Description

Since the MR technique is becoming more popular due to its non-invasive principle, the imaging of biological structures by MR equipment’s is a routine investigating procedure today. For this reason, the automatic processing of this kind of images is getting more attention. Nowadays, the issue of automatic analysis of brain tumors is of great interest. It is the first step in surgical and therapy planning. And the very first step of the automatic analysis of brain tumor is its detection and subsequent segmentation. The difficulty of the tumor segmentation is in its shape variability in each case. The automatic segmentation of brain tumors is still a challenging problem, even though several different and interesting fully- or semi-automatic algorithms have been proposed in recent years. The existing algorithms can be classified into semi- and fully-automatic methods from a user viewpoint and into region- and contour-based methods from technical view-point. Existing system encoded knowledge of the pixel intensity and spatial relationships in the images to create a fully automated segmentation system know n as the KG (know ledge- guided) method. In this project the system proposed a novel semi-automatic segmentation method based on population and individual statistical information to segment brain tumors in magnetic resonance (MR) images. The probability of each pixel belonging to the foreground (tumor) and the back ground is estimated by the k NN classifier under the learned optimal distance metrics. A new cost function for segmentation is constructed through these probabilities and is optimized using graph cuts.