ehottl/blood_dataset

Dataset Card for Blood Cell Dataset

Description

This dataset contains images and labels for blood cells across eight distinct classes:

• basophil
• eosinophil
• erythroblast
• plasma cell
• lymphocyte
• monocyte
• neutrophil
• platelet

This dataset is primarily based on the Falah/Blood_8_classes_Dataset. To enhance its diversity and robustness, it has been augmented with additional White Blood Cell (WBC) data from five external sources, which are detailed below:

• Blood Cell Detection (BCD) dataset (Aslan): Comprises 100 annotated PNG images. Of these, 2,237 cells are labeled as Red Blood Cells, and only 103 as White Blood Cells. Each image is 256x256 pixels with RGB channels. (For more details, visit https://github.com/draaslan/blood-cell-detection-dataset. Accessed: June 30, 2020)

• Complete Blood Count (CBC) dataset (Alam et al.): Includes 360 blood smear images along with their annotation files. (Find more information at [54] and https://github.com/MahmudulAlam/Complete-Blood-Cell-Count-Dataset. Accessed: June 20, 2021)

• White Blood Cells (WBC) dataset (Zheng et al.): Features 300 images sized 120x120 pixels, and an additional 100 color images sized 300x300 pixels. (More information is available at http://www.doi.org/10.17632/w7cvnmn4c5.1. Accessed: May 15, 2019)

• Kaggle Blood Cell Images (KBC) dataset (Mooney): Consists of 12,500 augmented JPEG images of blood cells, accompanied by CSV files containing cell type labels. There are approximately 3,000 images for each of four different cell types. (Explore more at https://www.kaggle.com/paultimothymooney/blood-cells. Accessed: May 15, 2019)

• Leukocyte Images for Segmentation and Classification (LISC) dataset (Rezatofighi et al.): A dataset of 250 blood smear images in BMP format, including a specific set of 25 basophil images. (Further details can be found at http://users.cecs.anu.edu.au/~hrezatofighi/Data/Leukocyte%20Data.htm. Accessed: May 15, 2019)

Usage

Originally developed by Falah G. Salieh, this dataset is designed for blood health classification in healthcare applications. It is suitable for various tasks in computer vision, image classification, and machine learning research.

Citation

I acknowledge and appreciate the foundational work of Falah G. Salieh. If you utilize this dataset in your research or work, please cite the original author:

@misc{blood_8_classes_dataset,
  title = { Blood_8 Classes_Dataset},
  author = {Falah G. Salieh},
  year = {2023},
  publisher = {HuggingFace Hub},
  url = {\url{https://huggingface.co/datasets/Falah/Blood_8_classes_Dataset/}}
}