Cardiac And Arterial Motion Tracking Based On Snake Model

Coronary artery disease has been the first cause of death threatening people inmost industrial countries. X-ray angiography is one of significant imaging techniquesto diagnose coronary artery disease. Since configuration and dynamics of coronaryarteries in one or several cardiac cycles are recorded in angiogram sequences, analysisfocused on detection and tracking of moving arteries can be carried out, which ishelpful for diagnosing coronary artery disease in clinic. Coronary arterial motiontracking from angiogram sequences has attracted increasing attention over the pastdecades. The main purpose of this dissertation is to extract 2D and 3D dynamicinformation of coronary artery using the technique of image analysis and snake basedmethod.Major work of this dissertation can be include in four parts as follows:1. Discussion of the principle and clinical application of X-ray coronaryangiogram. X-ray angiography is an important imaging modality for diagnosingcardiovascular disease and assessing cardiac function. The imaging principle, as wellas main components of the device is introduced in the dissertation.2. Introduction of the principal and application of active contour modelActive contour model(snake) is an energy minimization model, which has beenapplied in computer vision and image segmentation widely. Its position is updateddue to internal and external forces, derived from image features and user definedconstrains.3. Two-dimensional motion tracking of coronary arteriesA method for tracking dynamic coronary arterial vessels from X-ray angiogramsequences is proposed. It is based on active contour model, which deforming from itsinitial position to the optimal position through minimizing the proper energy function.The initial plan of the snake is obtained manually or automatically using the techniqueof ridge detection. Image features and similarity measurements of vessel axes atconsecutive instants are incorporated into the external energy to implement accuratetracking.4. Three-dimensional motion tracking of coronary arteriesA Top-down method for tracking 3D coronary artery axes using angiogramsequences on two approximately orthogonal views is proposed. 2D skeleton extraction,3D reconstruction and 3D motion tracking are implemented simultaneously, through3D-snake model deforming directly in 3D to adjust its two projections to image data.Hence, 4D (3D+T) representation of arterial vessels during the overall cardiac cycle canbe obtained.