The ROIs Segmentation Of The Lungs And The Detection Of Lung Nodules

Lung cancer is still one of the cancers threating human life and health, so earlydetection of lung cancer will help improve the survival rate of patients with lung cancer.Relying solely on the subjective diagnosis of doctors on large number of medical CTimages, the fatigue or other factors often cause missed diagnosis. Automatic detectionprograms of lung nodules can help doctors improve efficiency and reduce the intensityof work. Accurate segmentation of lung nodules is demanded to accurately detectnodules, reduce false positive rate of identification and classification and evaluate theclinical treatment effect. Based on application background of Computer-aided diagnosisand treatment evaluation system, this paper studies the segmentation of the ROIs in thelung CT images and detection of suspected nodules.First, the DICOM files of CT images are covered to facilitate computer processing.For the preprocessed images, region growing algorithm based on Otsu is applied toobtain lung parenchyma, and the SUSAN (Smallest UnivalueSegment AssimilatingNucleus) operator is used to compensate lung parenchyma. Then according to the imageintensity distribution, apply Gaussian mixture model (GMM) and Bayesian classifiermethod to realize binary segmentation of lung parenchyma images to obtain the initialROIs. Also, dividing the image into the same size patch, employ iteration algorithmwith adaptive threshold value to initialize the parameters of GMM and apply EMalgorithm to estimate GMM optimal parameters.Second, due to the gray difference between background and target objects, weproposed edge expansion algorithm based on gradient information for accuratesegmentation of ROIs according to the initial shape of ROIs. This paper exploits Sobeloperator combined with the regional boundary gradient to detect and locate edge untilobtain the approximate real boundary of ROI.In this paper, the proposed method is tested in the public LIDC (Lung ImageDatabase Consortium) database and data set of the Second Affiliated Hospital of HarbinMedical University. We evaluate objectively segmentation effect based on lung noduleslabeled by experts. Experimental results show that the proposed approach is effectivefor precise segmentation of ROIs and detection of suspected lung nodules, which is veryhelpful to identify nodules and reduce the false positives.