EulerianKnight/breast-histopathology-images-train-test-valid-split
收藏资源简介:
该数据集是对Kaggle上的原始乳腺癌组织病理学图像数据集的重排。原始数据集的结构是按照患者ID和类别(0和1)组织的,而当前数据集的结构被重新组织为训练集、验证集和测试集,每个集合下都有类别(0和1)的子目录。数据集的任务类别是图像分类,大小在100K到1M之间。
This dataset is a reorganized version of the original breast cancer histopathological image dataset hosted on Kaggle. The original dataset was structured based on patient IDs and two categories (0 and 1). In comparison, the current dataset has been restructured into training, validation, and test sets, each of which contains subdirectories corresponding to the two categories (0 and 1). The downstream task of this dataset is image classification, and its data scale ranges from 100K to 1M.
Breast Histopathology Image dataset 概述
数据集基本信息
- 许可证: Apache-2.0
- 任务类别: 图像分类
- 大小类别: 100K<n<1M
数据集来源
- 本数据集是对Kaggle上的原始数据集(https://www.kaggle.com/datasets/paultimothymooney/breast-histopathology-images)的重新组织。
数据引用
- 数据引用信息可参考:https://www.ncbi.nlm.nih.gov/pubmed/27563488 和 http://spie.org/Publications/Proceedings/Paper/10.1117/12.2043872
数据集结构
-
原始数据集结构:
|-- patient_id |-- class(0 and 1)
-
当前数据集结构:
|-- train |-- class(0 and 1) |-- valid |-- class(0 and 1) |-- test |-- class(0 and 1)




