RadOT-Eval is an optimal-transport framework that aligns attribute-structured evidence units between radiology reports and predicts human-annotated error burden with higher Spearman correlations than standard metrics or GREEN-radllama2-7B.
ReX- Err: Synthesizing clinically meaningful errors in diagnostic radiology reports,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
RadOT-Eval: Auditable Structured-Evidence Transport for Radiology Report Evaluation
RadOT-Eval is an optimal-transport framework that aligns attribute-structured evidence units between radiology reports and predicts human-annotated error burden with higher Spearman correlations than standard metrics or GREEN-radllama2-7B.