Visualization Of Brain Fiber Tracking Based On HARDI Model

Diffusion Tensor Imaging(DTI) is difficult to describe sophisticated microstructure of neural fiber, while high angular resolution diffusion imaging(HARDI) is an effective method for characterizing complex neural fiber paths in the human brain. In addition, visualizing and analyzing the fibers is often challenging because of the complexity of the fiber orientation distribution function used to describe the crossing, kissing, and fanning fibers. It is difficult by only using interactive methods like mouse and keyboard to handle the neural fibers in 3D space. Probabilistic fiber tractography generate redundant neural fibers, and some mistake neural fibers will overlap the correct result. In this paper, we propose a novel visual analytics approach to study brain fiber paths that allows users to explore fiber bundles to reveal the probability of fiber paths using a new visual classification method.This paper summarizes the domestic and foreign background and development in human brain neural fiber research, also discusses the MRI, brain fiber tractography and the visualization of the brain fiber. This paper analyses the development process of research results and their deficiency. The detail work is listed as follows:Firstly, due to DTI technology can only depict one direction of neural fiber orientation in single voxel, this paper use HARDI model to describe the crossing, kissing, and fanning fibers. In addition, we propose a novel prior probability method base on neural fiber tracking result. This method integrates the neural fiber tracking result and current tracking process.Secondly, in this paper, we propose a visualization tool, a pixel bar, to display multidimension information of the neural fiber. Pixel bar is a kind of convenient and intuitive interactive means. To solve the difficulty in handling 3D neural fibers, we propose a neural fiber and pixel bar mapping method, including color mapping and transparent mapping. In addition, DBSCAN is used to clustering the pixel bars, and sort by the transparent of pixel bars. The users obtain the confidence neural fibers by selecting low transparent pixel bars. This method solves the problem of redundant neural fibers.Thirdly, the object-oriented programmed visualization system uses Qt as user interface based on OpengGL and OpenGLSL developing libraries. The system provides necessary functions like the selection of seeding area, light effect and antialiasing of the neural fibers.After evaluating the system expert users provide positive feedbacks. Probabilistic fibers capture the main structure of the brain. The main anatomical structures in the brain, such as crossing, kissing, and fanning fibers, are clearly visible. The pixel bars are intuitive to use and easy to interact with.