases segmentation on the public CAMELYON16 dataset and show that itconsistently achieves higher sampling efficiency than the standard methodacross various AL step sizes. With only 2.6\% of tissue area annotated, weachieve full annotation performance and thereby substantially reduce the costsof annotating a WSI dataset. The source code is available athttps://github.com/DeepMicroscopy/AdaptiveRegionSelection.

Adaptive Region Selection for Active Learning in Whole Slide Image Semantic Segmentation