CT-IDP derives over 900 quantitative phenotypes from multi-organ CT segmentations and uses sparse logistic regression to classify diseases, achieving macro-AUCs of 0.897/0.877/0.780 on MERLIN/Duke-Abdomen/AMOS datasets and outperforming a DINOv3 vision transformer baseline.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
CT-IDP: Segmentation-Derived Quantitative Phenotypes for Interpretable Abdominal CT Disease Classification
CT-IDP derives over 900 quantitative phenotypes from multi-organ CT segmentations and uses sparse logistic regression to classify diseases, achieving macro-AUCs of 0.897/0.877/0.780 on MERLIN/Duke-Abdomen/AMOS datasets and outperforming a DINOv3 vision transformer baseline.