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

Deep learning based automatic medical image segmentation models often sufferfrom domain shift, where the models trained on a source domain do notgeneralize well to other unseen domains. As a vision foundation model withpowerful generalization capabilities, Segment Anything Model (SAM) showspotential for improving the cross-domain robustness of medical imagesegmentation. However, SAM and its fine-tuned models performed significantlyworse in fully automatic mode compared to when given manual prompts. Uponfurther investigation, we discovered that the degradation in performance wasrelated to the coupling effect of poor prompts and mask segmentation. In fullyautomatic mode, the presence of inevitable poor prompts (such as points outsidethe mask or boxes significantly larger than the mask) can significantly misleadmask generation. To address the coupling effect, we propose the decoupling SAM(DeSAM). DeSAM modifies SAM's mask decoder to decouple mask generation andprompt embeddings while leveraging pre-trained weights. We conductedexperiments on publicly available prostate cross-site datasets. The resultsshow that DeSAM improves dice score by an average of 8.96% (from 70.06% to79.02%) compared to previous state-of-the-art domain generalization method.Moreover, DeSAM can be trained on personal devices with entry-level GPU sinceour approach does not rely on tuning the heavyweight image encoder. The code ispublicly available at https://github.com/yifangao112/DeSAM.

DeSAM: Decoupling Segment Anything Model for Generalizable Medical Image Segmentation