Brain tumor is a tissue mass formed due to the uncontrolled growth of brain cells [1]. The presence of tumors at an early stage is very difficult to diagnose because the tumor boundary is generally still unclear, having low contrast, and sometimes similar to normal tissue. CT Scan (Computed Tomography Scanning) is a diagnostic supporting device utilized to examine organs such as the central nervous system, musles, bones, throat and abdominal cavity [2, 3]. Examination of brain tumor is mostly done by using a CT Scan [4].

Reading of brain CT Scan images was visually performed by doctors by observing the image [5]. While the determination of the location or area of object on CT Scan image that is affected by brain tumors cannot be done by means of human’s view or judgment, human valuation is not always right, then it is very worrying that the results of the diagnosis are inaccurate [6]. Hence, we need a technique that can evaluate, recognize, and detect the location or area of brain attacked by a tumor. Picture 1 (a) and (b) show normal image and image of brain tumor [7].

(a)                             (b) Figure 1. Brain Image (a) Normal (b) Brain Tumor [7]

Support Vecktor Machine (SVM) is a pattern recognition method aimed to find the best hyperplane [8]. The SVM method can be used to analyze brain tumors by determining the mean anda standard deviation as parameters in normal class and brain tumors class distinguishing.

Materials used in this study are 7 normal brain images and 7 brain tumor images from medical CT Scan. The process of implementing the mean and standard deviation methods with the SVM classification is described as follows:

The first step in analyzing brain image is pre-processing to eliminate noise following by histogram analysis. The histogram analysis had been performed in order to find out the distribution of pixel intensity values. The normal (a) and brain tumor (b) outcomed by images pre-processing and histogram analysis are shown in Figure 2.

Figure 2. Pre-processing and histogram analysis of brain image (a) Normal (b) Tumor

Subsequently, 7 normal images and 7 images of brain tumors were determined by the mean and standard deviation values sought in the program in each image. Mean and standard deviation values of normal (N) and brain tumor (T) images are represented in the following Table 1.

Table   1.   Mean   values   and standard deviations

Data in Table 1 show that the normal images have mean values with the range from 65.45 until 88.50 and standard deviation values with the range from 44.41 until 49.50, where as the brain tumor images have mean values with range from 69.46 until 93.29 and standard deviation values with range from 50.05 until 61.60.

The normal images have mean values with the range from 65.45 until 88.50 and standard deviation values with the range from 44.41 until 49.50, where as the brain tumor images have mean values with range from 69.46 until 93.29 and standard deviation values with range from 50.05 until 61.60. The mean and standar deviation of brain tumor images is greater than the normal images. Thus, a normal brain can be distinguished from brain tumor by using the mean and standar deviation values of its image.