Pulmonary nodule benign and malignant classification method and system based on multi-scale transfer learning

Abstract

The invention discloses a pulmonary nodule benign and malignant classification method based on multi-scale transfer learning, and the method comprises the following steps: S1, carrying out the multi-scale sampling of nodules in a lung CT image, and obtaining a multi-scale region of interest; S2, preprocessing the obtained multi-scale region of interest to synthesize a three-channel RGB image; S3,preliminarily constructing a transfer learning network model; and S4, training the preliminarily constructed transfer learning network model by using the synthesized RGB image to obtain a model capable of performing benign and malignant classification on pulmonary nodules. According to the pulmonary nodule benign and malignant classification model construction method based on transfer learning, benign and malignant judgment can be carried out by fully utilizing the imaging characteristics of the interior and exterior of the pulmonary nodule under different scales, only the approximate positionof the nodule needs to be provided, the contour information of the nodule does not need to be used during classification, and the segmentation step of a nodule region is avoided, so that the automation degree is higher, and the practicability is higher.

Description

technical field [0001] The invention relates to a method and system for constructing a benign and malignant pulmonary nodule classification model. Background technique [0002] Currently, cancer has become one of the most threatening diseases to human life and health. According to the statistics of the American Cancer Society in 2018, there will be 18.1 million new cancer cases in the world in 2018, while the number of deaths due to cancer is expected to be 9.6 million; among all cancers, the number and death rate of lung cancer patients rank first. Lung cancer is a disease with a poor prognosis, with an average five-year survival rate of less than 20%. When lung cancer enters locally advanced stage, or there is an unresectable part, patients can only receive radiotherapy and chemotherapy, and their average survival time is less than 12 months. It can be seen that the early diagnosis of lung cancer is of great clinical significance. [0003] Pulmonary nodules (Pulmonary N...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is to provide a method and system for classifying benign and malignant pulmonary nodules based on multi-scale transfer learning to solve the problem of deep learning technology in medical image recognition under the current situation of scarcity of high-quality data with labels. overfitting problem

Images