Automatic Counting of Infected White Blood Cells Using Multi-Level Thresholding

Abstract

The identification of leukocytes is the primary step to analyze diseases such as leukemia which are ordinarily performed by pathologists utilizing a light binocular. These are very time consumption, monotonous, expensive and require specialists. In this way, computer assisted symptomatic framework that helps pathologists within the demonstrative strategy can be exceptionally viable. White blood cell (WBC) fragmentation is ordinarily the primary stage in computer-assisted diagnostic framework development. This paper proposed an automated method for segmenting and counting white blood cell from microscopic images. For this purpose, Otsu technique was used in segmentation of WBCs from a microscopic image, Counting WBCs and diagnoses the patient case. The proposed system achieves an accuracy of 98.2%