Affine-based registration of CT and MR modality images of human brain using multiresolution approaches: comparative study on genetic algorithm and particle swarm optimization

Abstract

We present a non-linear 2-D/2-D affine registration technique for MR and CT modality images of section of human brain. Automatic registration is achieved by maximization of a similarity metric, which is the correlation function of two images. The proposed method has been implemented by choosing a realistic, practical transformation and optimization techniques. Correlation-based similarity metric should be maximal when two images are perfectly aligned. Since similarity metric is a non-convex function and contains many local optima, choice of search strategy for optimization is important in registration problem. Many optimization schemes are existing, most of which are local and require a starting point. In present study we have implemented genetic algorithm and particle swarm optimization technique to overcome this problem. A comparative study shows the superiority and robustness of swarm methodology over genetic approach.

Study on neurodegeneration at different stages using MR images: computational approach to registration process with optimisation techniques

“…In present paper, MRI brain data for patients having Alzheimer's diseases (ADs) (Horn et al 2009) have been used for experiment and for implementation of the proposed methodology of registration to assist the clinicians in prediction of the prognosis of neurodegeneration. This is a continuation of earlier work of authors where the dilation of ventricular region -a symptom of AD -was considered as region of interest (Bhattacharya and Dutta Majumder 2000;Das and Bhattacharya 2011). Changes in physiological process due to AD and how it may be captured during the process of registration are the objectives of study of the present experiment.…”

“…Affine (linear) transformation is one which includes scaling and shearing in addition to translation and rotation of rigid transformation (Das and Bhattacharya 2011). This type of transformation maps straight lines to straight lines and preserves the parallelism between lines.…”

“…To both speed up the optimisation process and avoid local optima, most currently used registration methods employ some form of multi-resolution optimisation (McGuire and Stone 2000; Jenkinson et al 2002;Das and Bhattacharya 2011). It is convenient to adapt a sequence of images at a progressively low resolution.…”

Non-rigid Multimodal Image Registration Based on the Expectation-Maximization Algorithm

“…The goal of these methods is to obtain the set of parameters describing a geometric transformation between the target and source images that optimizes the similarity metric (e.g. Mutual Information) [3], examples of these algorithms are those based on gradient descent [6], or more recent approaches based on global optimization techniques, such as particle filtering (PF) or genetic algorithms [7][8][9].…”

Dynamic brain magnetic resonance image registration based on inheritance idea and PSO

“…For practical applications, the optimization problems contain a lot of local optima, so the global optimization algorithm is welcome [1]. Evolutionary computation algorithms are global optimization algorithms and are applied in image registration problems with satisfying effect [2].…”