Research On The Key Technology Of Elastic Imaging Based On IVUS Image

According to the WHO statistics, death caused by cardiovascular disease has become one of the three major causes of death in humans. Medical research shows that vulnerable plaque rupture caused by Atherosclerotic(AS) is the leading cause of acute cardiovascular disease.But, there are no effective tools to predict plaque rupture at present. Plaque characteristics mainly depends on its composition and geometry, and these are related to the elastic properties of vascular wall, therefore,research on the vascular wall elasticity for the early identification of vulnerable plaques and reducing cardiovascular disease has important clinical significance.Intravascular Ultrasound(IVUS) as the "gold standard" for diagnosising cardiovascular disease can provide cross section image of the vessel,it can also help to recognise morphology of vascular wall, or to determine the position of plaque, but it still can not distinguish the composition of vulnerable plaque, Under certain blood pressure, plaque will exhibit different elastic properties from peripheral tissues and blood vessels. Therefore, analyzing the response of the vessel wall and plaque under blood pressure can evaluate the possibility of vulnerable plaque rupture.The main contents and innovative achivements are as follows:(1)Noise within IVUS image is mainly speckle noise and Gaussian noise,fefined a reduction method combined with median filter and image average comprehensive method. This method can effectively remove noise and keep image smoothing.In addition, on the basis of improved image edge extraction algorithm,the initial contour edge of outer membrane using an improved threshold segmentation algorithm, with weighted gray gradient instead of the simple average gray value method, the edge point gray brightness and statistical probability distribution as the constrained cost function for optimization calculation, get the final contour.(2) In meshing of the vessel wall, invalid triangles in the luminal area influence subsequent matches and elastic calculation accuracy, this paper proposes a method based on the limitted point set meshing method of the blood vessel walls, added a threshold angle between points as screening criteria edge points, and finally in accordance with the positional relationship between the edges and the center of the triangle delete invalid triangles within the lumen. On this basis, we propose amatching method based on B-spline deformation model,for traditional normalized cross-correlation similarity measurement might bring error for calculation of isolated areas, the cost function with gray weights is proposed; at last, I-BFGS optimization algorithm with fewer iterations to calculate the Hessian matrix to find the best match point.(3) For traditional vascular elasticity estimates rely on prosthesis simulation method, the experimental device is complex and difficult to ensure accuracy of the results, this paper makes full use of IVUS image information, without additional device, according to the existing deformation theory, put forward the position of the line element change instead of the surface element deformation method under two-dimensional rectangular coordinate system, and successfully deduced Green Strain formulas under small triangular element conditions and calculate the corresponding strain values, the strain value is converted to a value corresponding to RGB color space in the use of pseudo-color coding and fills the small triangle unit with corresponding color.Comparison of artificial deformation by Photoshop with the calculation results proved a series of image processing and elastricity analyzing method is feasible and accurate in this paper, be qualified to provide a reference for clinical diagnosis and treatment of cardiovascular diseases.