作者: Frank Heckel , Jan H. Moltz , Christian Tietjen , Horst K. Hahn
DOI: 10.1111/CGF.12193
关键词: Segmentation-based object categorization 、 Sketch 、 Segmentation 、 Artificial intelligence 、 Interface (Java) 、 Computer science 、 Context (language use) 、 Scale-space segmentation 、 Computer vision 、 Image quality 、 Image segmentation
摘要: In the past years sophisticated automatic segmentation algorithms for various medical image problems have been developed. However, there are always cases where fail to provide an acceptable segmentation. these user needs efficient editing tools, a problem which has not received much attention in research. We give comprehensive overview on three-dimensional (3D) images. For two-dimensional (2D) images, we discuss sketch-based approach modifies contour domain. Based this 2D interface, present image-based as well image-independent method intuitive and 3D context of tumour computed tomography (CT). Our tools evaluated database containing 1226 representative liver metastases, lung nodules lymph nodes different shape, size quality. addition, performed qualitative evaluation with radiologists technical experts, proving efficiency our tools.