GLRLM-based feature extraction for Acute Lymphoblastic Leukemia (ALL) detection

Abstract

This work presents gray-level run length (GLRL) matrix as feature extraction technique for the classification of Acute Lymphoblastic Leukemia (ALL). ALL detection in an early stage is helpful in avoiding fatal hematopoietic ailment which might cause death. The GLRL matrix extracts textural features from the nucleus of the lymphocyte image which is used with Support Vector Machine (SVM) for classification. The public dataset ALL-IDB1 is used for the experiment, and we have obtained 96.97% accuracy for GLRL feature with SVM classifier.