Density estimation for MR image elastic matching

Abstract

The problem of matching two images can be posed as finding a displacement field which assigns each point of the reference image to a point in the test image. In this paper we present an iterative algorithm to estimate the probability density function relating the intensity distribution of two MR scanners, based on the topological constraints embedded in the elastic matching model. The set of images used as input for the algorithm is the Harvard Atlas. The density estimation resulting from this method is compared with two other algorithms which do not assume any prior information about the media being imaged. The results show that the density estimation obtained with the elastic matching approach produce more realistic deformed images and is suitable to represent MR sensor models.