Multi-class pneumonia screening deep learning device based on CT images

Abstract

The invention provides a multi-class pneumonia screening deep learning system based on CT images. The method comprises the following steps of a CT image data preprocessing part preprocesses a CT imageto obtain a preprocessed CT image; the slice-level pneumonia dichotomy part analyzes the basic depth features to obtain a dichotomy result, the weakly supervised lesion positioning part obtains a lesion position map according to the class labels, and the slice-level pneumonia four-class part detects four classes of pneumonia to obtain a four-class pneumonia classification task result; the case-level pneumonia classification part performs analysis to obtain a multi-class pneumonia classification task result; and the pneumonia diagnosis comprehensive evaluation part outputs a pneumonia diagnosis comprehensive evaluation result and a focus positioning distribution map based on the slice-level four-classification result, the focus position map and the case-level multi-classification pneumoniaclassification task result. The method is advantaged in that the COVID-19 and other pneumonia diseases can be quickly and accurately distinguished on the premise of no medical staff besides obtainingthe common pneumonia diseases, which is helpful for screening the epidemic situation of the COVID-19; the COVID-19 and other pneumonia diseases can be quickly and accurately distinguished.

Description

technical field [0001] The invention relates to a deep learning device for multi-category pneumonia screening based on CT images, which belongs to the cross field of medical images and computer image recognition. Background technique [0002] Coronavirus disease (COVID-19) is caused by a new type of coronavirus (SARS-CoV-2, referred to as 2019-ncov), which is highly contagious and can cause acute respiratory distress or multiple organ failure in severe cases. One hundred thousand cases. Therefore, how to diagnose COVID-19 accurately and efficiently is crucial not only for the timely treatment of patients, but also for the allocation and management of hospital resources during the epidemic. [0003] Currently, the standard diagnostic method for 2019-nCoV is real-time polymerase chain reaction (RT-PCR), which detects viral nucleosides in specimens obtained from oropharyngeal swabs, nasopharyngeal swabs, bronchoalveolar lavage, or tracheal aspiration acid. The sensitivities ...

AI Technical Summary

Problems solved by technology

But as a priori studies, they also show some limitations: First, learning segmentation tasks using pixel-level annotations[1] is impractical in practice, especially in burst In the case of infectious diseases; secondly, since a patient’s CT examination usually contains hundreds of slices, the diagnosis or risk assessment of each slice’s prediction results [2] makes clinicians’ The application value is very limited. Although the currently used 3D convolutional neural network (3D CNNs) [3] is an option to solve these disadvantages, but its high requirements for hardware (such as GPU) , a large amount of computing cost and training time, making it inflexible and universal in practical applications

Images