Automated Leukemia Detection System

Abstract

The analysis and classification of White Blood Cells for the Detection of Acute Lymphocytic Leukemia (ALL), a blood cancer which can be fatal if left untreated. Currently, the morphological analysis of blood cells is basically performed manually by professional operators which have various drawbacks such as slow analysis, non-standard accuracy and operator’s dependencies. The idea of automated leukemia detection system is based on the image processing and artificial intelligence which provides automated results of the microscopic images from the dataset. This approach identifies the WBCs then separates them from grouped WBCs and then cleans the background. After this, each individual cell component is identified in detail. Then different shape and color features are extracted from these individual cells. Extracted features are used to train the Support Vector Machine classification model with Linear kernel in order to get the more accurate results. Using our method, 93% of the cells were properly identified and sensitivity of 90% is achieved. Therefore, this system helps to analyze the suspected blood samples automatically hence reducing the chances of human errors and provides the simpler, robust and convenient method for the results of the suspected blood samples. The method proposed permits the analysis of the cells and it represents the Automated leukemia detection system as the Medical tool for ALL Diagnosis.