Research On 3D Reconstruction Of Medical Image

3D reconstruction of Medical images technology is currently a hot topic in the field of medicine,which involves knowledge in the field of medicine,digital image processing and computer imaging.This technique converts the two-dimensional image sequence produced by medical imaging equipment(MR,CT)into a realistic three-dimensional medical image,which helps the doctor to diagnose the disease quickly and develop an accurate treatment plan.Thus,three-dimensional reconstruction technology has important clinical value.Based on magnetic resonance imaging(MRI)and computed tomography(CT),three-dimensional reconstruction of medical images has been studied.The main work is as follows:1.The 3D reconstruction and visualization system based on visualization toolkit(VTK)is built in C ++ programming language,and the advantages and limitations of the system are pointed out.2.As the resolution between slices of the short axis nuclear magnetic resonance image(CMR)sequence is much lower than the distance between pixels,the 3D reconstruction and visualization system based on VTK can not achieve good reconstruction effect.Therefore,a three-dimensional surface reconstruction method based on simplified pulse coupled neural network model(SPCNN)and moving cube(MC)algorithm is designed for 3D reconstruction of left ventricle.To ensure the accuracy of the reconstruction,we first use the cubic spline interpolation to increase the inter-layer resolution,and then use SPCNN to automatically obtain all the left ventricular endometrial contours.Finally,the left ventricular blood pool region was reconstructed by marching-cubes algorithm(MC).3.Four sets of experimental data in the MICCAI2009 database were tested.The results show that the left ventricular segmentation method based on SPCNN has higher segmentation accuracy and MC algorithm has the advantages of easy to implement and high imaging quality.By observing the results of three-dimensional reconstruction,we find that the effective combination of the two methods made the reconstruction results accurate and smooth.