Registración intermodalidad de imágenes médicas.

Abstract

It was investigated the intermodality registration of medical image in the Physical and Mathematics School of the Sciences Faculty of the Polytechnic School of Chimborazo. Through the implementation of algorithms acquired public basis of the National Library of Medicine: Insight Segmentation and Registration Toolkit (ITK), in the images of the public base data ATLAS, it is performed an analysis of the performance of the different techniques of registration, proposing a logarithm developed in MatLab (Matrix Laboratory) that minimize the mutual information. The images are pre-filtered in order to increase the robustness if the process against noise; a metric is applied to calculate the value of the intensity of the still image in an assigned position of the moving image, due to the computational cost, it was used an optimizer of gradient descent. The method is a result efficient in a 100% in the registration intramodality and intermodality images. In the intermodality images registration base on 2D images of Magnetic Resonance, Single Photon Emission Tomography and Computerized Tomography, under an automatic processing with transformations of Afin type with filter of Entropy Histogram, with metric of Mutual Information Histogram type with the optimizer of gradient descent wich proved to be the ore optimum development of this work. It is recommended that, if there is no a software in wich is executed the algorithms, that improve the result of registration of images are maintained the work original done.