Research On Personalized Modeling Method Of Human Soft And Hard Tissues Based On Medical Images

Due to medical models based on medical images can reflect the characteristics of patients,they play a very important role in the field of medical education,medical diagnosis,and prosthesis design.In the modeling process,for the hard tissue,the large difference in CT value from the soft tissue background makes its segmentation effective.At the same time,due to the high density,metal artifacts and beam hardening artifact are prone to occur during image acquisition.The usage of radiation has deepened the complexity of the problem.Compared with hard tissues,soft tissues use MRI for image acquisition,and the error introduced is minimal.However,due to the similar density between soft tissues and fat,take the heart for example,there are problems such as trabecular muscles and papillary muscles that are similar in strength to the myocardial tissue,plus the uneven brightness of the left and right ventricular chambers under the influence of blood flow.The segmentation has brought great difficulties.These problems will cause the medical model to fail to achieve the expected results due to the accuracy in surgical planning,medical diagnosis,and prosthesis design.Based on the above problems,this article starts from the steps of the medical model modeling,and then summarizes a set of personalized medical model modeling methods based on medical imaging from the perspective of soft and hard tissues according to different tissue characteristics and medical imaging characteristics.The main research will divide the object into two parts: image acquisition of hard tissue and image processing of soft tissue.After concentrating on solving the key problems in the modeling process of the two types of organizations,a set of unique personalized medical model modeling methods based on medical images have been formed.The main work of the paper is as follows:(1)Establish a one-to-one correspondence between bone density and radiation dose.The "personalization" element of the hard tissue in image acquisition is attributed to bone(bone density),and a standard phantom that can reflect the "personalization" element of the hard tissue is designed and manufactured,based on this phantom,bone density is the personality.The parameters of CT are attributed to radiation dose,and the corresponding relationship between bone density and radiation dose is found.It changes the view that age,height,and gender can represent individualization in previous studies,and can provides a guidance and suggestions in image acquisition.(2)From the perspective of bone physical parameters,three-dimensional volume,and mechanical properties,the above-mentioned method of obtaining personalized images of hard tissues is verified.The above-mentioned "personalized hard tissue image acquisition" method is applied to pig femur and bovine femur,and the geometric difference between the reconstruction model and solid bones of different radiation doses is compared,and the validity of the method is verified from a geometric point of view.Using the finite element method as a tool,the effects of radiation doses(24 m Gy,48 m Gy,64 m Gy)on the stress and strain distribution of the bone reconstruction model are analyzed and compared,it is found that there are expansion and shrinkage at the low doses model and the high doses model respective.At the same time,with the radiation dose increasing,the distribution and number of the model equivalent stress will also increase.The radiation dose just only changes the strain number,not the strain distribution.(3)For the special soft tissue structure of the heart,a segmentation method based on the segmentation threshold method of tissue density and the region growth method is proposed.During segmentation,the segmental threshold method is used to divide the heart model into six different tissues(left atrium,left ventricle,right atrium,right ventricle,aorta,pulmonary artery),and then segment the above six tissue structures by the method of region growth,threedimensional reconstruction.In this process,six different segmentation thresholds are used to solve the problems of tissue involvement and fat redundancy to the greatest extent.Comparing this model with a single threshold segmentation model,whether it is a two-dimensional size structure or a three-dimensional volume,the segmented threshold model is smaller than the single threshold model.