Purpose
Nonrigid image registration using a cubic B-splines transformation model is well known approach in medical imaging. Although, it is successfully applied to medical image registration applications, it exhibits computational complexity due to the lack of dedicated similarity measure and optimization algorithms. Therefore, the main purpose of this paper is to propose a simple and computationally efficient similarity measure with a suitable optimization algorithm.
Methods
In this paper, an efficient similarity measure is proposed for automatic nonrigid medical image registration. The proposed approach is based on Gerschgorin circle theorem and covariance matrix properties. In this approach, image registration is carried out by similarity calculation between two normalized images. The proposed similarity measure is optimized using the Levenberg-Marquardt back propagation (LMBP) algorithm.
Results
Experimental results for various 3-D magnetic resonance data volume and also clinically acquired 4-D CT image datasets are presented in order to show the effectiveness of the proposed approach. It is observed that the proposed approach is sensitive to a small nonrigid deformation field as well as the overlapping.
Conclusions
The various simulation results demonstrated that the proposed approach can be effective and promising option in complex 3-D medical image registration as well as in various image processing applications.