Research Of Blood Vessel And Guidewire Tracking Algorithm In Endovascular Interventions

Cardiovascular disease is a serious threat to human health and life.Computeraided diagnosis and treatment of cardiovascular disease based on contrast images provides a basis for quantitative analysis of cardiovascular disease.Therefore,the accurate segmentation and registration of blood vessel and guidewire centerline structure is of great significance in clinical medicine.Based on the analysis of the domestic and foreign tracking algorithm of blood vessel and guide wire,this thesis focused on the research of the algorithm of extracting the seed point of the blood vessel and the extraction method of the guidewire seed point,and used the extracted seed point to realize the blood vessel and guidewire centerline extraction algorithm.Finally,the simulation analysis of the vascular tracking algorithm in the sequence of contrast images was completed.The innovation of this thesis is to propose a blood vessel seed point extraction algorithm based on GVF field and an algorithm of blood vessel and guidewire centerline extraction based on OAC model.In the algorithm of seeding point extraction,the GVF field with stronger robustness and stability is used as the external force field of seed point response function.Compared with the traditional seed point extraction algorithm based on symmetrical edge and local extreme point,the algorithm can extract more and more accurate seed points.In the algorithm of the center line extraction of blood vessel and guide wire,the OAC model of the improved external force field is adopted,which effectively solves the problem that the curve of the active contour model is too short.The main work done in this thesis is as follows:(1)The research of vascular seed point extraction algorithm based on GVF field was completed.The algorithm uses multi-scale filtering algorithm to preprocess the image contrast and enhance the tubular structure.According to the GVF field of the contrast image,the seed point response function is obtained,so as to realize the automatic extraction of the blood vessel seed point.At the same time,an algorithm of guidewire seed extraction based on edge detection is realized,and the automatic screening and positioning of guidewire seed points is realized by using mathematical morphology corrosion expansion technique.Finally,the results of seed extraction were analyzed experimentally.(2)The research of the algorithm of the center line extraction of blood vessel and guide wire based on OAC model were completed.In the centerline initialization process,a centerline initialization algorithm based on adaptive tracking was studied.The initial results were substituted into the OAC model based on GVF to obtain the final centerline result.Finally,the results of the extraction of the centerline were analyzed experimentally.(3)The study of image registration based vessel tracking algorithm was completed.The algorithm includes a fast algorithm based on B-spline registration algorithm and sequential updating.With Hausdorff distance as the similarity measure,the fast blood vessel registration is completed by solving the B-spline control points,and the results of the registration are analyzed experimentally.At the same time,a symbolic language based on the interpretation of vascular movement was achieved,and quantitative data of vascular movement for doctors according to the registration results was provided.