Segmentation of Intervertebral Discs, Trachea and Spinal Cord from MRI Images

For preoperative planning and intraoperative support, segmentation of soft tissue structures from the cervical region is a prerequisite. Main activities have been started in the past to segment vertebrae from CT images. But soft tissue structures have been neglected. There is only little research work about segmentation of the spinal chord and the trachea, however as far as we know there is no work about segmentation of intervertebral discs. In this paper a new segmentation method is presented for cervical discs, trachea and the spinal chord. The approach implements a continuous process, which uses neighboring information and medical knowledge to accomplish automatic segmentation. Even seed points are not needed since localization of the discs is done using a generalized Hough transform and a cluster analysis. Additionally, for a complete view of the cervical region a rule based analyzer algorithm is briefly presented, which uses anatomical knowledge to find out attachment points on the surface of the cervical vertebrae.