Research and implementation of spinal MRI image segmentation algorithm

摘要

Using the method of image visual processing to assist clinical diagnosis is an important technology of medicine field, because of this, gradually become a hot research topic in recent years. Spine MRI image contains important information of spine and plays an active role in diagnosis, therefore, there must be a robust method for the image segmentation and extract the useful information, this is also helpful for the subsequent 3D reconstruction. In this paper, we propose a fully automatic, unsupervised segmentation algorithm of intervertebral disc using an improved marker controlled watershed transformation which combines intention and prior shape information. Experimental results show that our algorithm can accurately extract the outline of intervertebral disc from normal spine and different degree of deformation of the spine. Obviously, this method is robust enough to deal with different shapes of disc.