Abstracto

Fuzzy Local Information C Means Clustering For Acute Myelogenous Leukemia Image Segmentation

Meera V, Shammy Arun Mathew

Leukemia is a type of cancer that affects the blood cells and most commonly WBCs or leukocytes. There are two main types of acute leukemia: Acute lymbhoblastic Leukemia (ALL) and Acute Myelogenous Leukemia (AML). In this paper AML is only considered. Here microscopic blood smear images containing multiple nuclei are exposed to a series of preprocessing steps which includes color correlation and contrast enhancement. By performing FLICM clustering algorithm on the resultant images, the nuclei of the cells invested with cancer are obtained. The main objective is to demonstrate that the classification of peripheral smear images containing multiple nuclei can be fully automated and to validate the segmented images. The method has been evaluated using a set of 50 images (with 25 abnormal samples and 25 normal samples). The system robustly segments and classifies AML based on complete microscopic blood images. SVM is employed for classifying the nucleus images based on the extracted features in to healthy and leukemic. The developed system can be used as ancillary/backup service to the physician.