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/

Zendy Plus

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

Zendy Tools

Zendy Open

Medical image segmentation plays an irreplaceable role in computer-assisteddiagnosis, treatment planning, and following-up. Collecting and annotating alarge-scale dataset is crucial to training a powerful segmentation model, butproducing high-quality segmentation masks is an expensive and time-consumingprocedure. Recently, weakly-supervised learning that uses sparse annotations(points, scribbles, bounding boxes) for network training has achievedencouraging performance and shown the potential for annotation cost reduction.However, due to the limited supervision signal of sparse annotations, it isstill challenging to employ them for networks training directly. In this work,we propose a simple yet efficient scribble-supervised image segmentation methodand apply it to cardiac MRI segmentation. Specifically, we employ a dual-branchnetwork with one encoder and two slightly different decoders for imagesegmentation and dynamically mix the two decoders' predictions to generatepseudo labels for auxiliary supervision. By combining the scribble supervisionand auxiliary pseudo labels supervision, the dual-branch network canefficiently learn from scribble annotations end-to-end. Experiments on thepublic ACDC dataset show that our method performs better than currentscribble-supervised segmentation methods and also outperforms severalsemi-supervised segmentation methods.

Scribble-Supervised Medical Image Segmentation via Dual-Branch Network and Dynamically Mixed Pseudo Labels Supervision