Vector Field Streamline Clustering Framework for Brain Fiber Tract Segmentation

TL;DR

This paper introduces a novel vector field streamline clustering framework for brain fiber tract segmentation, enhancing brain disease analysis by improving segmentation accuracy and consistency.

Contribution

The paper presents a new clustering framework that combines vector field representation, streamline simplification, normalization, and advanced clustering algorithms for brain fiber segmentation.

Findings

Effective segmentation of brain fiber tracts demonstrated.

Quantitative and qualitative evaluations validate the approach.

Potential for automatic, robust fiber bundle template creation.

Abstract

Brain fiber tracts are widely used in studying brain diseases, which may lead to a better understanding of how disease affects the brain. The segmentation of brain fiber tracts assumed enormous importance in disease analysis. In this paper, we propose a novel vector field streamline clustering framework for brain fiber tract segmentations. Brain fiber tracts are firstly expressed in a vector field and compressed using the streamline simplification algorithm. After streamline normalization and regular-polyhedron projection, high-dimensional features of each fiber tract are computed and fed to the IDEC clustering algorithm. We also provide qualitative and quantitative evaluations of the IDEC clustering method and QB clustering method. Our clustering results of the brain fiber tracts help researchers gain perception of the brain structure. This work has the potential to automatically create a robust fiber bundle template that can effectively segment brain fiber tracts while enabling consistent anatomical tract identification.