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 techniques depend on large datasets whose annotation istime-consuming. To reduce annotation burden, the self-training (ST) andactive-learning (AL) methods have been developed as well as methods thatcombine them in an iterative fashion. However, it remains unclear when eachmethod is the most useful, and when it is advantageous to combine them. In thispaper, we propose a new method that combines ST with AL using Test-TimeAugmentations (TTA). First, TTA is performed on an initial teacher network.Then, cases for annotation are selected based on the lowest estimated Dicescore. Cases with high estimated scores are used as soft pseudo-labels for ST.The selected annotated cases are trained with existing annotated cases and STcases with border slices annotations. We demonstrate the method on MRI fetalbody and placenta segmentation tasks with different data variabilitycharacteristics. Our results indicate that ST is highly effective for bothtasks, boosting performance for in-distribution (ID) and out-of-distribution(OOD) data. However, while self-training improved the performance ofsingle-sequence fetal body segmentation when combined with AL, it slightlydeteriorated performance of multi-sequence placenta segmentation on ID data. ALwas helpful for the high variability placenta data, but did not improve uponrandom selection for the single-sequence body data. For fetal body segmentationsequence transfer, combining AL with ST following ST iteration yielded a Diceof 0.961 with only 6 original scans and 2 new sequence scans. Results usingonly 15 high-variability placenta cases were similar to those using 50 cases.Code is available at: https://github.com/Bella31/TTA-quality-estimation-ST-AL

Test-time augmentation-based active learning and self-training for label-efficient segmentation