胃癌诊断数据集|医学图像分析数据集

GasHisSDB是由东北大学等机构合作开发的胃癌组织病理学图像数据库，包含245,196张经H&E染色的正常与异常胃癌病理图像。数据集分为三个子库，对应不同图像尺寸，支持多尺度分析。该数据集规模大、标注清晰，适用于胃癌早期诊断研究中的深度学习模型训练与评估，可广泛应用于医学图像分类、分割及辅助诊断系统开发。

GasHisSDB is a gastric cancer histopathological image database jointly developed by Northeastern University and other institutions. It contains 245,196 H&E-stained normal and abnormal gastric cancer histopathological images. The dataset is divided into three sub-databases corresponding to different image sizes, which supports multi-scale analysis. Featuring large scale and clear annotations, this dataset is suitable for training and evaluating deep learning models in gastric cancer early diagnosis research, and can be widely used in the development of medical image classification, segmentation and auxiliary diagnosis systems.

搜集汇总

背景与挑战

背景概述

GasHisSDB是一个大规模的胃癌组织病理学图像数据库，包含24.5万张H&E染色的正常与异常病理图像，分为三种像素尺寸的子库以支持多尺度分析。该数据集标注清晰，适用于胃癌早期诊断的深度学习和医学图像分析研究。