← Back to Model Beat
Open Source·Aug 12·all news from August 12, 2026

RadFusion: Towards Threshold-Controllable Radiology Report Generation

Researchers have introduced RadFusion, a model designed to generate radiology reports while allowing users to adjust the balance between sensitivity and specificity in diagnostic findings. This development addresses a limitation in current automated reporting systems, which typically lack the nuanced control required to manage clinical error thresholds. By enabling customizable diagnostic output, the technology aims to better support clinical decision-making amid global radiologist shortages.

Covered by 1 source

  • AarXiv CS.AI↗Ying Jin, Noel C. F. Codella, John Corring, Mu Wei, Dinei Florencio, Eric HorvitzAug 12

Related stories

Open SourceChina's Zhipu Says Open-Source GLM-5.3 Outperforms Anthropic's Restricted Model in Vulnerability DetectionAug 14 · 7 sourcesOpen SourceMotif 3: Technical ReportAug 11Open SourceOh Lord, AI Reporters Are Actually Breaking Big NewsAug 12 · 3 sourcesOpen SourceSimple Token-Efficient Vision-Language Model for Case-level Pathology Synoptic Report GenerationAug 11