Lung Parenchyma Segmentation And Nodule Detection Based On Thin-section CT Scans

Solitary pulmonary nodule is an early form of most lung cancer.In the early diagnosis of lung cancer,the detection of pulmonary nodules is very important.However,as the accuracy requirements for clinical imaging of lesions increase,the CT scanning thickness decreases,which also calls ‘Thin-scanning’,and a large number of CT image sequences need to be produced.The massive amount of image data will be inevitably increase the workload of doctors and lead to misdiagnosis and missed diagnosis.Computer aided detection(CAD)using image processing technology and medical imaging method in the diagnosis of lesions aided detection in medical images,compared with manual reading,image can be objectively analysis and can effectively achieve the detection of pulmonary nodules.It will not only reduce the workload of doctors to improve the efficiency of diagnosis,to improve the survival of patients the rate of lung cancer has important significance.Therefore,the research of computer-aided detection of lung nodules is a hot spot in medical imaging research.Lung parenchyma segmentation is often performed as an important preprocessing step in the computer-aided diagnosis of lung nodules based on CT image sequences.However,existing lung parenchyma image segmentation methods cannot fully segment all lung parenchyma images and have a slow processing speed,particularly for images in the top and bottom of the lung and the images that contain lung nodules.We proposed a segmentation method for lung parenchyma image sequences based on superpixels and a self-generating neural forest.Our proposed method first uses the position of the lung parenchyma image features to obtain lung parenchyma ROI image sequences.A gradient and sequential linear iterative clustering algorithm(GSLIC)for sequence image segmentation is then proposed to segment the ROI image sequences and obtain superpixel samples.The SGNF,which is optimized by a genetic algorithm(GA),is then utilized for superpixel clustering.Finally,the grey and geometric features of the superpixel samples are used to identify and segment all of the lung parenchyma image sequences.The experimental results shows that our proposed method achieves higher segmentation precision and greater accuracy in less time.In the two-dimensional CT images,blood vessels and nodules are oval shape,and the density and CT value properties are very similar.There often have very high false positives in the detection of pulmonary nodules results,affecting the accuracy of detection of pulmonary nodules.For this reason,this paper proposed an automatic detection method for lung nodules based on multi-scale enhancement filter and 3D shape feature.First constructed the ideal model of pulmonary nodules and blood vessels,and then construct two kinds of multi scale 3D enhancement filter based on shape by using Hessian matrix,which are used to enhance the pulmonary nodule image and vascular image.Suspected nodule images will be obtained by removing most of the image in the image of vascular nodules.After that,a new feature descriptor of pulmonary nodules,surface normal orientation angles histogram(SNOAH)was proposed.Classification and feature extraction of nodules and suspected pulmonary nodules with SVM classifier.The experimental results show that the proposed method can effectively reduce the false positive results and improve the accuracy of detection of pulmonary nodules.The results show that the feature descriptor is effective and it can be used as the basis for the differentiation of nodules and blood vessels.