CBC (Complete Blood Count)

全血细胞计数 (CBC) 数据集包含 360 张血涂片图像及其注释文件，分为训练集、测试集和验证集。训练文件夹包含 300 张带注释的图像。测试和验证文件夹都包含 60 张带注释的图像。我们对原始数据集进行了一些修改以准备此 CBC 数据集，其中一些图像注释文件包含的红细胞 (RBC) 比实际值非常低，并且一个注释文件根本不包含任何 RBC，尽管细胞涂片图像包含 RBC .因此，我们清除所有错误文件并将数据集分成三个部分。在 360 张涂片图像中，先将 300 张带注释的血细胞图像作为训练集，然后将剩下的 60 张带注释的图像作为测试集。由于数据不足，训练集的一个子集用于准备验证集，其中包含 60 张带注释的图像。

The Complete Blood Count (CBC) dataset consists of 360 blood smear images and their corresponding annotation files, which are split into training, test, and validation subsets. The training folder contains 300 annotated images, while both the test and validation folders each hold 60 annotated images. We made revisions to the original dataset to prepare this CBC dataset: some of the original annotation files contained significantly fewer red blood cells (RBCs) than the actual count, and one annotation file had no RBCs at all even though the matched blood smear image did contain RBCs. Therefore, we eliminated all erroneous files and split the revised dataset into three parts. Out of the 360 total smear images, 300 annotated blood cell images were first designated as the training set, and the remaining 60 annotated images were used as the test set. Due to insufficient available data, a subset of the training set was employed to formulate the validation set, which contains 60 annotated images.

搜集汇总

背景与挑战

背景概述

CBC（全血细胞计数）数据集包含360张血涂片图像及注释，经过修正错误后划分为训练、测试和验证集。其中训练集有300张图像，测试集和验证集各含60张图像。