A Clustering Method of Cerebral White Matter Fiber Tracts Based on Fiber Midpoints and Endpoints

Abstract

The present invention proposes a method for clustering brain white matter fiber bundles based on fiber midpoints and endpoints, aiming at improving the accuracy and efficiency of brain white matter fiber bundle clustering. The implementation steps are: (1) obtaining the whole brain fiber set; ( 2) segment the whole brain fiber set; (3) determine the fiber midpoint; (4) midpoint clustering; (5) endpoint clustering; (6) determine the endpoint clustering result; (7) obtain the brain white matter fiber bundle clustering result. The present invention uses the position information of the fiber midpoint and end point as features, and divides the white matter fiber bundles of the brain into several fiber groups through DBSCAN density clustering. The accuracy of clustering is improved, while the amount of calculation is reduced, and the efficiency of clustering is improved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to a method for clustering cerebral white matter fiber bundles, in particular to a method for clustering cerebral white matter fiber bundles based on fiber midpoints and endpoints, which can be applied to the auxiliary research of cerebral white matter fiber bundles. Background technique [0002] White matter, gray matter and cerebrospinal fluid are the components of the brain, but their internal components are different. Therefore, the water molecules in different brain tissue structures will have different diffusion behaviors. The water molecules in the cerebrospinal fluid mainly show free diffusion. The water molecules in the gray matter are in a state of restricted diffusion, while the water molecules in the white matter are bound by the myelin fibers and neuron fibers, and suffer less resistance in the direction parallel to the fiber bundles, and the diffusion speed is no...

AI Technical Summary

Problems solved by technology

The disadvantages of the above two methods are: by calculating the distance matrix and then calculating the affinity map or calculating the similarity matrix to find the eigenvalue as the clustering feature, it may reduce the sensitivity of the feature, resulting in the accuracy of the clustering result. Insufficient, calculating the distance matrix at the same time needs to calculate the point-to-point distance between two fibers, which requires a large amount of calculation and low efficiency

Images