Gray level co-occurrence matrix and random forest based acute lymphoblastic leukemia detection

In this paper, we have proposed an acute lymphoblastic leukemia detection strategy from the microscopic images. The scheme utilizes all the steps associated with any other classification scheme, but our contribution lies on a marker-based segmentation(MBS), gray level co-occurrence matrix (GLCM) based feature extraction, and probabilistic principal component analysis(PPCA) based feature reduction. The relevant features are used in a random forest (RF) based classifier. Extensive experiments are carried out on the ALL-IDB1 dataset, and comparative analysis has been made with other existing schemes with respect to sensitivity, specificity, and classification accuracy. The proposed scheme (MBS+GLCM+PPCA+RF) achieves 96.29% segmentation accuracy and classification accuracy of 99.004% and 96% for nucleus and cytoplasm respectively.