The dissertation focuses on the nonrigid registration of medical images. We had developed some automatically algorithms of nonrigid medical image registration, which were based on intensity. The main contributions of this dissertation are as follows: 1. An effective method for nonrigid registration of medical images is proposed by introducing an adaptive force, which can enhance the forces as the displacements become small. The new scheme can decrease time cost greatly and don't lose its accuracy. The proposed method was applied to synthetic images and intersubj ect registration of brain anatomical structure images. The experimental results demonstrated that it is of the high efficiency and accuracy. 2. We assume that the local shape variations of brain subject to Maxwell model of viscoelasticity, the deformable fields are constrained by the corresponding partial differential equations. Based on the above assumptions, a new method for nonrigid registration of medical images is proposed. Experimental results showed that the performance of proposed method was satisfactory in accuracy and speed. 3. A free-form deformation based on a LSB B-Spline is proposed, which can enhance the shape-control capability of B-Spline. The experimental results indicate that the proposed method is much better to describe the deformation than the afflne algorithm and B-Spline techniques.
修改评论