Detection of Chronic Myelogenous Leukemia Applying Support Vector Machine and Feature Extraction

This paper employs image processing and artificial intelligence to detect Chronic Myelogenous Leukemia (CML) in peripheral blood smear images. The study's objective is to create a prototype that uses Otsu's Thresholding and SVM to create a device that can detect CML. Otsu's thresholding and Support Vector Machine Classification are the algorithms used for image processing and artificial intelligence. Roundness, Convexity, Solidity, and Eccentricity were the features that were extracted from the peripheral blood smear image. The results obtained show that 75 out of 78 samples were correctly classified. The system's overall accuracy via the confusion matrix is 96%. Hence, it can be concluded that Otsu's Thresholding, Feature Extraction, and SVM Classifier are algorithms that can detect CML.