Treating voxel-wise annotator agreement as an ordered target and combining Ranked Probability Score loss with binary segmentation objectives improves calibration to inter-rater variability without degrading accuracy.
Trustworthy clinical AI solutions: A unified review of uncertainty quantification in Deep Learning models for medical image analysis
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Multi-Rater Calibrated Segmentation Models
Treating voxel-wise annotator agreement as an ordered target and combining Ranked Probability Score loss with binary segmentation objectives improves calibration to inter-rater variability without degrading accuracy.