Estimation Of Cardiac Dynamic Parameters And Statistical Analysis Based On Heart Image

Cardiovascular disease is the highest morbidity and mortality of disease in the world, which is concealed and acute. So the early diagnosis and prevention becomes especially important. To estimate the global and motional functions of the heart via cardiac images is one of most effective way in clinical medicine, which is increasingly interesting and also an open research field.It is proved in clinical investigation that abnormal heart has the behaviors including the weakened, intensified and abnormal motion. However, due to the lack of criteria, the traditional methods only take the weakened region as the pathological. A discriminant function based on probability model of dynamic parameters is constructed to tell the pathological area from the normal. The main work and the results of this paper are as follow:(1) By comparing the cardiac function parameters, it is found that many kinds of heart diseases are strongly correlated to cardiac dynamic mechanics, so understanding the heart's mechanics of movement is crucial for clinical investigation as well as for diagnosis and patient care.(2) In this paper, the motion path of the key point in a cardiac cycle is successfully recovered via cube spline interpolation. And the results are used to estimate the velocity, acceleration of the key points and a novel parameter - Movement Extent of Myocardium firstly presented in this paper. These parameters can track the motion abnormity caused by cardiac fibrillation.(3) The data is extracted from normal heart images based on ASM method to build analysis set. After cutting out the error points produced by segmentation and estimation, a probability model of dynamic parameters is put forward to tell the abnormal range and the normal range.(4) A discriminant function based on the probability model of dynamic parameters is set up and used to judge the abnormality of two sets of cardiac images which have abnormal areas, the results show that the function can distinguish the abnormal heart from the abnormal efficaciously. It is also described in this paper how to re-project the determined pathological region to two-dimensional slices and mark the location on them in detail, which will sharply reduce the number of slices which clinicians have to read for clinical diagnosis, and help clinicians get more information from the original slices at the same time.As aforementioned, the approach presented in this paper can be used to aid the clinical diagnosis especially in early stage of the disease. If more patients' images and more parameters can be introduced to set up the sample set, the more accurate statistic method can be brought up to aid the clinical diagnosis efficiently.