Research Of Lung Nodule Segmentation Based On Lung Thin CT Image Sequences

In the early diagnosis of lung cancer,the most critical indicator is the early form of lung cancer-the detection rate of lung nodules.However,the use of ultra-finely spaced scans,ie,"thin scans," on CT scans can more accurately detect the key lesions in the early stages of lung cancer.Therefore,in order to improve the detection rate of lung nodules,it is necessary to use thin-section CT scan to accurately display the microscopic tissue of human body and the structural images of various primary lesions.However,there are two major problems in the diagnosis of lung cancer in the thin-scan mode of low-dose CT.First,there is a huge contradiction between the massive CT image data and the serious shortage of artificial diagnosis.Faced with large-scale mass CT image data,ideally,a large number of experienced and stable physicians need to complete the diagnosis without any fatigue.Otherwise,misdiagnosis and missed diagnosis will inevitably occur.Second,the conflict between the image quality of CT image data caused by thin-section scans and the increased detection rate of lung lesions.The use of low-dose CT thin-section scanning technology often leads to problems such as more image noise,inconspicuous boundary feature information,and lower image quality of image fogging,which affects the segmentation and diagnosis of early lesion areas.In addition,in the process of lung diagnosis,there are a wide variety of lung diseases and imaging features are complex,making it difficult to segment lung lesions under large-scale image data.In particular,the diagnosis of juxta-vascular nodules and cavitary nodules is highly likely to be malignant.If these difficultly segmented nodule images can be accurately segmented at an early stage,the current high lung cancer deaths will be effectively reduced.The fast and accurate segmentation of lung nodule image sequences is the basis of subsequent processing and diagnostic analyses.However,previous research investigating nodule segmentation algorithms cannot entirely segment cavitary nodules,and the segmentation of juxta-vascular nodules is inaccurate and inefficient.To solve these problems,we propose a new method for the segmentation of lung nodule image sequences based on superpixels and density-based spatial clustering of applications with noise(DBSCAN).First,our method uses three-dimensional computed tomography image features of the average intensity projection combined with multi-scale dot enhancement for preprocessing.Hexagonal clustering and morphological optimized sequential linear iterative clustering(HMSLIC)for sequence image oversegmentation is then proposed to obtain superpixel blocks.The adaptive weight coefficient is then constructed to calculate the distance required between superpixels to achieve precise lung nodules positioning and to obtain the subsequent clustering starting block.Moreover,by fitting the distance and detecting the change in slope,an accurate clustering threshold is obtained.Thereafter,a fast DBSCAN superpixel sequence clustering algorithm,which is optimized by the strategy of only clustering the lung nodules and adaptive threshold,is then used to obtain lung nodule mask sequences.Finally,the lung nodule image sequences are obtained.The experimental results show that our method rapidly,completely and accurately segments various types of lung nodule image sequences.To address the problem of existing segmentation algorithms cannot accurately segment juxta-vascular nodules and have a slow efficiency,a sequence segmentation method based on superpixels and sparse sub-space clustering is presented.Firstly,the lung parenchyma images sequences of the CT images are segmented.Then the regions of interest(ROIs)are obtained.The ROIs image sequences,is then segmented by the improved superpixel sequence segmentation method.Moreover,the new features of all the superpixel samples are extracted,including contrast enhancement histogram features,the texture features of the neighborhood of the superpixel samples and the location information features based on prior knowledge.The distance-constrained sparse sub-space clustering algorithm,is then utilized to perform superpixel samples clustering to obtain lung nodule mask sequences.Finally,the lung nodule image sequences are obtained.The experimental results show that the presented method can accurately and efficiently segment the juxta-vascular nodule image sequences.