Color Image Segmentation Using Centroid Neural Network

Abstract

Color image segmentation has been attracting more and more attention, mainly because color images can provide more information than gray level images. Many methods have been proposed so far to deal with the problem. However, most methods require fine tuning of parameters, which can be attained after repetitive trial and error. This paper discusses unsupervised learning in terms of Centroid Neural Network (CNN). In fact, CNN is the crucial algorithm to diminish the empirical process of parameter adjustment required for color image segmentation. The simulation results indicate that the proposed technique yields the reasonably segmented images in comparison with other conventional algorithms.