Research On Cardiac Vortex Identification Based On Vorticity Deviation In Temporal-spatial Domain

Cardiovascular disease is a serious threat to human life and a heavy burden on society and families.Accurate diagnosis of cardiovascular disease is crucial to save the lives of patients.With the development of medical imaging technology,Vector Flow Mapping(VFM)and Phase Contrast Magnetic Resonance imaging(PC-MRI)technologies make it possible to measure the blood flow field in hearts noninvasively.In the reconstructed blood flow field,the calculation of hydrodynamics indexes can assist the diagnosis of cardiovascular diseases.The research shows that the high intensity field of vortex has good reproducibility and can clearly reflect the state of blood flow in hearts.However,the existing methods to characterize cardiac vortices are inaccurate in spatial domain and incomplete in temporal domain,and it is difficult to establish a clear relationship with cardiac function.Aiming at the above problems,this thesis systematically studies key technologies such as vortex identification algorithm,image interpolation algorithm and automated identification based on the principle of fluid particle motion.A robust and accurate method for the identification of cardiac vortices is proposed,which effectively solves the problems that inaccurate and cumbersome identification of vortices in the heart,and provides strong support for the diagnosis and treatment of cardiovascular diseases for medical staff.The main contents and contributions are summarized as follows:1)To solve the problems that existing algorithms cannot describe the evolution of cardiac vortices in the time domain,an algorithm based on Lagrangian mean Vorticity Deviation(LAVD)is proposed to identify Lagrangian vortices in hearts.The evolution of vortex during left ventricular ejection period and the influence of vortex on blood flow field are depicted.By measuring the difference of Lagrangian vortex volume between healthy subjects and cardiac dilated patients,the validity and feasibility of the proposed method in the diagnosis of cardiovascular diseases are verified.2)Motivated by the fact that the LAVD is the integral of particle motion in time,it is necessary to construct the continuous blood flow velocity field in temporal domain.AS the quality of continuous velocity field directly affects the accuracy of LAVD in identifying vortex structures.In improve the accuracy of vortex identification,a novel method of constructing continuous blood flow based on optical flow in time domain is proposed.By analyzing the optical flow field between two consecutive frames of PCMRI,the continuous velocity field of blood flow is innovatively constructed,and the Lagrangian vortex structure in hearts is accurately identified according to the LAVD calculated in the continuous velocity field.Experimental results show that the proposed method can enhance the accuracy of vortex identification and describe the evolution of Lagrange vortex in the right atrium during contraction and diastolic periods.3)In order to solve the problem that existing algorithms for cardiac vortex identification have low accuracy in the spatial domain,Instantaneous Vorticity Deviation(IVD)is obtained based on the LAVD.The evolution of Euler vortex in systolic and diastolic period of left ventricle and the influence of dominant vortex on blood flow field are depicted.The value of the method in the diagnosis of cardiovascular disease is demonstrated by measuring the difference in the vorticity of the Euler vortex between healthy subjects and cardiac dilated patients.4)In order to solve the problem that LAVD and IVD cannot automatically identify cardiac vortices,a novel and robust ellipse detection method based on Hough transform is proposed.In this method,the vortex structure is extracted automatically by Hough transform,and the tubular characteristics of the identification results are maintained.Without setting threshold,the Lagrangian vortex and Euler vortex in the right atrium is identified automatically.Experimental results show that the proposed automated identification algorithm improves the stability,robustness and generalization ability of cardiac vortex identification.