作者: Thomas E. Hampshire , Holger R. Roth , Darren J. Boone , Greg Slabaugh , Steve Halligan
DOI: 10.1007/978-3-642-33612-6_1
关键词: Supine position 、 Source image 、 Process (computing) 、 Landmark 、 Matching (graph theory) 、 Markov random field 、 Image registration 、 Computer vision 、 Medicine 、 Artificial intelligence 、 Intensity (physics)
摘要: Matching corresponding location between prone and supine acquisitions for CT colonography (CTC) is essential to verify the existence of a polyp, which can be difficult task due considerable deformations that will often occur colon during repositioning patient. This induce error increase interpretation time. We propose novel method automatically establish correspondence two acquisitions. A first step segments set haustral folds in each view determines via labelling process using Markov Random Field (MRF) model. show how landmark correspondences used non-rigidly transform 2D source image derived from conformal mapping on 3D endoluminal surface mesh achieve full views. initialise an intensity-based non-rigid B-spline registration further increases accuracy. demonstrate statistically significant improvement over intensity based by composite method.