Fast Registration Algorithm for 3D Medical Images using Double Density Dual Tree Complex Wavelet Transformand Bi-directional Prediction

This paper discuss about 3D-3D CT-MRI image registration carried out using Double Density Dual Tree Complex Wavelet Transform (DD-DTCWT) with bi-directional prediction algorithm considering 2D frames. The 3D images of CT and MRI are sub sampled by considering odd frames. The selected odd frames are registered by computing the DD DTCWT features and mutual information metric. The registered 2D frames for which transformation parameters that are estimated are used in registering the even frames with bidirectional prediction algorithm. The algorithm is designed to be reconfigurable to select transformation type such as DWT, DTCWT, DD-DTCWT. The wavelet sub bands selection logic is set as control parameters between speed and registration accuracy. The registration algorithm performance is estimated considering mutual information and joint entropy with improvement of 10% and 6% respectively.