Detection of Acute Leukemia Cells in Color Images with RGB and HSI Format Using Image Processing Method

The word leukemia which is means white blood. Leukemia or blood cancer is a progressive disease and malignant hematopoietic organs of the body. The diseaseis caused by incomplete proliferation and development of white blood cells and their precursors in the blood and bone marrow. In patients with leukemia, the bone marrow produces abnormal white blood cells. At first, leukemia cells performance almost normal. After a while, the accumulation of leukemia cells instead of whitecells, red cells and platelets, cell performance is difficult [4]. There are four different models of leukemia that in patients with excessive leukemic, these cellsmultiply rapidly and develop [7]. The types of leukemia are grouped by type of white blood cell involved. The origin of leukemia can be lymphoid cells or myeloid [4]. Four models of leukemia include: - Acute LymphoblasticLeukemia (ALL); - Acute Myelogenous Leukemia (AML); - Chronic Lymphoblastic Leukemia (CLL); - Chronic Myelogenous Leukemia (CML); This disease is one of the four most common cancers among children. Detect andclassify cancer cells in leukemia images can be very helpful for rapid diagnosis andtreatment of leukemia. The early diagnosis of leukemia in these patients can also increase their chances of recovery. Since conventional diagnostic methods are verytime consuming and contains errors, so hematologists are looking for a way to detect leukemia with lower error rate and lower time of diagnosis of this disease.The conventional method of manually counting cells using a microscope performed the operation, in addition to being time-consuming, as the technicianinto a lot of stress. Also, each model of this disease has unique features that only trained professionals can see them using microscopic images of cells infected.However, it is usually due to a variety of feature detection and the uncertainty ofthe images is difficult. One of the existing diagnostic methods are the safety phenotype and cytogenetic abnormality. These problems can be time consumingand costly procedure. Therefore, a rapid and cost-effective method is needed to identify the different types of leukemia [7]. Recently, many methods have been proposed that all of them diagnostic by image processing. In these methods, microscopic images slides are processed of bone marrow and blood. Most of these methods are generally classified based on microscopic images which by processingorders, leukemic cells recognize. Detection of leukemic cells can be useful in diagnosing leukemia, because the detection and classification of blasts can beobtained features that using these extracted features, hematologists can quickly identify the type of leukemia. Presented an algorithm for accurate detection of blastcells and extracting their features can quickly identify the type of leukemia and in other words that triggers faster treatment and hope for improvement in the patient increase. But all this is depended on a perfectly reasonable and verificationalgorithms, the error rate is very low and the detection time is less than conventional diagnostics. The identification of blood cells in acute leukemia basedon color image is one of the most challenging about research areas in image processing. As a solution to this problem, are researching on image processingmethods for segmentation of color images with acute leukemia using different color spaces and evaluation of color spaces using an algorithm for counting thenumber of blasts with their sizes that excites in blood cancer images. Image processing is a part of computer science that used from this in many areas, such as space science, Speech recognition, handwriting recognition, document classification, Industry, optical character recognition, search engines and in the diagnosis of diseases such as cancer. Image segmentation refers to the process ofpartitioning a digital image into multiple pixels that in most cases, from segmentation in separate division or removing specific parts of an image wereused. Research in the image segmentation field, was leading to the creation of many techniques such as classification based on histogram, based on edge, basedon region, clustering and a combination of these techniques. The processing techniques in medicine is quickly diagnose of diseases, especially diseases such ascancer that early detection of this is very important that they are very useful and efficient (The main result is a faster onset of treatment). In this paper, the color segmentation process on acute leukemia images in HSI and RGB images formatperformed, and a segmentation algorithm for counting the number of blasts and extract features from images with cancer blood were offered. The proposed method helps to improve the image resolution. Generally, there are two conditions for achieving improved images with color format:The first condition the color maintain the structure of the original image, that this item can be maintained simply by the ratio between the R, G, B components of each pixel in images with RGB format. - The second condition is to provide more information than the original image; The case can be obtain by using information contained in the components of the intensity and also components of the color in images [6].