Description

Image registration means aligning two images so that common features overlap and differences are readily apparent. This paper presents a quantitative evaluation of three intensity based image registration methods applied to MRI brain images. These methods come under three categories such as global, local and improved(hybrid) methods. The aim of this study is to assess the suitability of these methods for brain image analysis. We will compare the accuracy and robustness of the three algorithms named as affine, demons and affine in combination with B-spline. Our proposal concludes and evaluates the appropriate registration method for MRI brain images.