Pulmonary nodule detection method and system based on image enhancement

Abstract

The invention provides a pulmonary nodule detection method and a system based on image enhancement. The method comprises steps: an anteroposterior view chest radiograph after pre-processing is dividedto a training set, a verification set and a test set according to a preset proportion; pulmonary nodule areas cut out from the anteroposterior view chest radiograph in the training set and the verification set are used as negative samples for model training, other areas in the anteroposterior view chest radiograph are randomly cut as positive samples for model training, and the negative samples are subjected to deformation processing to increase the number of negative samples; and a sliding window is used to divide the whole JPG anteroposterior view chest radiograph to non-overlapped image blocks, and the image blocks are inputted to a convolution neural network model to judge whether the image blocks are positive samples or negative samples. The pulmonary nodule detection efficiency canbe improved, the false positive rate is reduced, and through carrying out deformation processing on the negative samples, overfitting of the model due to insufficient chest radiograph data can be overcome.

Description

technical field [0001] The invention relates to the field of pulmonary nodule detection, in particular to an image enhancement-based pulmonary nodule detection method and system. Background technique [0002] Early lung cancer appears on medical imaging as small shadows called lung nodules. So far, many tools for detecting pulmonary nodules have appeared at home and abroad. Traditional X-ray is a simple, cheap and widely used method of chest radiograph detection. However, its ability to detect early lung cancer is limited by technology and physician level and the recognition rate cannot be well guaranteed. The most influential and the highest recognition rate is the multi-slice spiral CT technology, but its high radiation hazards also restrict the use of this technology. With the development of medical imaging technology, the resolution and contrast of chest X-rays have been greatly improved. Many hospitals still use chest X-rays for a preliminary diagnosis, and the radia...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the deficiencies of the prior art, the present invention provides a method and system for detecting pulmonary nodules based on image enhancement, which solves the defect of low detection efficiency of pulmonary nodules

Images