Description:

This dataset contains segmented images of white blood cells, derived from the Peripheral Blood Cell Dataset available. The original dataset comprises 17,092 images of individual normal cells, captured using the CellaVision DM96 analyzer in the Core Laboratory at the Hospital Clinic of Barcelona. These images are categorized into eight classes: neutrophils, eosinophils, basophils, lymphocytes, monocytes, immature granulocytes, erythroblasts, and platelets.

Download Dataset

In this dataset, we provide 195 images per class, each accompanied by a corresponding binary mask. These masks were created using OpenCV techniques, including GrabCut and morphological operations. This dataset is ideal for training segmentation models such as U-Net, Mask R-CNN, and other advanced image segmentation techniques.

Applications

Training Segmentation Models: This dataset is ideal for training models focused on segmenting white blood cells, such as U-Net and Mask R-CNN.

Educational Purposes: A valuable resource for learning and practicing advanced image segmentation techniques.

Benchmarking: Use this dataset to compare the performance of different segmentation algorithms and models.

Methodology

The segmentation masks were created using the GrabCut technique, followed by morphological operations for refinement. This method ensures precise and high-quality segmentation, making the dataset suitable for a wide range of image segmentation tasks.

This dataset is sourced from Kaggle.