UNSUPERVISED AUTOMATIC WHITE MATTER FIBER CLUSTERING USING A GAUSSIAN MIXTURE MODEL

Meizhu Liu (1) , Baba Vemuri (2) , Rachid Deriche (3)

1 Department of Computer and Information Science and Engineering
2 UF|CISE - Department of Computer and Information Science and Engineering [Gainesville]
3 ATHENA - Computational Imaging of the Central Nervous System

Fiber tracking from diffusion tensor images is an essential step in numerous clinical applications. There is a growing demand for an accurate and efficient framework to perform quantitative analysis of white matter fiber bundles. In this paper, we propose a robust framework for fiber clustering. This framework is composed of two parts: accessible fiber representation, and a statistically robust divergence measure for comparing fibers. Each fiber is represented using a Gaussian mixture model (GMM), which is the linear combination of Gaussian distributions. The dissimilarity between two fibers is measured using the total square loss function between their corresponding GMMs (which is statistically robust). Finally, we perform the hierarchical total Bregman soft clustering algorithm on the GMMs, yielding clustered fiber bundles. Further, our method is able to determine the number of clusters automatically. We present experimental results depicting favorable performance of our method on both synthetic and real data examples

Mots clés: Fiber clustering; Gaussian mixture model; Gaussian process; total Bregman divergence; total square loss.; total square loss

Domaines: Imagerie

Meizhu Liu, Baba Vemuri, Rachid Deriche. UNSUPERVISED AUTOMATIC WHITE MATTER FIBER CLUSTERING USING A GAUSSIAN MIXTURE MODEL. IEEE International Symposium on Biomedical Imaging (ISBI)., May 2012, Barcelona, Spain. pp.522-525. ⟨hal-00712673⟩