Bounding-object Segmentation

摘要

This paper presents a new method for the semiautomatic segmentation of anatomical or pathological structures in MRI, CT or ultrasound images. The concept of bounding-object segmentation is based on an efficient combination of a new interactive approach with well known automatic segmentation algorithms. The efficiency of this new method is based on the transparent interaction between a 3D scene as well arbitrary 2D views of the scene. Bounding-object segmentation can also be described as a combination of interactive 3D segmentation with region-based, level-set-based, and/or texture based 3D-segmentation algorithms.