Brain Image Segmentation Research Paper

1535 Words7 Pages
Abstract—Brain image segmentation is challenging task for proper clinical diagnosis. In this paper we present automatic segmentation of the brain into four classes namely background, cerebro spinal fluid, grey matter and white matter. Accurate segmentation of the tumor in the brain is also achieved using the proposed method. Segmentation is achieved by calculating the distance of pixel with respect to the mean and passing only those pixels having a specified variance by assuming a Gaussian distribution for each class. I. INTRODUCTION Neurological Disorders have been a challenging area for Medical Science. This is because of high complexity of the anatomical structures of the brain. From the clinical point of view, because of different forms…show more content…
MRI scanning is relatively safe because of its non invasive nature and unlike other medical imaging modalities, can be used as often as necessary. MRI images have good contrast in comparison to computerized tomography (CT). Therefore, most of research in medical image segmentation uses MRI images [1]. MRI is an important imaging technique for detecting abnormal changes in different parts of the brain in early stage. MRI imaging is a popular way to obtain an image of brain with high contrast. MRI acquisition parameters can be adjusted to give different grey levels for different tissues and various types of neuropathology [1]. The identification of brain structures in magnetic resonance imaging (MRI) is very important in neuroscience and has many applications such as: mapping of functional activation onto brain anatomy, the study of brain development, and the analysis of neuroanatomical variability in normal brains [2]. Brain image segmentation is also useful in clinical diagnosis of neurodegenerative and psychiatric disorders, treatment evaluation, and surgical planning…show more content…
4a shows the MR brain image, taken from IBSR database. It is called Image 3. It is image of Tumor infected brain. The proposed method segments the tumor region in the image. The segmented tumor and ground truth images are shown as follows Fig. 4b and Fig. 4c respectively. Fig. 5a shows the MR brain image, taken from BRATS database. It is called Image 4. The segmentation results and ground truth images are shown as follows Fig. 5b to Fig. 5i. A. Results of Evaluation The segmentation done by the proposed method is evaluated using the following segmentation evaluation metrics 1. Jaccard Coefficient: It measures the overlap of two sets. The Jaccard index is zero if the two sets are disjoint, i.e., they have no common members, and is one if they are identical. 2. Dice Coefficient: It measures set agreement. A value of 0 indicates no overlap; a value of 1 indicates perfect agreement. 3. Ratio of False Positives (RFP) and 4. Ratio of False Negatives(RFN) Table 1 shows the evaluation metrics for Image 1. Table 2 shows the evaluation metrics for Image 2. Table 3 shows the evaluation metrics for Image 3. Table 4 shows the evaluation metrics for Image

More about Brain Image Segmentation Research Paper

Open Document