UniToPatho
收藏资源简介:
UniToPatho是由都灵大学创建的高分辨率注释数据集,包含9536个从292张全切片图像中提取的H&E染色补丁,用于训练深度神经网络进行结直肠息肉的分类和腺瘤异型增生分级。数据集内容包括不同类型的结直肠息肉样本,大小不一,数据量丰富,来源于癌症筛查中的患者。创建过程涉及从全切片图像中提取关键补丁,并由专家进行注释。该数据集主要应用于自动化结直肠息肉特征分析,旨在提高诊断准确性和效率。
UniToPatho is a high-resolution annotated dataset developed by the University of Turin. It comprises 9536 H&E-stained patches extracted from 292 whole-slide images, and is tailored for training deep neural networks to conduct colorectal polyp classification and adenoma dysplasia grading. The dataset encompasses colorectal polyp samples of diverse types and varying sizes, with a large volume of data sourced from patients undergoing cancer screening. The dataset construction process entails extracting critical patches from whole-slide images and performing expert annotations. This dataset is primarily utilized for automated colorectal polyp feature analysis, with the objective of improving diagnostic accuracy and efficiency.




