Automatic detection of spine axis in 3D CT data

The thesis deals with the topic of automatic detection of spine axis in 3D CT data. The theoretical part discusses the basic principles of image processing and focuses on issues of collection and processing of medical image data, particularly data from X-ray computed tomography. The entire chapter is devoted to the principle of the creation of 2D and 3D CT images. Then the article about the anatomy of the axial skeleton is elaborated focusing on the anatomy of the spine. The second part is devoted to the proposal and implementation of methods of segmentation of the spine using a simple thresholding and morphological operations to locate the initial position of the spine axis with the definition of the beginning and the end of the axis using cross-correlation with created binar masks. Subsequent refinement of the position of the spine is done by segmentation of the spinal canal using a region growing method, when the starting points, located inside the spinal canal, are acquired by finding the centre of gravity of a vertebra or by Hough transform