Classification of White Blood Cells Based on Morphological Features

The extraction of nucleus from the blood smear images of white blood cells (WBC) provides the valuable information to doctors for identification of different kinds of diseases as most of the diseases present in body can be identified by analyzing blood. Manually it very soporific and tiresome to segment the nucleus and after that classification is done on the basis of that besides that the instruments which are used by experts for segmentation and classification of white blood cells are not affordable by every hospitals and clinics, so automatic system is preferable which reduces the times of segmentation and classification. In our research we focus on segmentation of nucleus from blood smear images using Otsu's thresholding technique applied after contrast stretching and histogram equalization of image followed by minimum filter for reducing noise and increasing brightness of nucleus, mathematical morphological is done to remove the components which are not WBCs, then shape based features are extracted on the basis of that classification rule is applied to classify them in their five category. The classification of nucleus is necessary as they are used to identify different kind of diseases which are related to each type of white blood cells and also help in differential blood count of cells.