The blood sample images consist of Red Blood Cells, Platelets, White Blood Cells, and Plasma. White Blood Cells are classified into different types and their sub type based on its features. The white blood cells are further classified as granulocytes, Monocytes and lymphocytes. Leukaemia is a type of blood cancer which causes severe damage to human body if it is left untreated. The counting and classification of blood cells allows in evaluating and diagnosing of vast number of diseases. These diseases can be detected and treated at early stage Analysing of white blood cells (WBCs) allows for detecting acute lymphoblastic leukaemia (ALL). The morphological analysis of blood cells is performed manually by skilled operators. However, this method has many drawbacks such as slow analysis, non-standard accuracy, and dependencies depend on the skills of the operator. This paper presents an automated method for identifying WBC and classifies those using microscopic images. The other approaches identify the leucocytes first and then the other components but the proposed system isolate the whole leucocytes and then separates the cytoplasm and nucleus. This approach is necessary to analyse each cell component and extract its features. From each cell component, various features like shape, colour and texture, are extracted using a new approach for background pixel removal. The feature set obtained was used to train different classification models to determine which one is most suitable for the detection of acute and chronic leukaemia.
Cite this article:
K. Dhana Shree, B. Janani. Classification of Leucocytes for Leukaemia Detection. Research J. Engineering and Tech. 2019;10(2):59-66. doi: 10.5958/2321-581X.2019.00011.4