Classification of Acute Myeloid Leukaemia using Deep Learning Features

Abstract

Acute Myeloid Leukaemia (AML) is a sub-category of leukaemia, which is prevailing among adults. One of the top ten most dangerous causes of mortality for humans is AML, which stands for Acute Myeloid Leukaemia. AML originates in the bone marrow but tends to rapidly spread to the bloodstream. In some cases, it can also affect other parts of the body such as lymph nodes, liver, spleen, central nervous system (which includes the brain and spinal cord), and testicles. Manual methods are time-consuming, and their accuracy depends on the operator’s ability. However, automatic classification can improve accuracy and speed up clinical decisions for haematologists and medical experts. However, state-of-the-art mechanisms are complex and their accuracy is less than 98 percent. The objective of our research was to introduce a novel approach that utilizes a Convolutional Neural Network (CNN) for extracting CNN features from images of blood smears. Moreover, it classified the images into two classes i.e., normal and abnormal. The proposed mechanism achieves high accuracy of 99.07%.