Segmentation White Blood Cells by Machine Learning Algorithms

Abstract

Blood and its elements have a vital position in human life and are the best indicator for deciding many pathological states. Specifically, white blood cells are of great significance for diagnosing hematological disorders. In this analysis, 350 microscopic blood smudge images have experimented with six machine learning algorithms for the sort of white blood cells, and their renditions have resembled. Thirty-five distinct geometric and statistical (consistency) features have been pulled from blood pictures for practicum and test parameters of machine learning algorithms. According to the outcomes, the Multinomial Logistic Regression (MLR) algorithm accomplished better than the other techniques, with an average of 95% test victory. The MLR can be utilized for the automatic classification of white blood cells. It can be used mainly as a source for diagnosing diseases for hematologists and internal medicine experts.