
A unified unsupervised domain adaptation framework, MAPSeg, leverages masked autoencoding and masked pseudo-labeling to enable versatile adaptation across various domain shifts in medical image segmentation, including cross-modality, cross-sequence, cross-site, and cross-age.


coremsg

masked-autoencoding-and-pseudo-labeling-for-versatile-unsupervised-domain-adaptation-in-heterogeneous-medical-image-segmentation


Masked Autoencoding and Pseudo-Labeling for Versatile Unsupervised Domain Adaptation in Heterogeneous Medical Image Segmentation


title_rewrite


The authors propose an enhanced Filtered Pseudo Label (FPL+) framework to effectively adapt a medical image segmentation model from one modality to another without requiring labeled data in the target domain. The key innovations include cross-domain data augmentation, a dual-domain pseudo label generator, and joint training with image-level and pixel-level weighting to mitigate the impact of noisy pseudo labels.


enhancing-unsupervised-cross-modality-adaptation-for-3d-medical-image-segmentation-with-filtered-pseudo-labels


Enhancing Unsupervised Cross-Modality Adaptation for 3D Medical Image Segmentation with Filtered Pseudo Labels
