A new framework in solving tailing and necking problems of thinned binary image

Abstract

A framework for solving tailing and necking problem in thinned binary image (TBI) is proposed. Tailing and necking are some of the classical problems occurred in thinned binary image. Artificial Neural Network (ANN) approach has been selected to be implemented in this study for obtaining a better thinned binary image. The identified TBI with tailing and necking problem are represented in a n x n dimensions of matrix and will be undergo a training of different set of neural network models that have been develop by using multiple layer perceptron and back propagation algorithm with different numbers of hidden layers. The experimental works show promising results.