Research On Segmentation Algorithm Of Lung Fissure Image Based On Skeletonization

In recent years,environmental problems have become increasingly prominent,and many harsh environments such as smog have not only affected people's food,clothing,housing and transportation,but also seriously affected people's health.There are more and more people with lung-related diseases such as pneumonia,tuberculosis,emphysema,and lung cancer,which have a great physiological and psychological impact on patients.When medically diagnosed with lung diseases,doctors usually look at a large number of CT images to identify the diseased parts of the lungs,which is extremely time consuming and energetic.The lung fissure is the physical boundary of the lung lobe.The lung lobe is divided into five parts.The segmentation of the lung fissure can more accurately locate the lesion.Therefore,segmentation of the lung fissure has important research value.In order to effectively segment the lung fissure,this paper proposes to use the skeleton-based method to segment the lung fissure according to the characteristics of the three-dimensional CT image,and obtain a good segmentation effect,which has important practical value.In this paper,the lung CT image is preprocessed,the lung parenchyma is extracted and the region of interest is defined,and then the lung fissure is displayed as a linear structure in the 2D CT image.The lung fissure is displayed as a surface structure in the 3D CT image.The lung fissure enhancement filter is used to enhance the lung fissure,and then the image is skeletonized,the branch points are removed,the joint meta-analysis is processed,and finally the hole is filled to complete the post-treatment of the lung fissure.The specific research work of this paper is as follows:In this paper,a simple and effective common lung parenchyma extraction algorithm is adopted in the extraction of lung parenchyma.The lung parenchyma extraction algorithm is divided into two steps.The first step is the initial segmentation of the lungs,using a thresholding segmentation algorithm,while the initial mask in the lung region is the initial segmentation of the lungs;the second step is the lung repair,using morphological principles and three-dimensional connectivity analysis methods,and finally The lung mask is the result of lung repair.Through the above two steps,the lung parenchyma can be extracted and the lung area of interest can be isolated.Because of the numerous vascular morphology of the lungs,in order to accurately segment the lung fissure,it is necessary to first strengthen the blood vessel and suppress the background noise,and obtain accurate blood vessel enhancement results to segment the lung fissure efficiently.Based on the Fissureness filter-based lung fissure enhancement algorithm,the Hessian tubular reinforcement of the three-dimensional image is used to distinguish the tubular structure from other structures(disc,spherical,etc.).The main idea is to use the relationship between the eigenvalues of the Hessian matrix and the geometric structure to define the description operators of different structures to distinguish the vascular structure and the tracheal structure,so as to effectively enhance the lung fissure and suppress the noise.After the pulmonary fissure is enhanced,the lung fissure is post-treated.Through the comparison of the experimental effects of various post-processing methods,this paper adopts the skeleton-based post-processing method of lung fissure image.For the three-dimensional lung fissure image,this paper improves the skeletonization method and adopts the skeletonization method based on the fast marching method(FMM).After the lung fissure image is skeletonized,the branch points are removed,the joint meta-analysis is performed,and finally the holes are filled,and finally the complete lung fissure is obtained.By analyzing the experimental results,the experimental comparison and evaluation of this paper demonstrates that the proposed algorithm has an ideal effect in segmentation of lung fissure.In order to verify the accuracy of the skeletonized lung splitting algorithm in this paper,it is compared with the artificial gold standard.The experimental results show that the segmentation algorithm based on the fast marching method is better,and the segmentation accuracy of the lung fissure image is effectively improved.