2D Information Processing And 3D Representation Of DSA Images For 3D Reconstruction

Digital subtraction angiography (DSA) is a widely used technique for the visualization of blood vessels in the human body and has been used in clinic for more than 20 years,which has significant values in the diagnosis and therapy of all kinds of vessel diseases.In this paper we mostly research vessel segmentation, vessel centerline extraction and the three dimensional representation of vessels in the three dimensional reconstruction .Vessel segmentation is the base of 3d reconstruction on Digital Subtraction Angiography images. In this paper we propose local thresholding approach of vessel segmentation based on knowledge for DSA images. The original DSA image is divided into several appropriate subimages according to a priori knowledge of the diameter of vessels. Then we apply local thresholding techniques to choose an optimal threshold for every subimage respectively. Finally an overall binarization of the original image is achieved by syncretizing the thresholded subimages. We do experiments on cerebral and hepatic DSA images, showing that our proposed methods yield satisfactory binary results efficiently.In the aspect of vessel centerline extraction, we propose multiscale vessel tracking using Gabor filters to extract and track vessels. The flexible frequency bands and enhancement effects of Gabor filters,are fully utilized to enhance centerlines of the blood vessels in various size. Than we choose a source and a goal in the enhancement response image to track the vessel by minimum cost path .This method should be able to cope with close proximity or crossing of vessels as well as with varying vessel widths including large stenoses and severe imaging artifacts.In the three dimensional representation of vessels, we firstly discuss the registration of two views and the skeleton reconstruction, than mainly introduce three dimensional representation of vessels based on the model of generalized cylinders. The second step includes acquirement of the three dimensional radius of vessels, smoothing centerlines and space orthogonal coordinate transformation. The proposed method is simple and efficient, which can show full three dimensional vessels.Finally, surface rendering is used to show the whole three dimensional vessels by VTK. This method is fast and show living result.