Automated Leukemia Disease Classification using Machine Learning on Microscopic Blood Images

Abstract

Leukemia is a pathology that affects teenagers and adults, which leads to several other symptoms and premature death. Computer-aided system (CAD) is used to assist specialists in the diagnosis of this disease and lower the risk of prescribing improper treatment. Microscopic analysis is an efficient strategy for performing the initial screening of patients with leukemia. This kind of test can be manually done, generating fatigue in operators. Consequently, an economical method that is robust and automatic is needed to prevent the influence of operators. Several CAD systems were designed by using computational intelligence and image processing methods. In this paper, we present an Automated Leukemia Disease Classification utilizing Machine Learning on Microscopic Blood Images (ALDC-MLMBI) method. The ALDC-MLMBI technique aims to employ ML approaches for the identification of leukemia. As a primary step, the ALDC-MLMBI technique follows median filtering (MF) based noise elimination and adaptive histogram equalization (AHE) based contrast enhancement. Besides, the segmentation process can be performed by watershed segmentation. Meanwhile, the ALDC-MLMBI technique involves a set of feature extractors namely local binary pattern (LBP), histogram of gradients (HOG), scale-invariant feature transform (SIFT), and gray level co-occurrence matrix (GLCM). Furthermore, the classification of leukemia can be made by the use random forest (RF) method. The simulation analysis of the ALDC-MLMBI system can be performed using the Kaggle image dataset. The experimental outcomes highlighted the superior performance of the ALDC-MLMBI system compared to existing classifiers.