Using Neural Networks and Hough Transform for Leukocytes Differentiation in Blood Count Images

Abstract

Automated equipments for blood count have facilitated the clinical routines in medical laboratories. However, when grave anomalies are identified, some samples still need to be analyzed in a traditional way using a microscope. This paper proposes a method to classify the five most common types of leukocytes, present in the human blood using blood smear images. This allows reducing the amount of work, the time spent in the blood test as well as cost of such a test. The results are satisfactory and should contribute positively to the development of an automatic system and/or a smart electronic device, which would come handy for this kind of applications.