Researchers have introduced RAM-H1200, a new dataset and evaluation benchmark designed for comprehensive analysis of rheumatoid arthritis (RA) from hand radiographs. The dataset comprises 1,200 radiographs from six medical centers, featuring detailed annotations for whole-hand bone segmentation, pixel-level bone erosion masks, and joint-level scores for bone erosion and joint space narrowing. This resource aims to enable quantitative analysis of bone erosion, which has been a significant challenge, and to jointly evaluate anatomical structure modeling and clinical severity assessment. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Introduces a new benchmark for AI models to improve diagnosis and quantitative analysis of rheumatoid arthritis from radiographs.
RANK_REASON This is a research paper introducing a new dataset and benchmark for medical image analysis. [lever_c_demoted from research: ic=1 ai=1.0]