Segmentation of Bone Marrow Cell Images for Morphological Classification of Acute Leukemia

Abstract

This paper presents a two-phase methodology to analyze the morphology of abnormal leukocytes images for the classification of acute leukemia subtypes using image processing and data mining techniques. In the first phase we propose a segmentation algorithm that uses color and texture information in order to extract leukocytes and their respective nucleus and cytoplasm from bone marrow images with heterogeneous staining. As usual, these images show a high-cell population, we suggest using conical shapes to separate overlapped blood elements. In the second phase we perform feature extraction to the regions segmented and use these attributes to classify the cells into leukemia subtypes. In our experiments we achieved an average accuracy of 95% in the evaluation of the segmentation process. An overall accuracy of 92% was reached in the supervised classification of acute leukemia types, 84% in lymphoblastic subtypes, and 92% in myeloblastic subtypes.