Image Annotation using Moments and Multilayer Neural Networks
Mustapha Oujaoura, Brahim Minaoui, Mohamed Fakir · 2012
This document presents a system in order to annotate image content by using the region growing segmentation, as a method to separate different objects within an image, and the multilayer neural network to classify these objects and to find the appropriate keywords for them. In many applications, different kinds of moments have been used as features to classify the images and objects’ shapes. The Hu moments, Legendre moments and Zernike moments are used, in this paper, as features to describe an image. The experiments are done through using ETH-80 database images. General Terms Neural Networks, Moments, Image Annotation.