Modelos neuronales pulsantes adaptados para el mejoramiento de luminosidad de imágenes cerebrales de gran resolución

Abstract

In this paper it is propose, implement and experiment with two Pulsed Neural Networks based on the Intersection Cortical Model for the enhancement of human brain medical images luminosity . Digital images are widely used in the medicine area, but these can be degraded by several factors. The degradation in its luminosity generates a problem for its correct analysis, since they have a short dynamic range and low contrast. The actual necessity to obtain good quality images joined to the resolution increase tendency, demand new techniques to improve the quality, but in less time. That is why it is necessary to search for paradigms that can take advantage of parallel computing such as Pulse-Coupled Artificial Neural Networks. Experimentation shows that the proposed methods are highly competitive compared versus other techniques mentioned in the specialized literature.