IEEE - ICSCN 2007, MIT Campus, Anna University, Chennai, India. Feb. 22-24, 2007. pp.521-522. Computer Aided Medical Diagnosis for the Identifi'cation of Malaria Parasites S.F. Tohal and U.K. Ngah2 Abstract: This paper presents one of the applications of digital image processing in artificial intelligence particularly in the field of medical diagnosis system. Currently in Malaysia the traditional method for the identification of Malaria parasites requires a trained technologist to manually examine and detect the number of the parasites subsequently by reading the slides. This is a very time consuming process, causes operator fatigue and is prone to human errors and inconsistency. An automated system is therefore needed to complete as much work as possible for the identification of Malaria parasites. The integration both soft computing tools has been successfully designed with the capability to improve the quality of the image, analyze and classify the image as well as calculating the number of Malaria parasites. I. INTRODUCTION Nowadays, as the computational power increases, the role of automatic visual inspection becomes more important. There are four identified species of this parasite causing human Malaria, where there are normally three main parasites usually found in Malaysia namely, P. Falciparum, P. Vivax, and P. Malariae [1]. It is a disease that can be treated in just 48 hours, yet it can cause fatal complications if the diagnosis and treatment are delayed. Image processing and artificial intelligence techniques are introduced that may provide a valuable tool for improving the manual screening of specimens. There are two types of blood available for the detection of Malaria parasites [1] that are thin blood smear, to find the species of Malaria parasites while thick blood smear is done to find the density of Malaria parasites per red blood cells (used in this study). There have been a number of relevant approaches in the literature using computational intelligence and microscopic images. Toha and Ngah [2] presented the use of fuzzy logic and two soft computing tools to identify the type and development stage of Malaria parasites using thin blood smear image. Ngah and Ho [3] developed a pi fuzzification algorithm to analyse the intensity value of micro calcification. They use a fuzzy c-mean clustering in order to classify the mammography micro calcification image into few image clusters Department of Mechatronics, Faculty of Engineering, International Islamic University Malaysia, Malaysia 2 School of Electrical and Electronic Engineering, Universiti Sains Malaysia, Malaysia Email: tsfauziah iiu.edu.my, umik 4eng.usm.my II. RESEARCH APPROACH The Thick Blood Smear Image analysis is focusing on how to count the number of Malaria parasites exist in each digitized red blood specimen image. The image analysis can be done in both manual and automatic. Figure 1 shows the overall process of the system. l ~~~~~~~Imag)e Clustering Thick Segmentation And Blood Tresholding Euclidean Distance ..........Number of ....................................................................~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........................................... 11y111 +,iiiiiiiiiiiiiiillllllllllllllllllla ll1||1ilG~~............................................. m iii ii t................................................................... Fig. 1. Program Use-Case Diagram A. Histogram The information about an image is known directly from the result of image histogram. In order to calculate histogram of an image, an array with the same size as total of grey level is developed with a starting value of 0. The image is then scanned pixel by pixel and the pixel's grey level is used as the array index. Refer to Equation (1) for histogram hi. h(i = n(i) (1) n n(i)= Summation of the number of pixels within the same grayscale value, i. n Summation of all the number of pixels in an image. i = 0 to N which represent image intensity. B. Image segmentation - threshold Thresholding is a non-linear operation that converts a gray- scale image into a binary image where the two levels are assigned to pixels that are below or above the specified threshold value [4]. The threshold value within 1 to 255 gained is used as an input for image segmentation. All the intensity value which lies below the threshold value will be set as 0 while for the intensity value above the input value will be set as 255. This will result to a segmented image where black color will represent the background of the digitized red blood specimens while the white color will characterize the Malaria parasites existed in the image. Equation (2) is used to calculate threshold value: 1-4244-0997-7/07/$25.00 C2007 IEEE 521 Authorized licensed use limited to: Sheffield University. Downloaded on July 15, 2009 at 09:46 from IEEE Xplore. Restrictions apply.