Three Dimensional Reconstruction Based On A Series Of Angiography Images

Angiography technique is an unparallel visualization tool which being used widely for the visualization of blood vessels in the human body as an significant values in the diagnosis and therapy of all kinds of vessel diseases. How to accomplish three dimensional reconstruction of vessels to provide more powerful diagnostic value is still an active research topic in the field of medical imaging. In this thesis, some problems of three dimensional reconstruction based on a series of angiography images are discussed.In this paper, we do some research on the vessel segmentation. Trace the vessel centerlines making use of the MRF theory, and detect the scales of the vessel using of morphological operators. Base on the information of the imprecise vessel centerlines and the information of vessel scale, an OTSU method being using for selecting a local threshold to accomplish the vessel segmentation.In order to settle the problem of the cardiac cycle correspondence of two coronary artery angiogram sequences with different viewpoints, instead of traditional methods which use electrocardiogram(ECG) signal, a new method based on the curve energy being proposed to matching the time sequences. In this paper, we introduce a self-adaptation method to pick up the curve of vessel centerlines, and then the curve information is used to calculate the curve energy of each frame image. In this way, we translate the problem of the cardiac cycle correspondence into the curve energy correspondence.Because of the intrinsic parameters and extrinsic parameters of camera will not change in the process of heart flopping, so that the foundational matrix will not change in theory. In a creative way, we make use of the RANSAC algorithm which being used widely in the scenery reconstruction to estimate the foundation matrix. By comparing the errors of this algorithm and the improved 8-point algorithm, we can know that the foundation matrix is estimated more accurate using the RANSAC algorithm.In the end of the paper, the theory of three dimensional skeleton reconstruction of vessel be researched. Match the feature points first and then join these points using the B-sample to get others three dimensional points. Introduce three dimensional representation of vessels based on the model of generalized cylinders and display it by dint of VTK.