Automated Screening System for Acute Myelogenous Leukemia Detection using Layer Subtraction

Abstract

Leukemia is one type of blood cancer. It mainly attacks the blood, bone marrow etc. One of the subtype of acute leukemia is Acute myelogenous leukemia (AML). Leukemia is detected by the microscopic analysis of peripheral blood smear. But this microscopic assessment is time consuming and is governed by haematologists. To overcome this drawback automatic screening system for leukemia detection arises. For automatic detection of leukemia in peripheral blood samples, an efficient methodology such as pre-processing, nuclei segmentation, feature extraction, classification is proposed. The advantage of this technique is its simplicity, classification of complete blood smear images and helps to segment and detect nucleated cells. Fourty microscopic blood images were tested and the proposed method obtains 97.56 percentage accuracy for the localization of cells and to separate it from the complete images.