BCI乳腺癌免疫组化图像

BCI数据集由北京朝阳医院与首都医科大学合作构建，专注于乳腺癌免疫组化图像生成，包含4872对结构级对齐的H&E与IHC染色图像，涵盖HER2四种表达水平（0、1+、2+、3+），涉及51对全切片图像。数据经脱敏处理，支持HE到IHC的图像转换算法研究。该数据集为深度学习驱动的病理图像合成提供了高质量配对数据，已应用于Pyramid Pix2pix等模型，推动非商业学术研究与病理诊断辅助技术发展。

The BCI dataset was collaboratively constructed by Beijing Chaoyang Hospital and Capital Medical University, focusing on the generation of breast cancer immunohistochemical (IHC) stained images. It contains 4872 pairs of structure-level aligned Hematoxylin and Eosin (H&E) stained and IHC stained images, covering four HER2 expression levels (0, 1+, 2+, 3+) and including 51 pairs of whole-slide images. The data has been de-identified, supporting research on image translation algorithms from H&E to IHC. This dataset provides high-quality paired data for deep learning-driven pathological image synthesis, and has been applied to models such as Pyramid Pix2pix, advancing non-commercial academic research and the development of auxiliary technologies for pathological diagnosis.

搜集汇总

背景与挑战

背景概述

BCI乳腺癌免疫组化图像数据集包含4872对H&E与IHC染色图像，涵盖HER2四种表达水平，支持乳腺癌免疫组化图像生成研究。数据集由北京朝阳医院与首都医科大学合作构建，适用于医学图像处理和计算机视觉研究。