Accurate red blood cells automatic counting in microscopic thin blood smear digital images

Abstract

The aim of this study is to automate the counting process of Red Blood Cells in thin Blood smear images in more accurate, efficient and universal way. The Red Blood Cells have important role in the blood; their counting is part of the complete blood count test and is frequently suggested by the Physician because the Red Blood Cells have hemoglobin, responsible to carry oxygen to various tissues of the body.. The number of Red Blood Cells both (Low and High) deviations from normal range is an important indicator about any disorder existence in the body. At present mostly the counting process is performed manually which is laborious, error prone and time consuming. The automated diagnosing gain the attention of the researchers from the last two decades because it assist the experts to reduce the burden of errors, labor and time of examination. In this regard, too much research has been performed on the automation of the counting process of the Red Blood Cells but still the test demands to be done in a proper, efficient, accurate and realistic way. The proposed method achieved an average True Positive Rate (TPR) of 94%, True Negative Rate (TNR) of 6%, average accuracy of 97% and average error rate of 3%.