Formation of parametric images in positron emission tomography using a clustering-based kinetic analysis with statistical clustering

摘要: A method is proposed for forming parametric images in positron emission tomography, using clustering kinetic analysis. To overcome the dual problems experienced voxel-based data, of signal noise and very long computational time, data are clustered before parameter estimation, then an estimation procedure applied to averaged each cluster. Using this algorithm, PET optimally clustered, depending on that present, by hierarchically applying a statistical-clustering algorithm based mixed Gaussian model. In computer simulation, correctly noise-contaminated data. Applying /sup 18/F-FDG clinical physiologically acceptable glucose metabolism brain were obtained practical calculation time.