Research On Coronary Artery Segmentation And Hemodynamic Simulation Based On Improved U-Ne

Since the beginning of the 21 st century,the number of patients with coronary artery disease has been increasing.Although the elderly are the main patients with coronary heart disease,according to the hospital treatment in recent years,the age of patients tends to be younger,which seriously affects the national health.By the medical research found that coronary artery because of its unique shape structure with hemodynamic characteristics different from common blood vessels,so you can through the medical simulation analysis software of three-dimensional model of coronary arteries simulation experiments,in order to realize the under different circumstances,the internal blood flow of coronary arteries impact analysis purposes.Therefore,based on the U-Net,this paper designed a kind of internal and external nested U-shaped network with attentional mechanism,which deepened the training depth of the network and improved the segmentation effect of coronary artery vessels through the nested structure design.After the segmentation,3D reconstruction of coronary artery vessels was carried out to build a real vascular model,and the hemodynamic analysis of coronary artery vessels was completed by applying real vascular physiological parameters to each vessel to simulate blood flow.The main work of this paper is as follows:(1)In order to reduce the noise in coronary artery images,improve the definition of vascular images and the contrast between coronary artery and surrounding vessels,bilateral filtering and vascular enhancement processing were performed on coronary artery images in this paper The definition and contrast of blood vessels after processing are further improved,and irrelevant effects around coronary vessels are filtered out,laying a foundation for the precise segmentation task in the following.(2)Due to the complex region of coronary artery images,redundant vessels have a great influence on the segmentation of coronary arteries,which brings some difficulties to accurate segmentation.Based on the U-Net network,on the basis of attention by adding a space module,segmentation of coronary arteries has been designed nested U-Net network,through the nested structure design further enhance the depth of the network model,in the case of without the training model to get good segmentation model,thus may accurately from complex coronary arteries CTA images extracted.Using Io U as the quantitative analysis standard of segmentation results,the segmentation accuracy can reach 88.36% by comparing with other models,and the segmentation effect is further improved.(3)To the existing cases of coronary arteries to complete 3D reconstruction,three-dimensional vascular model,and on the basis of fluid mechanics knowledge,vascular parameters are simulated through setting of coronary artery blood flow under maximum congestion state,blood dynamics simulation analysis was carried out on the coronary arteries.Finally,the correlation between vascular stenosis and blood pressure was obtained.The nested U-shaped network with attentional mechanism designed in this paper fully considers the location and shape characteristics of coronary arteries and can achieve the purpose of complete vascular segmentation.The advantages of this network are verified by comparison with other networks.The correlation between vascular stenosis and blood pressure can be approximated through hemodynamic simulation analysis of 3D reconstructed vessels.