Journals Proceedings

Abstract

The aim of this paper is to automate the process of detection of leukocytes using image processing techniques such as automatic thresholding and mathematical morphing. We have used the segmentation to detect white blood nucleus. Now a day’s the automatic system is preferred as manual segmentation is very tedious, tardy and sometime prone to error, besides that the medical instruments which are used to detect white blood cells are very costly and may not be exist in all the hospitals and clinics. In automatic process, localization and segmentation of white blood cell are the most important stages. In this research we focus on white blood cell nucleus segmentation in which we firstly applied the mathematical techniques to identify the dark objects in an image like white blood cell and platelets then automatic segmentation on blood smears using the global Ostu thresholding technique which convert the segmented image to binary image and then mathematical morphing is applied on the binary image which removed the components which are not white blood cells and then the removal of border objects which are incomplete white blood cells results in final segmentation. The result obtained in this research shows a good accuracy rate as compared to other techniques for automatic segmentation.