White Matter Fiber Tracking And Application In Diffusion Tensor Imaging

Diffusion Tensor Imaging (DTI) is one noninvasive imaging modality based on Magnetic Resonance Imaging (MRI). By detecting the diffusion anisotropy of water in cerebrum, DTI can reveal the information of the structure of the tissues and the exchanges of water molecules with each tissue in pathology. One of the important applications in magnetic resonance diffusion imaging is to reconstruct neural fibers in the brain white matter with the diffusion tensor imaging data, and it is the unique way to reconstruct neural fibers in vivo.First, this paper analyzes the line propagation algorithms and discusses their disadvantages. Then we proposed an adaptive step of FA vector criterion tracking method under the context of multi resolution. This method yields smoother results by using Taylor 2nd order series for performing intravoxel interpolation and this improvement increases speed of the method and reserves more details.Second, as a global energy minimum algorithm, fast marching method gets a lot of false pathway and can't reflect the fiber tendency correctly in particular region. We proposed a topology preserved fast marching method. During the propagation, this method adds topology information in the model and a curvature weighted speed function is proposed to adapt the new model. It can be assumed that bending energy is included in the global energy. The original FM method just uses the local likelihood as the measure. Topology preserved FM method can better control the propagation of interface and describe the fiber directions. Comparisons with standard FM, this imp is more accurate and smooth in reconstructing trajectories and more robust to noise.Finally we apply our fiber tracking method to the brain tumor. By using the MRI and DTI images of the same patient, the fusion result can reflect the abnormal position of the fiber due to oppression and pathological changes of tumors, and it can show the spatial relationship of the fibers and tumors. This application improves the utility of clinical application and research.