Reproduce “Clinically-Grounded Hierarchical Classification for Consistent Chest X-ray Interpretation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jong Hak Moon, Minjun Kim
CHASE links chest X-ray findings to broad anatomical regions and sub-regions in a 54-label, three-level hierarchy. A shared backbone predicts each level, while alignment and hierarchy-violation penalties encourage coherent probabilities. On 3,403 test images, it achieves Leaf AUC of 0.848±0.004 and probabilistic consistency of 74.26±0.115%, outperforming adapted H-CAST on both measures; results are limited to the evaluated MIMIC-CXR-derived task.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.