A pulmonary nodule analysis method and device

Abstract

The invention discloses a pulmonary nodule analysis method and device. The method comprises the following steps: acquiring a pulmonary nodule image of a lung of a patient and three-dimensional coordinates of pulmonary nodules in the pulmonary nodule image; according to the three-dimensional coordinates of the pulmonary nodules, determining ROIs containing the pulmonary nodules from the pulmonary nodule images, and inputting the ROI of the pulmonary nodules and the three-dimensional coordinates of the pulmonary nodules into a preset feature extraction neural network model for feature extractionto obtain feature vectors of the pulmonary nodule images, and inputting the feature vectors of the pulmonary nodule images into a preset classification neural network model to obtain a pulmonary nodule analysis result of the patient output by the preset classification neural network model. Compared with a traditional doctor diagnosis mode, the diagnosis error rate caused by the horizontal difference of doctors can be reduced, and therefore the accuracy of pulmonary nodule analysis is improved.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of machine learning, and in particular, to a method and device for analyzing pulmonary nodules. Background technique [0002] With the increasingly serious environmental pollution, more and more diseases show a trend of high outbreak rate. The development of modern medical technology has been very mature, and doctors can diagnose various diseases with medical knowledge and medical experience. That is to say, the diagnosis of diseases in the prior art mostly relies on the diagnosis of doctors. However, since the medical level in various regions is very inconsistent, and the personal experience level of doctors is also uneven, the traditional methods of diagnosing diseases by doctors are easily affected. The regional medical level and the image of the doctor's personal experience level lead to the problem of large diagnostic error. [0003] Taking pulmonary nodules as an example, doctor...

AI Technical Summary

Problems solved by technology

That is to say, most of the determination of diseases in the prior art depends on the diagnosis of doctors. However, because the medical level in various regions is very inconsistent, and the level of personal experience of doctors is also uneven, the traditional method of diagnosing diseases by doctors is vulnerable. Imaging of regional medical level and doctor's personal experience level leads to large diagnostic errors

Images