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

Noisy labels collected with limited annotation cost prevent medical imagesegmentation algorithms from learning precise semantic correlations. Previoussegmentation arts of learning with noisy labels merely perform a pixel-wisemanner to preserve semantics, such as pixel-wise label correction, but neglectthe pair-wise manner. In fact, we observe that the pair-wise manner capturingaffinity relations between pixels can greatly reduce the label noise rate.Motivated by this observation, we present a novel perspective for noisymitigation by incorporating both pixel-wise and pair-wise manners, wheresupervisions are derived from noisy class and affinity labels, respectively.Unifying the pixel-wise and pair-wise manners, we propose a robust JointClass-Affinity Segmentation (JCAS) framework to combat label noise issues inmedical image segmentation. Considering the affinity in pair-wise mannerincorporates contextual dependencies, a differentiated affinity reasoning (DAR)module is devised to rectify the pixel-wise segmentation prediction byreasoning about intra-class and inter-class affinity relations. To furtherenhance the noise resistance, a class-affinity loss correction (CALC) strategyis designed to correct supervision signals via the modeled noise labeldistributions in class and affinity labels. Meanwhile, CALC strategy interactsthe pixel-wise and pair-wise manners through the theoretically derivedconsistency regularization. Extensive experiments under both synthetic andreal-world noisy labels corroborate the efficacy of the proposed JCAS frameworkwith a minimum gap towards the upper bound performance. The source code isavailable at \url{https://github.com/CityU-AIM-Group/JCAS}.

Joint Class-Affinity Loss Correction for Robust Medical Image Segmentation with Noisy Labels