Computer-aided Analysis Of Pulmonary Function Based On Medical Image Segmentation Technology

With the development of computed tomography(CT)imaging technology and computer technology,computer-aided lung function assessment system based on medical images has played an important role in the clinical practice of lung function assessment.Traditional methods of pulmonary function assessment mainly rely on visual scale,symptoms and signs,and pulmonary function tests(PFT).However,this method is not reliable in judging the severity of lung scales,and it cannot reflect the local pulmonary function status.At present,the computer-aided pulmonary function assessment system based on medical images has become a research hotspot in the field of lung function assessment because of its simple operation,reliable results,and strong reproducibility.In view of the shortcomings of the traditional pulmonary function assessment,an in-depth study on the medical image segmentation techniques is conducted for the pulmonary function assessment.Then,a computer-assisted system of the quantitative analysis of pulmonary function is built to assist doctors in assessing patients’ pulmonary function.The main research contents of this paper are divided into four aspects:1.Pulmonary tissue segmentation method based on optimal regional growth,grayscale reconstruction and fractional differential.Pulmonary tissue segmentation mainly includes segmentation of lung parenchyma,lung trachea and pulmonary vascular.Firstly,a lung parenchymal segmentation scheme based on adaptive thresholding is proposed for the segmentation of lung parenchyma.Next,based on the structure of the tracheal tree anatomy,a lung trachea segmentation algorithm based on optimal threshold growth and morphological gradation reconstruction is proposed.Then,pulmonary blood vessels are segmented based on fractional differential algorithm and shape index.Finally,the results of crude segmentation of lung parenchyma are repaired and the left and right lungs are separated.2.The detection of pulmonary fissure based on improved Random Walk.Pulmonary fissure detection is the basis for the segmentation of the lobe,so it is necessary to detect the fissure first,and then use the shortest path as the standard to achieve segmentation of the lobe.As for the pulmonary fissure segmentation,improved multi-scale retinex algorithm based on bilateral filter is proposed for enhancing the pulmonary fissure.The bilateral filter and weight β are employed to estimate the illumination and improve the resolution of images.Then,a modified segmentation method of random walker algorithm is implemented for the segmentation of pulmonary fissure.To decrease the false rate of seeds,Candy edge detection is employed to detect the region of pulmonary fissure.The pixel location and gradient difference information between a pair of pixels are integrated to define a weight function.Eightneighborhood information is employed to measure the similarity between a pair of adjacent pixels.As a result,the segmentation accuracy is significantly enhanced.3.The 2D ROI and 3D VOI segmentation based on interactive method.In order to quantitatively analyze the lung function regarding local tissue,the VOI and ROI segmentation algorithms are studied,which can evaluate lung function of any two-dimensional or threedimensional lung tissue of interest and assist doctors to observe the distribution and severity of local lung lesions.4.From the abovementioned technical result,a quantitative computer-aided lung function assessment system is designed based on quantitative lung function measurements and medical image visualization techniques.The total lung volume,functional lung volume,non-functional lung volume,pixel index and pixel distribution histogram,mean lung density and standard deviation are used as quantitative indexes for evaluating lung function in this system,which can be achieved based on single-CT images and CT sequences.The pulmonary function of the whole lung,unilateral lung,and lobe of the image is evaluated.At the same time,quantitative analysis of local lung function is performed on the VOI and ROI regions.The clinical test cases demonstrate that the system can be used for the diagnosis of pulmonary diseases such as emphysema and chronic obstructive pulmonary disease,and can assist doctors in assessing patients’ pulmonary function.