Classification method of bacterial and viral pneumonia in children based on deep learning algorithm

Abstract

The present invention provides a classification method for bacterial and viral pneumonia in children based on a deep learning algorithm. The method first manually labels the source data set, and then, on the basis of full convolutional network semantic segmentation and convolutional neural network algorithm, First, the full convolutional network semantic segmentation algorithm is used to foreground the image to segment the lung region to obtain the region of interest, and the extracted region of interest is input into the convolutional neural network model to train the classifier, so as to predict the category of the unknown chest X-ray image Extract the high-dimensional features of the region of interest, and at the same time use traditional image processing methods to extract the low-dimensional features of the region of interest, respectively use the high-dimensional and low-dimensional features to train nonlinear classifiers, and predict the category of unknown X-ray images, thereby Determine the type of pneumonia the patient is suffering from. The principal component analysis algorithm is used to reduce the dimensionality of the features to reduce the amount of calculation, and then the mixed dimensionality-reduced features are input into the nonlinear classifier to predict the category of the unknown X-ray image.

Description

technical field [0001] The present invention relates to the fields of computer vision technology and medical image processing, and more specifically, to a method for classifying bacterial and viral pneumonia in children based on deep learning algorithms. Background technique [0002] Pneumonia is a common and frequently-occurring disease in children, and it is also the leading cause of death in children. Pneumonia in children is mostly caused by bacteria and viruses, and a few are caused by mycoplasma and fungi. Pathogen diagnosis is an important basis for clinically correct antibiotic selection. Chest X-ray is one of the most common aids in the diagnosis of pneumonia. With the rapid development of the computer field and the success of deep learning algorithms, it has gradually set off an upsurge in the field of computer-aided diagnosis, and a large number of image processing algorithms and disease predictions based on X-ray images have emerged. Classification algorithm, th...

AI Technical Summary

Problems solved by technology

However, there are still serious noises in chest X-ray images, such as occlusion of human bones, interference of bronchi and blood vessels, human organs, etc., and images vary with individual differences

And there is currently no relevant technology that uses deep learning methods to judge whether children's pneumonia is infected by bacteria or viruses from images

Images