| In recent years,with the continuous improvement of people’s living standards,lung diseases have become one of the biggest threats to human health.Early diagnosis and treatment of lung diseases have become an important means to alleviate this problem.The development and advancement of medical image processing and the diversification of medical imaging equipment have caused the amount of(Computed Tomography CT)image data to increase exponentially.At present,it is common to rely on the subjective judgments of doctors and other professionals to interpret CT images.Because massive amounts of data require a lot of time to process,work fatigue or high-pressure work often occurs,which leads to errors in judgment,misdiagnosis,and missed diagnosis,which make patients Missing the best period of treatment or misusing the treatment plan.In this case,we urgently need to conduct in-depth research on(computer-aided diagnosis CAD),realize preliminary analysis of medical images,and explore high-accuracy and high-speed automatic lung parenchymal segmentation methods based on CT images.By assisting judgment,the accuracy and efficiency of doctors and other professionals’ judgments can be improved,and the work pressure and work intensity of doctors can be reduced,and more accurate and comprehensive information can be provided for the subsequent further processing and judgment of lung areas.In CT images,how to extract lung regions quickly and accurately still has the problem of segmentation accuracy.At present,in the field of medical imaging,many researchers have proposed detection methods for lung parenchymal extraction.However,combined with the current research methods proposed,due to the high variability of the appearance of lung tissue,a reliable solution cannot be provided.Classification methods based on local descriptors can achieve good results in image local segmentation,but this method is usually computationally expensive,which limits its wide use in real-time or near-real-time clinical applications.In view of the above situation,this paper combines the region-based segmentation method and optimized sampling grid for the classification of local descriptors,and proposes a fast,accurate,and reliable method for segmenting lung parenchymal regions from CT scan images.The method in this paper is mainly carried out in two stages;in the first stage,the target area is located based on the threshold segmentation combined with the connected region growth method,and then combined with the mathematical morphology calculation to separate the left and right lung regions,and finally the lung parenchymal boundary is smoothed to realize the lung parenchyma The area is initially divided.In the second stage,based on the lung parenchyma area initially extracted in the first stage,the local area information is calculated based on(fuzzy connectivity)FC method,and the local area information is calculated by approaching the best key points of the supervoxels of the sampling grid.Descriptors,so as to achieve local accurate segmentation of lung parenchymal regions,improve segmentation accuracy,and achieve better results.The method in this paper evaluates and analyzes the segmentation results of lung parenchymal region on a data set of 20 patients(512dpi*512dpi*300 sheets/case),which verifies that the proposed method is robust and robust.Comparing the results of lung parenchymal segmentation algorithm,qualitative and quantitative data comparison analysis,comprehensive evaluation of various indicators,prove that the method proposed in this paper is more efficient,accurate and universal for lung parenchymal region segmentation in CT images.Therefore,the comprehensive performance of our proposed method is better and it is of great significance for computer-aided medical diagnosis. |
PDF Full Text Request