RAM-W600
收藏资源简介:
RAM-W600是一个多任务腕骨数据集和基准,用于类风湿性关节炎(RA)的研究。该数据集由东京理科大学、北海道大学和东京大学的研究团队创建,包含来自四个医疗中心的227名患者的621张腕部常规放射照片。其中,443张高分辨率照片带有专家验证的像素级实例分割注释,用于腕骨分割,以及548张照片带有SvdH BE评分。该数据集旨在支持与RA相关的广泛研究任务,如关节间隙狭窄(JSN)进展量化、骨侵蚀(BE)检测、骨变形评估和骨赘检测等。数据集的创建过程遵循了严格的伦理标准和专业的标注流程,以确保数据的高质量。该数据集的发布有望降低腕部RA研究门槛,加速计算机辅助诊断(CAD)在RA相关领域的进展。
RAM-W600 is a multi-task wrist bone dataset and benchmark for rheumatoid arthritis (RA) research. This dataset was created by research teams from Tokyo University of Science, Hokkaido University, and The University of Tokyo, containing 621 routine wrist radiographs from 227 patients across four medical centers. Of these, 443 high-resolution radiographs have expert-validated pixel-level instance segmentation annotations for wrist bone segmentation, while 548 radiographs are equipped with SvdH BE scores. This dataset aims to support a wide range of RA-related research tasks, such as quantification of joint space narrowing (JSN) progression, bone erosion (BE) detection, bone deformation assessment, and osteophyte detection. The dataset was developed following strict ethical standards and professional annotation workflows to ensure high data quality. The release of this dataset is expected to lower the barrier to entry for wrist-related RA research and accelerate the progress of computer-aided diagnosis (CAD) in RA-related fields.




