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

In medical imaging, the heterogeneity of multi-centre data impedes theapplicability of deep learning-based methods and results in significantperformance degradation when applying models in an unseen data domain, e.g. anew centreor a new scanner. In this paper, we propose an unsupervised domainadaptation framework for boosting image segmentation performance acrossmultiple domains without using any manual annotations from the new targetdomains, but by re-calibrating the networks on few images from the targetdomain. To achieve this, we enforce architectures to be adaptive to new data byrejecting improbable segmentation patterns and implicitly learning throughsemantic and boundary information, thus to capture disease-specific spatialpatterns in an adversarial optimization. The adaptation process needscontinuous monitoring, however, as we cannot assume the presence ofground-truth masks for the target domain, we propose two new metrics to monitorthe adaptation process, and strategies to train the segmentation algorithm in astable fashion. We build upon well-established 2D and 3D architectures andperform extensive experiments on three cross-centre brain lesion segmentationtasks, involving multicentre public and in-house datasets. We demonstrate thatrecalibrating the deep networks on a few unlabeled images from the targetdomain improves the segmentation accuracy significantly.

Domain Adaptive Medical Image Segmentation via Adversarial Learning of Disease-Specific Spatial Patterns