Research On Lung Airway Tree Segmentation Algorithm Based On T-prim Model

With the aggravation of environmental pollution and the tension of modern life style,human beings are facing more and more severe challenges of medical health problems.The threat of lung diseases to human beings has become a problem that can not be ignored.Computer-based medical imaging technology has made great contributions to the visualization and quantification of medical diagnosis.Segmenting lung tracheobronchial tree from CT data plays a very important role in the diagnosis and treatment of lung diseases.Therefore,it is of great significance to segment lung tracheobronchial tree accurately and quickly from CT image sequence,which is an essential part of computer aided diagnosis(CAD)system and has very important research value.The difficulty of lung tracheobronchial tree segmentation lies in the segmentation of high-order bronchioles.The structure of lung tracheobronchial tree itself is complex.Noise generated in imaging and pathological changes of lung and trachea interfere greatly with segmentation.Therefore,this paper proposes a lung tracheobronchial tree segmentation method based on T-prim tree and optimal watershed segmentation framework,which can quickly and accurately segment the complete lung tracheobronchial tree,and there is almost no leakage and error segmentation in the whole process.The innovative performance of the algorithm in this paper is as follows:(1)Aiming at the artifact and noise of lung CT images,a region growing algorithm based on simple surface fitting is used,which greatly optimizes the traversal process of the algorithm,improves the robustness of the algorithm and greatly optimizes the traversal process.(2)The segmentation framework of watershed algorithm is optimized,and the optimized segmentation framework is further modified from the perspective of Markov random field,which can be used to segment specific classes from multiple object backgrounds(3)To improve the prim maximum spanning tree algorithm,a T-prim model is proposed.According to the optimal energy function obtained from the above optimization framework,the reliability evaluation criteria for evaluating the bronchial gray data are constructed,and the results are used as thresholds to implement MST algorithm to obtain the complete structure of the pulmonary trachea tree.(4)The algorithm is extended on the platform of MITK to realize the segmentation of the algorithm and the three-dimensional visualization of the segmentation results.