Reproduce “Learning Anatomy-Grounded CT Vision-Language Representations with Organ-Hierarchical Report Knowledge” 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
Guoliang You, Hongming Li, Yuanwang Zhang, Yong Fan
OKA-CT trains a chest CT model using organ-specific fields extracted from radiology reports. It uses these fields to supervise visual features, then guide scan–report matching, assigning weak positive weight to cases with shared organ-level evidence. On CT-RATE, it reaches zero-shot abnormality macro AUROC 84.9 versus fVLM’s 77.8; findings are bounded to evaluated chest CT settings and derived weak labels.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.