Research On 3D Model Feature Recognition Of Human Spinal Lumbosacral Segment And Its Mechanical Characteristics

In recent years,the incidence of lumbar spine disease in our country’s population has increased significantly,and there is a trend towards the development of this disease at a younger age.Because of the complexity of the lumbar spine,the lumbar spine surgery is difficult and involves a significant risk.As the level of scientific and technological development continues to advance,the diagnostic methods of lumbar spine diseases and the techniques of lumbar surgery continue to improve.Yet,the existing surgical simulation and surgical navigation still have not reached the desired level of clinical diagnosis and treatment.The key to solving this problem is by developing a high precision biomechanical modeling,combined with 3D geometry model for an accurate positioning and simulation of the actual operation process of the lumbosacral spine segment.This will assist the surgery preview and help in recognition of the related lesions.Thus,the process of diagnosis and treatment plan becomes quantitative,and there is a significant reduction in the existing risk in the surgery process.This dissertation takes into account the National Natural Science Foundation and other projects as research background,and builds a series of three-dimensional visualization of human spine models,combined with computer three-dimensional visualization technology,and carries out a series of research on biomechanical modeling and analysis methods of spine lumbosacral sections based on CT data.This research includes the denoising and optimization post-processing of CT images of the spine,the accurate annotation of the feature points of the 3D model of the spine and the construction of a statistical model considering the 3D geometry of the spine based on the feature points annotation.Furthermore,the integrated mechanics model of a lumbosacral segment of the spine and the biomechanical analysis was also done.The related technological research yielded some innovative research results,in the following areas:1)In this study,aimed at the problem of insufficiency in the reservation of medical spine images,This dissertation proposed a method of denoising and adaptive diffusion based on the bilateral filter operators.In this method,The model initially uses bilateral filtering to extract the contour lines of the vertebral 3D model under multi-scale conditions.This design improves the adaptive diffusion coefficient to better control the entire diffusion process.Subsequently,according to the discrete characteristics of the image,the corresponding discrete iteration was established.The equations discretized the iterative process,and iterative stopping criteria were designed to stop the iteration when the denoising-smoothed image model and the noise correlation were minimized,thereby finally establishing a medical spine image denoising model.In comparison with the classical PM method and Catte method,this method achieved a better denoising filtering effect in solving denoising,and also maintained the local edge detail feature of the image.2)To address the problem of low accuracy of existing feature points of the human vertebrae,this disertion proposes an improved method for automatic annotation of the spine feature points,based on the maximum value of normal curvature and vector inner product.This method can dynamically adjust the curvature of the pick point,so as to maximize the precision of the pick-up feature points.Initially,the curvature maxima of a manually selected point is obtained by using the multi-curvature features like the Gaussian curvature and Mean curvature flow,and the relative maxima of the normal curvature of all the model points within the specified minimum radius(r)are calculated.The maximum values of the curvature are sorted in a descending order,and n candidate points(n<=5)with a higher value of the curvature are selected.The vector inner product between the candidate points and the manually selected points is chosen to obtain the angle between the vectors.The higher the curvature is,the greater is the bending degree of the 3D model surface at that point.In other words,it is the better representation of the geometric contour information of the 3D model.The smaller the geometric angle between the vector inner products indicates that the Euclidean distance between the two points is closer,so the manually picked points are replaced by the candidate vertices with the smallest angle,which can reflect the local characteristics of the point accurately.After comparing and analyzing the experimental results,it was found that the new algorithm improves the tagging accuracy of the feature points by about 35%,which ascertains the effectiveness of the proposed algorithm.3)Due to the lack of high precision integrated spinal lumbosacral model and the corresponding high accuracy mechanical analysis and comparison,an integrated biomechanical model of the lumbosacral segment was proposed.Initially,the 3D model of the vertebral body in the sample database of geometric vertebral body statistics model was used.On the basis of the model,the model was used to simulate the attachment of the soft tissue to the main vertebrae such as the anterior ligament,posterior ligament,yellow ligament,fibrous ring,and nuclear medulla.Subsequently,on the basis of the mechanics analysis model,and using the ANSYS finite element analysis software,the youth lumbar and the lumbar intervertebral disc herniation lesions and normal intervertebral disc bulge lesions in the lumbar spine and the elderly lumbar three-dimensional spine model was established.Furthermore,from the simulation calculation of the different operating conditions and the finite element mechanics analysis,it can be concluded that the model established the experimental data under different motion states,namely the youth lumbar,the lumbar intervertebral disc herniation lesions and the lumbar intervertebral disc bulge lesions,and the elderly lumbar in applying the same load condition.Also,the model exhibited the simulation proneness after buckling,stretching,left bending,the right side of the bending in the process of deformation,displacement and stress nephogram.Thus,from the mechanics analysis model and comparison of experimental data simulation software,the present study summarizes the biomechanics characteristics of the spine,the maximum intensity that different vertebral model can withstand to,and the differences in the spine of different individuals.The mechanics theory provided by our research can help in the lesion recognition,thereby assisting the clinical medicine.Moreover,a precise mechanics model can provide a fair reliability in the diagnosis,the safety of the surgical procedures,and can provide insights about the study of the spinal cord lesions,which will ultimately offer an important guiding significance in the reasonable treatment for different spine disorders.