Sistem saringan dan diagnosis automatik penyakit leukemia akut berasaskan sampel darah

Leukemia is a blood cancer disease that contributes to the increment of mortality rate in Malaysia each year. There are two main categories for leukemia which are acute and chronic. Acute leukemia can be divided into two types which are acute myelogenous leukemia (AML) and acute lymphocytic leukemia (ALL). The production and development of acute leukemia cells happen rapidly and uncontrollable. ALL is a common type of leukemia that occurs amongst children between the ages of 0 to 13 years old. If the identification technique for acute leukemia could be done fast and effectively in the early stage therefore proper treatment and medicine could be delivered. Nowadays, haematologists or technologists will screen for leukemia using microscopic blood image. Meanwhile, microscopic bone marrow images are used for diagnosing purposes. However, there are many techniques other than analyzing microscopic images that are not used especially in the third world countries due to high cost in both purchases and maintenance of the required equipments. Therefore, screening and diagnosing using computer aided system becomes more popular choice amongst medical image processing researchers. This research involved the development of an automated screening and diagnosis of acute leukemia based on blood samples. In this study microscopic blood images are used to diagnose leukemia as an alternative solution towards the use of microscopic bone marrow images. With this approach, it can indirectly reduce the sense of fear and stress of an acute leukemia patient during bone marrow biopsy needle process. The system developed is also equipped with an automatic capturing system for capturing images automatically with 10X and 40X lenses. Besides that, this study also suggested the combination of image processing with artificial neural networks in creating a new procedure for screening and diagnosing process of acute leukemia blood sample. For screening process, a new procedure for segmentation technique was proposed based on the combination of S-component of HSI colour model with threshold method, 5×5 median filter and SRG algorithm by using image capture with 10X magnification lens. Meanwhile, the combination of S-component of HSI colour model with clustering algorithm, 7×7 median filter and SRGAE algorithm has created a new procedure for segmentation technique by using image capture with 40X magnification lens. The study has also incorporated feature extraction where features such as size, colour and shape were chosen in classifying acute leukemia cells with the aid of artificial neural networks which can be done automatically. The proposed system is able to classify acute leukemia for both ALL and AML with percentage accuracy of up to 99%.