Semi-automatic Spine Segmentation Method of CT Data

Computer Tomography modality perfectly visualizes hard tissues - cancellous and cortical bone. For purpose of spine segmentation CT data were chosen. Main goal of the presented work is to achieve independent sets of segmented vertebrae and intervertebral discs by minimum effort of manual user interaction. Additionally, vertebrae structures that are built with two bony tissues, should be segmented independently. To detect discs, volume between vertebral endplates is analyzed and then classified as a disc. The key points in the proposed algorithm is to prepare manual area correction of joints area and separation in intervertebral level and automation in seed detection for region growing method based on anatomical vertebra characteristic. Segmented data can be used for generating surface and volume meshes for custom spine models for Finite Element Modeling and analysis.