Research On CT Images Segmentation For Extracting Lung Parenchyma And Hepatic Vessels

CT imaging technology is an important means for radiologists to carry out clinical medical image assisted diagnosis,which has the advantages of high contrast and resolution.Due to the interference of tissue lesions,bronchus and abnormalities near pleura,it is necessary to accurately segment the relevant anatomical structure when carrying out thoracoabdominal examination.Aiming at the segmentation of pulmonary parenchyma and hepatic blood vessels in CT images,this paper proposes improved algorithms to improve the accuracy of detection and segmentation.The main work is as follows:(1)A 3D lung parenchymal segmentation algorithm is presented based on the integration of the Surfacelet transform and the pulse coupled neural network(PCNN).First,the 3D CT lung volume is decomposed into Surfacelet transform domain to obtain multi-scale and multi-directional sub-band information.The edge features are then enhanced by filtering sub-band coefficients using local modified Laplacian operator.Second,Surfacelet inverse transform is implemented and the reconstructed image is feedback to the input of the PCNN.Finally,iteration process of the PCNN is carried out to obtain final segmentation result.The proposed algorithm is validated on the samples of two public dataset,namely,Kaggle and LIDC.The experimental results demonstrate that the proposed algorithm can get more complete lung parenchyma and the average Dice similarity coefficient,the average over-segmentation rate,and the average under-segmentation rate can achieve 97.85%,0.31%and 1.87%respectively.(2)This paper proposes an automatic segmentation method of 3D vessel CT images to obtain better segmentation results.First,the single Gaussian kernel of Hessian matrix in the Jerman’s algorithm is replaced by bi-Gaussian kernel.Then,a histogram-based method is adopted to adaptively estimate the threshold value of the region growing as the growth criterion.Finally,the region growing algorithm is used to extract the hepatic vessels in the enhanced CT images.The experimental results show that the proposed method achieves a significant enhancement of hepatic vessels segmentation with an average accuracy 98.1%.