Gaussian Mixture Model based Statistical Segmentation of Neuroimages
Jiabin Wang, Dagan D. Feng · 2007
2. Introduction • Dual-modality PET-CT imaging are now a routine component of clinical practice. However, medical image segmentation methods, as listed below, have generally only been applied to single modality images. • Expectation-maximisation segmentation (EMS) algorithm [1]; • Statistical Parametric Mapping (SPM) package (version 8); • Voxel-based morphometry (VBM) package (version 5) [2]. • The difficulties of brain PET-CT image segmentation consist of • Low spatial resolution and high levels of noise in positron emission tomography (PET) images • Low contrast computed tomography (CT) component • Use the complementary functional (PET) and anatomical (CT) information.