Segmentation of Brain Tumors MRI Images using a Nonparametric Method

Parzen window technology was used to segment a set of magnetic resonance brain images, which are segmentation by the threshold of the image to identify tumors in the brain. This hypothesis indicates that the gray level contains two or more values and that there is a marginal value to separate them, so that the area where the gray level is below the minimum cut value is Background and the value of the area where the gray level is higher than the minimum cut value is objects, and vice versa. Finding the threshold limit is by finding the density function of gray image data by dividing the original image into levels that represent the level of the object (the image being determined) and the background of the image. Through these levels, the density function is calculated by using the function of Gaussian Epanechnikov, and other functions in order to determine the threshold limit on which the image is divided.