Image segmentation with improved watershed algorithm using radial bases function neural networks

This paper proposes an improved watershed segmentation algorithm that uses RBF Neural Networks for the segmentation of image target objects. Instead of using catchment basin minima in order to define object regions, the technique developed throughout this work deploys RBF neural networks to predict the end boundaries of the segmentation clusters which are formed from the watersheds created in the image histogram topography. The RBF initial parameters, such as centers, and widths, are automatically set upon the histogram peaks and minima respectively. Experimental results of this leaning algorithm make it viable for different applications of gray scale image classifications.