Research on the Automatic Detection Method of Pulmonary Nodules Based on Deep Learning

Abstract

Lung cancer has become the highest incidence and mortality of malignant tumors in China and even the world. It is truly the “first killer of cancer”. Pulmonary nodules are the early manifestation of lung cancer. Early detection, early treatment and early diagnosis of pulmonary nodules can greatly reduce the mortality of lung cancer. Due to their different sizes and complex structures, it is difficult for doctors to recognize the specific characteristics of pulmonary nodules. In order to help doctors detect and analyse pulmonary nodules faster and more accurately, based on deep learning, this paper proposed a new method to detect pulmonary nodules by cascades of improved VGG network model and improved Resnet network model. The improved network was evaluated on the LIDC-IDRI database, and the experimental results demonstrate the effectiveness and robustness of the model, which can help doctors accurately and effectively detect pulmonary nodules.