A COMPARISON OF HOPFIELD NEURAL-NETWORK AND BOLTZMANN MACHINE IN SEGMENTING MR-IMAGES OF THE BRAIN

In this paper we present contributions to improve a previously publish ed approach for the segmentation of magnetic resonance images of the h uman brain, based on an unsupervised Hopfield neural network, We formu late the segmentation problem as the minimization of an energy functio n constructed with two terms: the cost-term as a sum of squared errors and the second term temporary noise added to the cost-term as an exci tation to the network to escape certain local minima, with the result of being closer to the global minimum, Also, to ensure the convergence of the network and its utilization in the clinic with useful results, the minimization is achieved with a step function that permits the ne twork to reach stability corresponding to a local minimum close to the global minimum in a prespecified period of time. We present segmentat ion results of our approach for data of patient diagnosed with a metas tatic tumor in the brain, and we compare them to those obtained from p revious work using Hopfield neural networks, the Boltzmann machine, an d the conventional ISODATA clustering technique.