A Brain Fiber Visualization System Based On HARDI Data

Medical data visualization is a very important application in the field of medical-assisted diagnosis.With the development of Diffusion Weighted Magnetic Resonance Imaging(DW-MRI),the researchers use the Diffusion Tensor Imaging(DTI)model to reconstruct the micro-sized brain fibers in the white matter by measuring the anisotropy of the diffusion of wat er molecules in the fibers.This technique can be used in the medical-assisted diagnosis of Parkinson's disease,depression,Alzheimer's disease and some other diseases.Due to the limitations of the DTI model itself,the brain fiber image can't show the fiber cross.High Angular Resolution Diffusion Tensor Imaging(HARDI)is used to provide more accurate auxiliary images to the medical personnel.The technique,which uses high-order tensor models and other models,can accurately reflect the cross-flow of fibers in low anisotropy areas.The raw data obtained by the modeling method is also known as HARDI data.In order to visualize,interact and analyze the HARDI data,this paper designs and implements the brain fiber visualization system based on HARDI data.In the system,the visualization technique is used to enhance the rendering effect,and the visual result is analyzed by mathematical methods such as scatter matrix and clustering,and the external somatosensory interaction device is added for interactive enhancement.The main work and achievements of this paper are as follows:(1)A system framework for visualization of brain fiber data is designed.The framework includes a hardware base layer,a data storage layer for storing data,a data processing layer for preprocessing data and drawing operations and a display layer for visualizing display,analysis,and interaction.(2)The method of drawing enhancement and visual analysis is added to the fiber drawing results.The volume of the brain fiber is large and the density is high,which leads to the decrease of the perceived ability of the fiber.The system enhances the effect of fiber drawing by adding light and anti-aliasing,and filters the fiber by local analysis and scatter matrix.In addition,the clustering method of fast peak density search is used to cluster the results.(3)Add somatosensory devices for the system.Traditional interaction is based on the mouse and keyboard,etc.,the development of somatosensory technology marks a more ergonomic means of interaction will appear.The system adds space gesture interaction through Leap Motion,and on this basis,it presents an interactive gesture for sphere data.At last,we implement the prototype system of brain fiber visualization,demonstrated the rationality of our design method,and the system has a good scalability,easy for follow-up to further development and upgrading.