Medical image pulmonary nodule detection based on depth learning

Abstract

The present invention discloses a medical image pulmonary nodule detection based on depth learning, in order to solve the problem of low detection accuracy due to the insufficient data amount of effective medical images extracted in the prior art. The implementation scheme comprises: 1) obtaining a medical image; 2) introducing Gaussian noise in the medical image to expand a data sample set; 3) constructing a new feature extraction network; 4) using the new feature extraction network, and combining with the existing regional recommendation network and the classification network to obtain a detection model; 5) using the expanded data sample set to train the detection model; and 6) using the trained detection model to carry out pulmonary nodule detection. According to the method disclosed by the present invention, the new feature extraction network is established, so that the degree of over-fitting of the network is reduced, medical image pulmonary nodule detection precision is improved, and the method can be applied to the computer-aided diagnosis system.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to a method for detecting pulmonary nodules in medical images. It can be used in computer-aided diagnosis system. Background technique [0002] Pulmonary nodules are one of the most important early signs of lung cancer. According to the lesion characteristics of pulmonary nodules, the lesion characteristics of lung lesions can be inferred. Therefore, early detection and treatment of pulmonary nodules in patients with lung diseases is a key measure to reduce lung cancer mortality. Due to its high morbidity and mortality, lung cancer has become the deadliest cancer among cancers. With the change of people's living habits and the deteriorating environment, the number of people with lung cancer is increasing, and the society is paying more and more attention to it. Combined with the medical characteristics of pulmonary nodules, using deep learning technology to process and stud...

AI Technical Summary

Problems solved by technology

The common theoretical basis of these methods is traditional statistics, which belong to shallow structural models. They usually require strong prior knowledge or different feature attempts and parameter selection to obtain satisfactory features, which brings complexity to the entire classification problem. As a result, the detection speed of pulmonary nodules in existing medical images is slow and the accuracy is low

Images