Feasibility of Atlas-Based Segmentation of the Brain in the Presence of Tumor by a Weighted Least-Squares Demons Algorithm

Keywords

abnormally according advantageous alignment allows altered although analogy another appears application around assumed assumption atlas better bias brain brightness choice coincides common computing conic consequently constant cost database deformation demons displacement easy efficacy efficient electrical embedded engineering equals errors estimation estimator estimators experimental explicit expression fail feasibility field fixed force forces function functions gives gradient gross ground healthy hospitals identical implement implementation improve influence insertion intensities intensity introduced iterative iteratively largely leads least like locally locations long maps maxwell measure medical merely minimization minimizes minor modifications modified moving noise nonparametric notation obvious orange outliers pathologies pathology plenty possibilities possibility practice preferably preliminary presence process processing propagate qualitative radiology reduce registration regression regularization relies replacing residuals residues respectively rise robustness segmentation segmented setup similarity simulate simulated smooth solved speech squared squares starts step subject target term towards truth tumor tumors tutorial undesirable update validation varying vector vicinity violated vision visually volume workshop