Research And Implementation Of Segmentation Algorithm For Multi Type Pulmonary Nodules Based On CT Image

Lung cancer is a common cancer disease,and its mortality rate ranks first among all cancer diseases.As far as the current medical technology,the key of lung cancer treatment lies in early detection and early treatment.The early stage of lung cancer is in the form of pulmonary nodules.Therefore,the key point of diagnosis is whether the pulmonary nodules can be found in the lung imaging in time and whether there is a tendency of deterioration through medical imaging technology.At present,chest CT image and biopsy are still used as the main methods of pulmonary nodule diagnosis.The traditional diagnosis method has a long diagnosis cycle,which not only takes time and efforts of doctors,but also delays the treatment time of patients.To this end,a computer-aided diagnosis(CAD)system was developed,which was expected to help doctors diagnose diseases through computer technology,improve the efficiency and accuracy of diagnosis,and provide better treatment for patients.However,pulmonary nodules are complex in morphology,heterogeneous in structure,fuzzy in edge,uneven in density distribution,sensitive to noise and are easily confused with other surrounding tissues,which make it difficult for CAD system to take into account the diagnosis of different types of pulmonary nodules.Therefore,the lung nodule segmentation in CAD system was conducted,and a multi-type lung nodule segmentation algorithm based on CT image was proposed.On the basis of ensuring the accuracy,it realizes the effective segmentation of different types of lung nodule,so as to improve the performance of CAD system in multi-type lung nodule diagnosis.The main work of this paper is as follows:(1)Otsu lung parenchyma segmentation algorithm based on optimal symmetric particle swarm optimization is proposed.Firstly,the lung CT image is preprocessed by bilateral filtering and top-hat cap transform enhancement algorithm.Then,the improved particle swarm optimization algorithm is introduced based on Otsu algorithm to find the best threshold with its optimization ability,so as to improve the segmentation efficiency of the algorithm.(2)An improved adaptive edge matching algorithm is proposed.This algorithm is mainly used to solve the problem of lung edge depression caused by juxta-pleural nodules pulmonary nodules.Firstly,the edge of the lung is obtained by edge tracking algorithm,and then the edge matching algorithm is improved,the fixed threshold is changed to adaptive threshold,and the adaptive step size is redefined.The improved adaptive edge matching algorithm is used to repair the edge of the lung to obtain a complete lung area,and at the same time,the segmentation problem of juxta-pleural nodules pulmonary nodules is solved.(3)The WFCM algorithm based on local spatial correlation is proposed,which combined with the concept of local image feature correlation and the spatial local correlation,so that it can take into account the spatial information and gray information of the image at the same time,and improve the accuracy and anti-noise of the algorithm segmentation.What’s more,the fuzzy clustering idea is used to solve the edge fuzzy problem of pulmonary nodules and the segmentation challenge of ground glass nodules,which could improve the over segmentation or under segmentation of pulmonary nodules.