An Investigational Study of Detecting Acute Lymphoblastic Leukemia using Computer Vision

Abstract

Among all divesting cancers, Hematologists predict that the Leukemia is mostly occur on the children, teenagers, and young adults. Moreover 85% of cancer cases are detected younger than the age of 15. Due to a Genetic abnormality in the bone marrow, Acute lymphoblastic leukemia (ALL) is particularly susceptible to life-threatening infections. The laboratory method to detect the ALL is prolonged and slow process. Reviewing prior research on detection methods in ALL is the goal of this study. This review includes (a) the parameters necessary for the ALL detection (b) methods involved in the detection process (c) which algorithms gives the accurate prediction and (d) finally all the necessary context for the detection of the ALL-affected cells. The proposed system possesses the automated machine learning (Chabot) approach used to categorize the infected and healthy cells in the blood smears (microscopic images) in order to detect ALL. The Blood smears are converted into the CMYK color space and separated into clusters using K-means Algorithm. A cell from each cluster is picked and detected whether the cell is affected with ALL or not using the nuclei of the cell using supervised learning algorithm like SVM, XGBoost Classifier, Etc., The proposed system aids in enhancing the acute lymphoblastic leukemia detection system.