Research On Detection And Diagnosis Methods For Pulmonary Nodules Based On Restricted Boltzmann Machine

Lung cancer is one of the biggest malignant tumors that threaten human health and life in the world.The application of cutting-edge technologies such as big data and deep learning has become a trend in the medical field.Many deep learning algorithms have been used in the early screening and diagnosis of lung cancer.Not only it can save the lives of patients,but it can also ease medical resources.In order to assist radiation doctors to efficiently improve the detection rate of early-stage lung cancer and to diagnose pulmonary benign nodules with high accuracy,the advent of computer-aided detection systems(CADe)and computer-aided diagnosis systems(CADx)has brought patients and physicians new opportunities and challenges,the use of computer-aided diagnostic techniques to help radiologists read and understand medical images,reduce the workload of doctors to process data,improve the diagnostic efficiency and accuracy.This article has carried on the detailed research and design to the above two systems,the concrete work is as follows:(1)For the traditional computer-assisted detection system for lung cancer,artificially designed features are used.Not only is the procedure complex and accompanied by a large number of false-positive data,this article introduces the deep belief network into the lung cancer detection model.In this paper proposed a method for pulmonary nodules detection based on multi-view deep belief network.Because pulmonary nodules are a three-dimensional sphere in space.Firstly,the three-dimensional reconstructed pulmonary nodules are resized to different sizes of cube,and 2.5D slices of different angles are used as the input data of the deep belief network.Finally,the fusion strategy is used to identify the pulmonary nodules.Compared with traditional CAD system,the sensitivity of this method is 92.8±0.25% and the 2.4±0.3 false positives per scan.The large number of experiments on dataset shows that this method can effectively reduce the false positive rate of pulmonary nodule detection.(2)In order to solve the complex problem of lung nodule feature extraction in traditional computer-aided diagnosis system,this article introduces the deep belief network into lung cancer diagnosis model and a method for classifying benign and malignant pulmonary nodules based on deep belief network is proposed.For the characteristics of the data set,first,the original CT image was preprocessed using the threshold probability maps method.Next,the deep feature of the pulmonary nodules were extracted using the multi-hidden deep belief network and form feature vectors.Finally,an extreme learning machine was used as the classifier for benign and malignant classification.The experiment discussed and designed the structural parameters of a custom deep belief network in detail.In addition,comparing the advantages and disadvantages of other deep learning methods in the diagnosis of pulmonary nodules,it is proved that this method is superior to other algorithms in terms of accuracy,specificity and sensitivity.