An intelligent diagnosis system for COVID-19 based on deep learning

Abstract

The invention relates to an intelligent diagnosis system for novel coronavirus pneumonia based on deep learning, which belongs to the field of medical image processing and includes a control unit, an intelligent detection and diagnosis unit, a storage unit and a three-dimensional display unit. The control unit is used for inputting and modifying system data; intelligent detection The and diagnosis unit is used to predict new coronary pneumonia cases; the intelligent detection and diagnosis unit includes a data preprocessing module, a lesion area detection module, a false positive removal module, and a case prediction module. The invention adopts a deep learning network diagnosis system that combines detection, false positive removal and prediction to output diagnosis results, and solves the problems of too small lesions and difficult detection, too little information of a single local lesion, and too high misdiagnosis rate, etc., and solves medical problems. The problem of too small sample size greatly improves the diagnostic efficiency, helps to improve the treatment efficiency of patients and accumulate clinical data.

Description

technical field [0001] The invention belongs to the field of medical image processing, and relates to an intelligent diagnosis system for novel coronavirus pneumonia based on deep learning. Background technique [0002] The global nature of the novel coronavirus pneumonia (COVID-19) infectious disease has quickly attracted widespread attention from medical experts due to its strong infectivity and high morbidity rate, and accurate diagnosis of patients is helpful for disease control and recovery. At present, there are two main methods for detecting new coronary pneumonia: nucleic acid detection and CT detection. However, due to the shortage of reagents and the relatively high false negative and misdiagnosis rate in nucleic acid detection, CT detection is usually used as the main diagnostic basis. Usually, the CT of a case contains 200-500 tomographic images. The traditional way of doctor reading is often time-consuming and labor-intensive, and some lesions are small in size ...

AI Technical Summary

Problems solved by technology

At present, there are two main methods for detecting new coronary pneumonia: nucleic acid detection and CT detection. However, due to the shortage of reagents and the relatively high rate of false negative and misdiagnosis in nucleic acid detection, CT detection is usually used as the main diagnostic basis.

Usually, the CT of a case contains 200-500 tomographic images, and the traditional way of reading images by doctors is often time-consuming and laborious, and some lesions are small in size and difficult to distinguish with the naked eye, resulting in missed diagnosis and misdiagnosis

Images