Segmentation Method Of Risk Organ Lung With Radiotherapy Lung Cancer Complete Sequence CT Images

Lung cancer radiotherapy planning plays an important role in the treatment of lung cancer.With the development of medical technology,there are higher requirements for improving the cure rate of lung cancer and reducing the radioactive injury of organs at risk.Therefore,it is very helpful for accurate segmentation of organ lung at risk to increase the dose of tumor target and reduce the dose of normal lung tissue,which is important for the treatment of clinical lung cancer.The main contents of this dissertation include:(1)The technology of medical window adjustment is investigated.Pretreatment of lung CT image is studied.And mathematical expression of CT image is abstracted by using the matrix.The lung cancer image is simplified and depicted in detail through the centroid method and wavelet unsharp masking method after window adjustment.A general method of lung cancer image model pretreatment is put forward,which is to solve the poor relevance of lung CT image segmentation method with medical knowledge and its low rate of clinical application.(2)Segmentation method of organ lung at risk in pulmonary window is proposed.Aiming at the poor edge effect of traditional algorithm segmentation,a hybrid algorithm based on the three-dimensional Otsu segmentation and Kirsch operator is proposed.According to the clinical characteristics of lung cancer CT images,the one dimensional matrix is transformed to co-occurrence matrix consisting of edge information to increase the weight of edge information in threshold segmentation.The accuracy of the algorithm is verified by segmentation of CT images of organ lung at risk.(3)Segmentation method of organ lung at risk in mediastinal window is presented.For the poor effect of segmentation for CT images of lung at risk in pulmonary window,an adaptive region growing method is proposed based on the mediastinal window.With the seed point coordinates determined automatically through failed segmented CT images in pulmonary window,the growth criterion is established by using regional consistency test.The reliability of the segmentation algorithm is verified through the twice segmentation of CT images,which are not well segmented in pulmonary window.(4)The refining of segmented images of organ lung at risk is studied.The segmented masking images of organ at risk are processed by morphology and rolling method to solve the problem of burr and hole which are liable to occur in segmentation.And the complete sequence of refined images is compared with that manual segmentation of clinicians.Results confirmed that the algorithm can quickly and efficiently segment the images of organ lung at risk,which is beneficial to clinicians and physicists for radiotherapy planning.