Medical Image Segmentation Using the Kohonen Neural Network

Abstract

An important definition of image segmentation is to divide an image into its features. This is a very important area in image processing field. The nontrivial image segmentation task is one of the most difficult challenges, since and there isn’t a unique suitable method to target all kinds of images so far. There are several researches related to image segmentation, some of them is the application of Self-Organizing Maps - SOM. In order to explore this method, that differs of the most used, such as the spatial filtering using derivative filters as detectors of edge Sobel, Prewitt, Roberts and Canny, thresholding techniques and Otsu method, this work aimed the medical image segmentation using the Kohonen neural network so itself adopts a non-supervised learning in the training process.