Evaluation of the Twofold Gaussian Mixture Model Applied to Clinical Volume Datasets

Abstract. Volume representations of blood vessels acquired by 3D rotational angiography are very suitable for diagnosing a stenosis or an aneurysm. For optimal treatment, physicians need to know the shape of the diseased vessel parts. Binary segmentation by thresholding is the first step in our shape extraction procedure. Assuming a twofold Gaussian mixture model (GMM), the model parameters (and thus the threshold for binary segmentation) can be extracted by the Expectation-Maximization (EM) algorithm. The question is whether this GMM threshold gives a good segmentation. Therefore we compared segmentations induced by the GMM threshold with segmentations induced by thresholds derived in a different way. It appeared that a twofold Gaussian mixture model is not always a correct assumption for the distribution of the gray values.

• Type: Conference
• Year: 2009
• Where: BILDMED
• Authors: Jan Bruijns
• Keywords: BILDMED 2007, BILDMED 2009, Binary Segmentation, Gaussian Mixture Model, Twofold Gaussian Mixture