A Brain Fiber Clustering Method Based on Spatial Path Similarity

Abstract

A brain fiber clustering method based on spatial path similarity, including the following steps: 1), import fiber data, track the fiber path, and calculate the similarity of each fiber; 2), due to the diversity of fiber data, there are The fibers of some tissues are relatively scattered and the number of fibers is small, while the fibers of some tissues are relatively dense and the number of fibers is very large; according to the size of the fiber data, two different fiber clustering methods are used: DPC (NGK) fiber Clustering method and DPC(GK) fiber clustering method; 3) Map the clustered fiber data to the corresponding color space for color coding and visual display; so as to achieve the effect of clearly distinguishing different types of fibers.

Description

technical field [0001] The invention relates to the research of brain nerve fibers, and is a brain fiber clustering method based on spatial path similarity. Background technique [0002] Water molecule dispersion is the basis of NMR technology. DWI, DTI and HARDI non-invasive detection technologies are all developed on the basis of water molecule dispersion and diffusion. Through MRI technology, people can already track the direction of white matter fibers. DTI is the most commonly used MRI technique to detect brain white matter structure, but its model is limited by the Gaussian assumption and cannot solve the problem of multi-fiber crossing. HARDI technology describes the displacement process of water molecules as a mixed Gaussian model, and compensates for the defects of DTI technology by calculating the characteristic direction and eigenvalue of the diffusion tensor of each component. The orientation of the fibers can be better described. [0003] The above-mentioned ...

AI Technical Summary

Problems solved by technology

Through the tracking of fibers, people have been able to obtain thousands of fibers in the brain. However, these fibers are intricate and obstruct each other, and people cannot directly observe the organizational structure of brain fibers with the naked eye.

Due to the complexity of fibers, people cannot visualize and analyze fiber trajectories well. How to cluster and render a large number of fiber trajectories is still a very important issue.

Images