Image registration based on neural network
Shuxiu Wang, Lei Sheng, Faliang Chang · 2008
Image registration is to seek a spatial alternation which can make a 2D or 3D image to consist with another 2D or 3D image in space; it is an important technique in image processing. The concept of image registration is introduced in detail in this paper, and the main emphasis is how to choose the feature points. The author raises a new algorithm to choose the feature point, it is easy and fast. At the same time to solve the nonlinear distorting using direct mapping transformation method, this paper registers image based on neural network. By simulations of MRI images, it demonstrates that the new algorithm is of great effectiveness.