An efficient EGWO algorithm as feature selection for B-ALL diagnoses and its subtypes classification using peripheral blood smear images

Abstract

Acute lymphoblastic leukemia (ALL) is the most serious type of leukemia that develops because it causes an abnormal increase in the production of immature white blood cells in the bone marrow. ALL spreads rapidly in children's bodies and leads to their death. The main objective of this paper is to introduce an enhanced methodology based on the k-means clustering algorithm for classifying all subtypes of ALL. Image preprocessing is the first step of the proposed methodology. For obtaining the images' descriptive features, feature extraction is the second step that has been used. To select the most vital features that can characterize the histology of the blood cells, Enhanced Grey Wolf Optimization (EGWO) algorithm has been used in the third step. Hence, EGWO will start to select the search agents based on certain criteria as the best cluster center by using the k-means clustering algorithm. Several supervised classifiers as Random Forest (RF), K Nearest Neighbors (KNN), Support Vector Machine (SVM), and Naïve Bayes (NB) have been compared. The proposed methodology achieves a high degree of accuracy of 99.22%, precision of 99%, and sensitivity of 99%. A comparative study has been established in order to verify the effectiveness of the proposed methodology.