Automated morphometric classification of acute lymphoblastic leukaemia in blood microscopic images using an ensemble of classifiers

Subrajeet Mohapatra,D. Patra,S. Satpathy,R. Jena,Sudha Sethy

DOI: https://doi.org/10.1080/21681163.2014.897650

2016-01-02

Abstract:Leukaemias are neoplastic proliferations of haemopoietic cells which affect both children and adults and remain one of the leading causes of death around the world. Early diagnosis and classification of such malignant disorders are necessary, and it has always been a challenge in the field of haematology and laboratory medicine. Accurate and authentic diagnosis of leukaemia is essential for the confirmation of the disease, prognostic classification and effective treatment planning. Visual microscopic examination of blood slides is considered as an indispensable diagnostic technique for the screening and classification of leukaemia across continents. However, manual investigation is often slow and limited by subjective assimilation and reduced diagnostic precision. In this study, we investigate the use of image morphometry and pattern recognition techniques for subtyping leukaemic lymphoblasts as per French–American–British classification. Reliable classification results were obtained using the robust segmentation methodology, prominent morphological features and an ensemble of classifiers. To evaluate the performance of the proposed methodology, a comparative study is realised over the available image data-set. The classification rates achieved with the standard classifiers, that is naive Bayesian, K-nearest neighbor, multilayer perceptron, probabilistic neural network and support vector machines, were compared with that obtained using an ensemble of classifiers. It is observed that the classification rate is improved with the use of multiple classifier ensemble and is expected to assist clinicians in making the diagnostic process faster and more accurate.