LUMIC

LUMIC数据集损害了564例患者和1043例结肠镜检查程序的11276图像，这些患者在马尔马拉大学医学院的胃肠病学系接受了2011年12月和2019年7月之间的溃疡性结肠炎的结肠镜检查。两名经验丰富的胃肠病专家根据内窥镜Mayo评分 (EMS) 对所有图像进行了盲目审查和分类。由两个审阅者标记不同的图像也由第三个经验丰富的审阅者独立标记，而没有看到他们以前的标签。使用多数投票确定了不同标签图像的最终EMS。

The LUMIC dataset comprises 11,276 images derived from 1,043 colonoscopy procedures performed on 564 patients. These patients underwent colonoscopies for ulcerative colitis at the Department of Gastroenterology, School of Medicine, Marmara University between December 2011 and July 2019. All images were blindly reviewed and categorized by two experienced gastroenterologists using the Endoscopic Mayo Score (EMS). For images where the two reviewers assigned discrepant labels, a third experienced gastroenterologist independently annotated the images without access to the prior labels from the two initial reviewers. The final EMS classification for these discrepantly labeled images was determined via majority voting.

搜集汇总

背景与挑战

背景概述

LUMIC数据集包含来自564例患者的11276张结肠镜检查图像，用于溃疡性结肠炎的内窥镜Mayo评分(EMS)分类。数据集由马尔马拉大学和中东科技大学于2018年发布，图像由经验丰富的胃肠病专家进行标记和审查。