Turning Healthcare Data into Innovative Solutions. Raabin brings together medical expertise, real-world health data, researchers, engineers, and technology teams to create new opportunities for research, artificial intelligence, software, medical devices, and healthcare innovation. Our mission is to build an open and collaborative ecosystem where valuable health data can become the foundation for meaningful solutions.
The Raabin-WBC dataset provides approximately 40,000 microscopic images of white blood cells and blood-cell artifacts, along with cell-type labels, localization information, and ground-truth segmentation data for selected cells. To make access easier, the dataset is organized into separate downloadable files and subsets. You can download only the data you need for classification, cell detection and localization, or nucleus and cytoplasm segmentation, without downloading the entire collection.
Please consult the original publication for dataset structure, labeling details, and recommended use. When using Raabin-WBC, please cite the original dataset publication.
Need only specific data? Browse the six dataset components below and download individual files or subsets without downloading the entire collection.
Cropped white blood cell images with expert annotations for training and evaluating cell classification models. Includes training and test subsets.
Microscopy data collected using the first microscope, provided in separate downloadable archives for more flexible access.
Microscopy data collected using the second microscope, organized into individual archives for targeted download and analysis.
Example files demonstrating how to extract cell images and associated information from JSON data generated using the first microscope.
Example files demonstrating cell extraction and access to associated information from JSON data generated using the second microscope.
Ground-truth data and related resources for white blood cell nucleus and cytoplasm segmentation, supporting the development and evaluation of image segmentation algorithms.
If you utilized the ground truth data, in addition to the dataset paper, you should cite the following paper: Sajad Tavakoli et al. New segmentation and feature extraction algorithm for classification of white blood cells in peripheral smear images
Currently, there is a possibility for presenting more than 10,000 images of 4 types of leukemia.
Download HereRaabin Data www.raabindata.com

Associate Professor, Machine Learning and Robotics Group, University of Tehran.
“Raabin-WBC is a valuable contribution to medical image analysis and AI research, providing researchers with accessible and high-quality blood cell image data. More importantly, Raabin is now moving beyond a dataset toward building an ecosystem connecting health data, artificial intelligence, research, and practical diagnostic solutions. I believe this direction has significant potential to support innovative and locally relevant healthcare technologies.”