White Blood Cell Detection and Segmentation from Fluorescent Images with an Improved Algorithm using K-means Clustering and Morphological Operators

Cell detection is the most basic and essential step for the analysis of cells. There are enormous types of blood disorders that can be identified by analyzing blood cells. There are several approaches used for this purpose. However, every method has their pros and cons. Improvement of segmentation of cells can increase the performance of cell classification and cell counting in later stages. The main concern of our paper is to segment the white blood cells from fluorescent images using K-means Clustering and Morphological Operators. We detect the cluster with WBC and refine the result depending on the presence of nucleus in the segmented cells. Non-WBCs are the cells without a nucleus and smaller in size. Presence of nucleus in a cell can be an indicator of WBC. We segment nucleus in cells and calculate the average area of the nucleus. We then refine the segmentation result based on the presence and size of the nucleus. Our analysis on WBCs demonstrates a comparison with ground truth values. We achieved our result with sensitivity of 96.4932% and precision of 9S.3584%. Our algorithm and analysis of results outperform state of the art method in several aspects.