glomeruli segmentation

The description of the proposed glomerulus segmentation dataset is as follows: The dataset contains 200 WSI (H&E, PAS, MAS, PASM) images provided by Peking University Shenzhen Hospital and manually labeled by pathologists. The slides of the dataset are basically similar in tissue structure but not pix-level paired. The magnification of the slides is 40×, and the resolution is 0.2528μm/pixel.This dataset can be used as a segmentation dataset to evaluate the glomerular segmentation performance of different stained kidney tissues and can be used as a virtual staining dataset to evaluate the stain transfer performance of different methods. This dataset will encourage research on artificial intelligence (AI) methods, create and compare fresh algorithms, and measure their usability in quantitative nephropathology.

所提出的肾小球分割数据集的描述如下：本数据集包含由北京大学深圳医院提供的200张WSI（H&E、PAS、MAS、PASM）图像，并由病理学家进行人工标注。数据集中的切片在组织结构上基本相似，但在像素级别上并非成对。切片的放大倍数为40×，分辨率为0.2528μm/像素。本数据集可作为分割数据集，用于评估不同染色肾脏组织的肾小球分割性能，同时也可作为虚拟染色数据集，用于评估不同方法的染色迁移性能。本数据集将激励人工智能（AI）方法的研究，促进新算法的创建与比较，并测量其在定量肾病理学中的实用性。

搜集汇总

背景与挑战

背景概述

该数据集包含200张不同染色的肾组织WSI图像，由专业病理学家标注，适用于肾小球分割和虚拟染色方法的研究与评估。