作者: R. Ogniewicz , T. Vehkomäki , G. Székely
DOI: 10.1007/978-3-642-49351-5_56
关键词: Computer vision 、 Engineering drawing 、 Edge detection 、 Segmentation 、 Canny edge detector 、 Path (graph theory) 、 Image (mathematics) 、 Voronoi diagram 、 Edge (geometry) 、 Closing (morphology) 、 Artificial intelligence 、 Computer science
摘要: Segmentation techniques which employ edge detection suffer from the fact that generated maps usually do not represent connected boundaries. However, in order to perform a segmentation of image one needs regions defined by closed contours. We present an approach segment 2D radiological images (MR and CT slices) based on Canny detector followed gap closing operation exploits geometrical characteristics Voronoi tessellation points. Thereafter, contours are extracted global path optimization dynamic programming technique.