2D U-Net for segmentation of the four muscle compartments on MRI.
journal contribution
posted on 2024-09-06, 17:56 authored by Seung-Ah Lee, Hyun Su Kim, Ehwa Yang, Young Cheol Yoon, Ji Hyun Lee, Byung-Ok Choi, Jae-Hun Kim2D U-Net for segmentation of the four muscle compartments on MRI.
History
Usage metrics
Categories
Keywords
per subject exhibitedbased labeling strategyaverage dice coefficients97 ± 984 ± 379 ± 442 ± 439 ± 329 ± 403 ± 3slices strategy showedslices per subjectground truth segmentationautomated muscle segmentationleveraging unlabeled datausing axial t1labeled set models07 %; lowermostusing unlabeled datasupervised model showedsignificantly higher adcdiv >< p46 %) compared41 %; centralsupervised learning modelslabeling three sliceshighest segmentation performancesupervised model usinglower leg mrislowermost sliceslabeled datasupervised learningsegmentation performancemodel performanceusing pairedstrategies usingweighted mristooth diseasep net architectureefficient methodbonferroni correction82 %,71 %,2d u
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC