erythroSight: Open-access image dataset of blood cells from individuals with sickle cell disease and/or beta-thalassemia or without known hemaglobinopathies, captured using automated microscope Octopi

This dataset contains over 300,000 microscopic images of blood cells from 138 participants in Nepal and Canada, captured using automated microscope Octopi. The images (0.9 mm by 0.9 mm) were captured for research studies in Nepalgunj, Nepal and Vancouver, Canada to evaluate low-cost tools for detecting sickle cell disease and beta-thalassemia. The dataset includes raw image pairs (left and right half illuminated images), processed images (differential phase contrast images), cell segmentation outlines, and morphological characterization of cells. The majority of images are captured 2 hours after sample preparation of the sickling test at room temperature, while other configurations (such as other time instances, different temperature settings, time series data) are also included.

本数据集涵盖来自尼泊尔与加拿大共138名受试者的逾30万张血细胞显微图像，所有图像均通过自动化显微镜Octopi采集。单张图像幅面为0.9毫米×0.9毫米，采集自尼泊尔内帕尔根杰（Nepalgunj）与加拿大温哥华的相关研究项目，旨在评估用于检测镰状细胞病（sickle cell disease）与β地中海贫血（beta-thalassemia）的低成本检测工具。本数据集包含原始图像对（左右半侧照明图像）、处理后图像（微分相差显微图像）、细胞分割轮廓以及细胞形态学表征数据。绝大多数图像采集自室温下镰状化试验样本制备完成2小时后，数据集同时涵盖其他配置场景的数据，例如不同时间节点、不同温度设置的样本以及时序数据集。

搜集汇总

背景与挑战

背景概述

该数据集是一个开放获取的血液细胞图像集合，包含超过30万张显微镜图像，来自尼泊尔和加拿大的138名参与者，重点关注镰状细胞病和β-地中海贫血患者，以及无已知血红蛋白病的个体。图像使用自动化显微镜Octopi捕获，包括原始图像、处理后的图像、细胞分割轮廓和形态特征，主要用于研究低成本检测工具，总大小达15.64 TB，发布于2024年。