LIMUC

该数据集是根据梅奥内窥镜评分（MES）进行标注的，公开发布的数据集，用于训练卷积神经网络（CNN）模型，以估算溃疡性结肠炎的严重程度。所有图像都由至少两名专业的胃肠病学家进行审查和标注。梅奥评分分布情况如下：6105张（54.14%）为梅奥0级，3052张（27.7%）为梅奥1级，1254张（11.12%）为梅奥2级，以及865张（7.67%）为梅奥3级。该数据集在保持类别比例的同时，以患者级别随机分为训练集和测试集。数据集规模包括来自564名患者的11276张图像。任务是对溃疡性结肠炎严重程度的序数回归估计。

This publicly released dataset is annotated according to the Mayo Endoscopic Score (MES), and is intended for training convolutional neural network (CNN) models to estimate the severity of ulcerative colitis. All images were reviewed and annotated by at least two professional gastroenterologists. The distribution of Mayo scores is as follows: 6,105 images (54.14%) belong to Mayo Grade 0, 3,052 images (27.7%) to Mayo Grade 1, 1,254 images (11.12%) to Mayo Grade 2, and 865 images (7.67%) to Mayo Grade 3. This dataset was randomly split into training and test sets at the patient level while preserving the original class distribution proportions. The dataset comprises 11,276 images from 564 patients. The core task is ordinal regression for estimating the severity of ulcerative colitis.

搜集汇总

背景与挑战

背景概述

LIMUC是一个包含11276张溃疡性结肠炎结肠镜图像的数据集，按照Mayo内镜评分系统分类，并提供了标准化的训练/验证/测试集划分。该数据集采用开放许可协议，适用于计算机辅助诊断研究。