CN112349425A - Novel artificial intelligent rapid screening system for coronavirus infection pneumonia - Google Patents
Novel artificial intelligent rapid screening system for coronavirus infection pneumonia Download PDFInfo
- Publication number
- CN112349425A CN112349425A CN202010083930.3A CN202010083930A CN112349425A CN 112349425 A CN112349425 A CN 112349425A CN 202010083930 A CN202010083930 A CN 202010083930A CN 112349425 A CN112349425 A CN 112349425A
- Authority
- CN
- China
- Prior art keywords
- coronavirus infection
- image
- infection pneumonia
- novel
- pneumonia
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 206010035664 Pneumonia Diseases 0.000 title claims abstract description 73
- 208000001528 Coronaviridae Infections Diseases 0.000 title claims abstract description 65
- 238000012216 screening Methods 0.000 title claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 37
- 238000013473 artificial intelligence Methods 0.000 claims abstract description 34
- 238000007781 pre-processing Methods 0.000 claims abstract description 24
- 238000000034 method Methods 0.000 claims abstract description 8
- 210000004072 lung Anatomy 0.000 claims description 21
- 238000012545 processing Methods 0.000 claims description 21
- 201000010099 disease Diseases 0.000 claims description 15
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 15
- 230000003902 lesion Effects 0.000 claims description 12
- 230000008859 change Effects 0.000 claims description 7
- 210000000621 bronchi Anatomy 0.000 claims description 6
- 230000000750 progressive effect Effects 0.000 claims description 6
- 208000002151 Pleural effusion Diseases 0.000 claims description 5
- 238000002591 computed tomography Methods 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 239000005338 frosted glass Substances 0.000 claims description 4
- 210000004224 pleura Anatomy 0.000 claims description 4
- 238000012795 verification Methods 0.000 claims description 4
- 208000003898 Mediastinal Emphysema Diseases 0.000 claims description 3
- 206010050184 Pneumomediastinum Diseases 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000007797 corrosion Effects 0.000 claims description 3
- 238000005260 corrosion Methods 0.000 claims description 3
- 210000001370 mediastinum Anatomy 0.000 claims description 3
- 238000012512 characterization method Methods 0.000 claims description 2
- 239000005337 ground glass Substances 0.000 claims description 2
- 238000000605 extraction Methods 0.000 claims 1
- 239000003550 marker Substances 0.000 claims 1
- 238000011017 operating method Methods 0.000 claims 1
- 238000005516 engineering process Methods 0.000 abstract description 6
- 238000013170 computed tomography imaging Methods 0.000 abstract description 4
- 241000711573 Coronaviridae Species 0.000 description 11
- 238000001514 detection method Methods 0.000 description 9
- 108020004707 nucleic acids Proteins 0.000 description 8
- 150000007523 nucleic acids Chemical class 0.000 description 8
- 102000039446 nucleic acids Human genes 0.000 description 8
- 210000000038 chest Anatomy 0.000 description 6
- 238000003745 diagnosis Methods 0.000 description 6
- 241000700605 Viruses Species 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 239000007787 solid Substances 0.000 description 3
- 230000009385 viral infection Effects 0.000 description 3
- 208000036142 Viral infection Diseases 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000005034 decoration Methods 0.000 description 2
- 239000011521 glass Substances 0.000 description 2
- 238000002955 isolation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 201000001178 Bacterial Pneumonia Diseases 0.000 description 1
- 241000112286 Bat SARS-like coronavirus Species 0.000 description 1
- 201000005019 Chlamydia pneumonia Diseases 0.000 description 1
- 206010011224 Cough Diseases 0.000 description 1
- 206010011409 Cross infection Diseases 0.000 description 1
- 208000000059 Dyspnea Diseases 0.000 description 1
- 206010013975 Dyspnoeas Diseases 0.000 description 1
- 241000709661 Enterovirus Species 0.000 description 1
- 241000342334 Human metapneumovirus Species 0.000 description 1
- 208000006083 Hypokinesia Diseases 0.000 description 1
- 206010021143 Hypoxia Diseases 0.000 description 1
- 208000019693 Lung disease Diseases 0.000 description 1
- 206010025102 Lung infiltration Diseases 0.000 description 1
- 241000202934 Mycoplasma pneumoniae Species 0.000 description 1
- 208000031662 Noncommunicable disease Diseases 0.000 description 1
- 206010029803 Nosocomial infection Diseases 0.000 description 1
- 208000002606 Paramyxoviridae Infections Diseases 0.000 description 1
- 206010035737 Pneumonia viral Diseases 0.000 description 1
- 206010037660 Pyrexia Diseases 0.000 description 1
- 241000725643 Respiratory syncytial virus Species 0.000 description 1
- 241000315672 SARS coronavirus Species 0.000 description 1
- 241000008910 Severe acute respiratory syndrome-related coronavirus Species 0.000 description 1
- 206010047115 Vasculitis Diseases 0.000 description 1
- 210000001552 airway epithelial cell Anatomy 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 208000023367 bronchiolitis obliterans with obstructive pulmonary disease Diseases 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 239000003153 chemical reaction reagent Substances 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000000280 densification Methods 0.000 description 1
- 201000001981 dermatomyositis Diseases 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 208000017574 dry cough Diseases 0.000 description 1
- 238000011841 epidemiological investigation Methods 0.000 description 1
- 208000018875 hypoxemia Diseases 0.000 description 1
- 238000000338 in vitro Methods 0.000 description 1
- 238000011534 incubation Methods 0.000 description 1
- 208000015181 infectious disease Diseases 0.000 description 1
- 230000008595 infiltration Effects 0.000 description 1
- 238000001764 infiltration Methods 0.000 description 1
- 210000004698 lymphocyte Anatomy 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 231100000915 pathological change Toxicity 0.000 description 1
- 230000036285 pathological change Effects 0.000 description 1
- 208000030773 pneumonia caused by chlamydia Diseases 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 230000002685 pulmonary effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 230000008719 thickening Effects 0.000 description 1
- 241000701161 unidentified adenovirus Species 0.000 description 1
- 241000712461 unidentified influenza virus Species 0.000 description 1
- 208000009421 viral pneumonia Diseases 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/80—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for detecting, monitoring or modelling epidemics or pandemics, e.g. flu
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30061—Lung
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Biomedical Technology (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
Abstract
The invention discloses a novel coronavirus infection pneumonia artificial intelligent rapid screening system and an operation method thereof, wherein the system comprises: the system comprises a database module (1), a data marking module (2), a preprocessing module (3), a training and learning module (4), an artificial intelligent network model (5) and a novel coronavirus infection pneumonia artificial intelligent identification model (6). The invention combines the artificial intelligence technology with the traditional CT imaging technology, and greatly improves the speed and the reliability of identifying the novel coronavirus infection pneumonia.
Description
Technical Field
The invention relates to the technical field of application of artificial intelligence in medicine, in particular to a novel artificial intelligence rapid screening system for coronavirus infection pneumonia.
Background
The diagnostic standard of the suspected patient of the novel coronavirus infection pneumonia comprises a definite epidemic history, the chest imaging examination of the patient shows that the total number of positive and/or blood conventional leucocytes is normal or reduced, the appearance of the reduction of the lymphocyte count can be diagnosed as a suspected case, and the nucleic acid detection is carried out on the suspected case to judge the diagnosed case.
At present, nucleic acid detection is a precise means for diagnosing 2019-nCoV infected patients, but most hospitals in China are lack of nucleic acid detection reagents and related technologies, a few hours are needed from detection to detection result obtaining, and huge time cost and medical cost are needed for rapid popularization in technology. The use of nucleic acid detection for the initial suspected case is likely to cause waste of medical resources.
Based on the reasons, the invention provides an artificial Intelligence system which integrates CT flat scanning of the chest of a patient with an Artificial Intelligence (AI) technology and builds for rapidly screening 2019-nCoV infected patients. The CT imaging examination of the patient with the epidemic history is accurately and quickly scanned through intelligent software, and whether the patient has multiple small spot shadows and interstitial changes or not is judged, so that the extrapulmonary zone is obvious, or whether the double lungs have multiple glass shadows and infiltration shadows and lung excess changes or not is judged. Therefore, suspected cases can be screened out accurately and quickly, and the working efficiency is improved.
Disclosure of Invention
The embodiment of the invention provides a novel artificial intelligent rapid screening system for coronavirus infection pneumonia, so that the technical problem of low reliability of a mode for identifying the novel coronavirus infection pneumonia in the related technology is solved.
According to an aspect of an embodiment of the present invention, the present invention provides a novel coronavirus infection pneumonia artificial intelligence rapid screening system, which is characterized by comprising: the system comprises a database module, a data marking module, a preprocessing module, a training and learning module, an artificial intelligent network model and a novel coronavirus infection pneumonia artificial intelligent identification model;
the database module is used for collecting and storing the image data of the novel coronavirus infection pneumonia in a preset quantity; the data marking module is used for analyzing, summarizing and marking the image characteristics of the stored data; the preprocessing module is used for preprocessing the marked medical image in the database module; the training and learning module is used for training and learning a part of the preprocessed medical image data; training a preset artificial intelligence network model by utilizing the preprocessed training image to obtain a novel artificial intelligence identification preliminary model of coronavirus infection pneumonia.
Optionally, each image of the image data of the predetermined number of new coronavirus infection pneumonia comprises: and (5) identifying areas of the pneumonia foci.
Optionally, wherein the pre-processing comprises: carrying out format conversion on the medical image images of the preset number to obtain gray level images corresponding to the medical image images of the preset number; and carrying out binarization processing on the gray-scale image in a binarization processing mode to obtain a standardized gray-scale image, and simultaneously carrying out corrosion and expansion processing on the standardized gray-scale image to filter interference points in the standardized image to obtain a filtered gray-scale image.
Optionally, the pre-processing further comprises: carrying out convolution processing on the preprocessed training image by utilizing a convolution neural network to obtain a convolution processing result; and inputting the convolution processing result into a regional candidate network for result demonstration.
Optionally, the method for performing result demonstration is to perform artificial intelligence image recognition on a predetermined number of lung CT images, and when the accuracy of the trained artificial intelligence recognition model for the new coronavirus pneumonia exceeds 90% for the predetermined number of CT images, we successfully train the artificial intelligence recognition model for the new coronavirus pneumonia.
Optionally, the marking is based on:
(1) in the early stage of lesion, the lung is scattered in a small piece shape, and the density and the real variation of focal frosted glass distributed under the pleura or distributed along the bronchial tree are reduced;
(2) in the progressive stage of lesion, the focus is increased, the range is enlarged, a plurality of lung lobes are involved, and the ground glass density shadow can be fused into a plurality of real change shadows scattered in the progressive stage of lesion, wherein the signs of gas filled bronchi can be seen;
(3) in the severe stage of disease, diffuse disease of both lungs, and a few of them show "white lung"; taking the real change shadow as a main shadow, merging GGO, and carrying out multiple bougie shadows; air bronchus characterization;
(4) in the convalescent period of the disease, the lung interstitium is changed into the main part, and the main part is accompanied by lymphadenectasis in the mediastinum, pleural effusion and mediastinal emphysema.
Alternatively, the markers need to be identified from CT images of other diseases.
According to another aspect of the embodiment of the invention, the operation method of the novel coronavirus infection pneumonia artificial intelligent rapid screening system comprises the following steps:
step S102: collecting and storing the image data of the novel coronavirus infection pneumonia with a preset quantity by using a database module; then, the data marking module is used for analyzing and summarizing the standardization of the image characteristics of the stored data, and marking the image characteristics, wherein the marks comprise: the characteristics of a marking region and a focus of the novel coronavirus infection pneumonia;
step S104: preprocessing the marked medical image in the database module by using a preprocessing module, processing an interference area outside the mark to obtain a processed medical image, and taking one part of the preprocessed medical image data as a training image in a training and learning module;
step S106: inputting the training image into a preset artificial intelligence network model, and training the preset artificial intelligence network model by utilizing the preprocessed training image to obtain a novel artificial intelligence recognition preliminary model of coronavirus infection pneumonia.
Optionally, the method for forming the final version of the artificial intelligent rapid screening system for pneumonia infected by the novel coronavirus includes:
after step S106, CT raw data of a new predetermined number of patients is imported; correcting the preliminary model to obtain a novel coronavirus infection pneumonia artificial intelligent identification final version model, and preliminarily forming an artificial intelligent rapid screening system; the novel coronavirus infection pneumonia artificial intelligent rapid screening system is connected with a clinical work CT scanning platform for verification, and is finally corrected and optimized by cooperation of a clinician; finally, a final version of the novel coronavirus infection pneumonia artificial intelligent rapid screening system is formed.
The beneficial effects of the invention include:
1. suspected cases can be accurately and quickly screened out, the stay time of a patient in a hospital is reduced, and the risk of cross infection is reduced;
2. human errors in the work of medical workers are reduced, the diagnosis precision is improved, the single CT examination time is shortened, and the work efficiency is improved;
compared with nucleic acid examination, CT examination is more popular nationwide, and has the advantages of more benefited people, reduction of blind utilization rate of nucleic acid detection, more accurate people facing nucleic acid detection, and reduction of medical cost and time cost.
4. The unified image diagnosis standard is formed, and the influence of the medical treatment level irregularity in each region is avoided.
Drawings
FIG. 1 is a schematic flow chart of the novel artificial intelligent rapid screening system for coronavirus infection pneumonia according to the invention;
FIG. 2 is a schematic diagram of the formation of a final version of the novel coronavirus infection pneumonia artificial intelligence rapid screening system of the invention;
FIGS. 3 and 4 are CT images of patients with the novel coronavirus having typical characteristics;
FIGS. 5 and 6 are images of early CT images of patients with the novel coronavirus;
FIGS. 7-10 are CT images of the progression of a patient with the novel coronavirus;
FIGS. 11-12 are CT images of patients with the novel coronavirus in the severe stage;
FIG. 13 is a CT image of the convalescent phase of a patient with the novel coronavirus;
FIGS. 14-16 are CT images of patients with other viral infections.
Detailed Description
In order to make those skilled in the art better understand the technical solution of the present invention, the technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention.
The novel coronavirus (2019-nCoV) belongs to the genus beta, and has an envelope, a round or oval particle shape, usually a polymorphism, and a diameter of 60-140 nm. The gene characteristics of the strain are obviously different from those of SARSr-CoV and MERSR-CoV, and the current research shows that the homology with bat SARS-like coronavirus (bat-SL-CoVZC45) reaches more than 85 percent. In vitro isolation culture, 2019-nCoV96 hours or so can be found in human airway epithelial cells, while in Vero E6 and Huh-7 cell lines, isolation culture takes about 6 days. Based on epidemiological investigation, the incubation period is 1-14 days, mostly 3-7 days, the main clinical manifestations are fever, hypodynamia and dry cough as main symptoms, and severe patients often have dyspnea and/or hypoxemia after one week of onset. The early stage of chest imaging can show multiple small spot images and interstitial changes, and the lung with the exterior zone is obvious. Further, the lung disease develops into a double lung multiple-wear glass shadow and a infiltrative shadow, and the severe cases can cause lung excess change, so that pleural effusion is rare. The imaging characteristics have a certain specificity (as shown in figures 3 and 4).
According to the invention, the CT image is combined with an artificial intelligence system, the artificial intelligence recognition system is utilized to rapidly recognize whether positive lesion exists in the CT image of the chest of the examined person, the positive lesion image is intercepted and derived, the whole recognition process only needs 2-3 seconds and diagnosis is made, and suspicious cases are rapidly screened. Greatly reduces the human operation identification error, reduces the missed diagnosis rate and improves the diagnosis efficiency. As the CT scanning is popularized in most hospitals nationwide, the beneficial population is wider, and the CT scanning is not influenced by the uneven medical level of each region, so that diagnosis and unified standards are easy to form.
Fig. 1 shows a schematic diagram of the novel artificial intelligent rapid screening system for coronavirus infection pneumonia of the invention, which comprises the following modules: the system comprises a database module 1, a data marking module 2, a preprocessing module 3, a training and learning module 4, an artificial intelligent network model 5 and a novel coronavirus infection pneumonia artificial intelligent identification model 6. The database module 1 is used for collecting and storing the image data of the novel coronavirus infection pneumonia with a preset quantity; the data marking module 2 is used for analyzing, summarizing and marking the image characteristics of the stored data; the preprocessing module 3 is used for preprocessing the marked medical image in the database module 1; the training and learning module 4 is used for training and learning a part of the preprocessed medical image data; and training a preset artificial intelligence network model 5 by utilizing the preprocessed training image to obtain a novel coronavirus infection pneumonia artificial intelligence recognition model 6.
The modules work according to the steps shown in fig. 1:
step S102: collecting and storing the image data of the novel coronavirus infection pneumonia with a preset quantity by using a database module 1; then, the data marking module 2 is used for analyzing and summarizing the standardization of the image characteristics of the stored data, and marking the image characteristics, wherein the marks comprise: the characteristics of a marking region and a focus of the novel coronavirus infection pneumonia;
wherein each image in the image data of the predetermined number of novel coronavirus (2019-nCoV) infection pneumonia comprises: and (5) identifying areas of the pneumonia foci.
Step S104: preprocessing the marked medical image in the database module 1 by using a preprocessing module 3, processing an interference area outside the mark to obtain a processed medical image, and taking one part of the preprocessed medical image data as a training image in a training and learning module 4;
the pretreatment comprises the following steps: carrying out format conversion on the medical image images of the preset number to obtain gray level images corresponding to the medical image images of the preset number; and carrying out binarization processing on the gray-scale image in a binarization processing mode to obtain a standardized gray-scale image, and simultaneously carrying out corrosion and expansion processing on the standardized gray-scale image to filter interference points in the standardized image to obtain a filtered gray-scale image.
Carrying out convolution processing on the preprocessed training image by utilizing a convolution neural network to obtain a convolution processing result; and inputting the convolution processing result into a regional candidate network for result demonstration.
Step S106: inputting the training image into a preset artificial intelligence network model 5, and training the preset artificial intelligence network model 5 by utilizing the preprocessed training image to obtain a novel coronavirus infection pneumonia artificial intelligence recognition preliminary model 6-1.
Wherein the predetermined number of medical image images are labeled according to the following:
(1) as shown in FIGS. 5 and 6, early stage lesions are mostly characterized by focal frosted glass density and solid lesions scattered in the lung in small pieces, distributed under the pleura or distributed along the bronchial tree.
Early CT presentation:
the limitation of the pathological changes, mainly the plaque, sublevel or segmental distribution
Distribution under pleura
GGO with or without thickening of leaflet spaces
(2) As shown in the accompanying figures 7-10, in the progressive stage of the disease, most patients reach the most severe stage of lung infiltration within 8-14 days of the disease, the frosted glass density images can be fused into the solid variable images scattered in the multiple stages, and the signs of the gas filled bronchi can be seen in the solid variable images.
The CT performance in the progressive stage:
the disease progresses, the focus is increased, the range is enlarged, and a plurality of lung lobes are affected
Partial lesion densification, GGO coexistence with real deformation shadow or streak shadow
A small amount of pleural effusion
(3) As shown in fig. 11-12, the disease progressed to severe stage, and a few patients reached severe stage in later stage of disease.
Severe phase CT presentation:
diffuse lesions of both lungs and a small number of "white lungs" manifestations
The real change shadow is dominant, GGO are merged and a plurality of funny shadows are formed
Air bronchus sign
(4) As shown in FIG. 13, the convalescent period of the disease is dominated by the change of the pulmonary interstitium. Some cases are accompanied by lymphadenectasis in mediastinum, pleural effusion, and mediastinal emphysema.
(5) As shown in FIGS. 14-16, are CT images of patients with other viral infections. The CT image of the novel coronavirus needs to be identified with other diseases, particularly with the CT image of other virus infections, so that the identification accuracy rate of the novel coronavirus infection pneumonia is improved.
The virus is mainly identified with other known viral pneumonia such as influenza virus, parainfluenza virus, adenovirus, respiratory syncytial virus, rhinovirus, human metapneumovirus, SARS coronavirus and the like, and identified with mycoplasma pneumoniae, chlamydia pneumonia, bacterial pneumonia and the like. In addition, non-infectious diseases, such as vasculitis, dermatomyositis, and organized pneumonia, are identified.
FIG. 2 is a schematic diagram of the formation of the final version of the novel coronavirus infection pneumonia artificial intelligent rapid screening system. As shown in fig. 2, the formation of the final version of the novel artificial intelligent rapid screening system for coronavirus infection pneumonia of the present invention comprises the following steps:
(1) and (3) finishing CT original data collection of confirmed cases: selecting CT flat scanning original images of a certain number (for example 100) of patients with positive nucleic acid detection in different development periods as original data, establishing a database through a database module 1, and collecting and storing the data;
(2) marking the CT imaging data by using a data marking module 2; preprocessing the CT imaging data by using a preprocessing module 3, and taking a part of the preprocessed medical image data as a training image in a training and learning module 4; a predetermined artificial intelligence network model 5 is trained using the training image,
(3) forming a novel coronavirus infection pneumonia artificial intelligent identification preliminary model 6-1;
(4) importing CT raw data of a new predetermined number (e.g. 100) of patients;
(5) correcting the preliminary model 6-1 to obtain a novel coronavirus infection pneumonia artificial intelligent identification preliminary final version model, and preliminarily forming an artificial intelligent rapid screening system;
(6) the novel artificial intelligent rapid screening system for coronavirus infection pneumonia is connected with a clinical work CT scanning platform for verification, and is finally corrected and optimized by cooperation of a clinician.
(7) Finally, a final version of the novel coronavirus infection pneumonia artificial intelligent rapid screening system is formed.
In clinical application, the initial model of the invention is corrected and optimized for many times through actual data to form a final version of the novel coronavirus infection pneumonia artificial intelligent rapid screening system, and the accuracy of the system can be improved.
Optionally, the artificial intelligent identification model for coronavirus infection pneumonia trained through the bottom neural network is professionally verified, the verification method is used for selecting a predetermined amount of CT chest radiographs, wherein the CT chest radiographs comprise confirmed novel coronavirus infection pneumonia and other virus induced pneumonia, and the accuracy of radiograph reading of the artificial intelligent identification model for coronavirus infection pneumonia trained is used for verifying whether the model is successful or not. When the accuracy of the trained novel coronavirus infection pneumonia artificial intelligent recognition model for recognizing CT images with the preset number exceeds 90%, the novel coronavirus infection pneumonia artificial intelligent recognition model is successfully trained.
Optionally, in the embodiment of the present invention, the preprocessing of the image data of the predetermined number may include extracting and classifying CT features of the new coronavirus infection pneumonia.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A novel coronavirus infection pneumonia artificial intelligence rapid screening system is characterized by comprising:
the system comprises a database module (1), a data marking module (2), a preprocessing module (3), a training and learning module (4), an artificial intelligent network model (5) and a novel coronavirus infection pneumonia artificial intelligent identification model (6);
the database module (1) is used for collecting and storing the image data of the novel coronavirus infection pneumonia with a preset quantity;
the data marking module (2) is used for analyzing, summarizing and marking the image characteristics of the stored data;
the preprocessing module (3) is used for preprocessing the marked medical image in the database module (1);
the training and learning module (4) is used for training and learning a part of the preprocessed medical image data;
training a preset artificial intelligence network model (5) by utilizing the preprocessed training image to obtain a novel artificial intelligence identification preliminary model (6) for coronavirus infection pneumonia.
2. The system of claim 1, wherein each image of the image data of the predetermined number of new coronavirus infection pneumonia comprises: and (5) identifying areas of the pneumonia foci.
3. The novel artificial intelligence rapid screening system for coronavirus infection pneumonia according to claim 1, wherein the preprocessing comprises: carrying out format conversion on the medical image images of the preset number to obtain gray level images corresponding to the medical image images of the preset number; and carrying out binarization processing on the gray-scale image in a binarization processing mode to obtain a standardized gray-scale image, and simultaneously carrying out corrosion and expansion processing on the standardized gray-scale image to filter interference points in the standardized image to obtain a filtered gray-scale image.
4. The novel artificial intelligence rapid screening system for coronavirus infection pneumonia according to claim 3, wherein said preprocessing further comprises: carrying out convolution processing on the preprocessed training image by utilizing a convolution neural network to obtain a convolution processing result; and inputting the convolution processing result into a regional candidate network for result demonstration.
5. The system for artificial intelligent rapid screening of coronavirus infection pneumonia according to claim 4, wherein the method for performing result demonstration comprises performing artificial intelligent image recognition on CT images of a predetermined number of lungs, and when the accuracy of the trained artificial intelligent identification model for coronavirus infection pneumonia exceeds 90% for the CT images of the predetermined number, we successfully trained an artificial intelligent identification final model for coronavirus infection pneumonia.
6. The system of claim 1, wherein the pre-processing of the image data includes CT feature extraction and classification of the coronavirus infection pneumonia.
7. The novel artificial intelligence rapid screening system for coronavirus infection pneumonia according to claim 1, wherein the marking is based on:
(1) in the early stage of lesion, the lung is scattered in a small piece shape, and the density and the real variation of focal frosted glass distributed under the pleura or distributed along the bronchial tree are reduced;
(2) in the progressive stage of lesion, the focus is increased, the range is enlarged, a plurality of lung lobes are involved, and the ground glass density shadow can be fused into a plurality of real change shadows scattered in the progressive stage of lesion, wherein the signs of gas filled bronchi can be seen;
(3) in the severe stage of disease, diffuse disease of both lungs, and a few of them show white lung; taking the real change shadow as a main shadow, merging GGO, and carrying out multiple bougie shadows; air bronchus characterization;
(4) in the convalescent period of the disease, the lung interstitium is changed into the main part, and the main part is accompanied by lymphadenectasis in the mediastinum, pleural effusion and mediastinal emphysema.
8. The novel artificial intelligence rapid screening system for coronavirus infection pneumonia of claim 7, wherein said marker needs to be identified with CT image of other diseases.
9. An operating method of the novel coronavirus infection pneumonia artificial intelligent rapid screening system according to any one of claims 1-8, which comprises the following steps:
step S102: collecting and storing the image data of the novel coronavirus infection pneumonia with a preset quantity by using a database module (1); then, a data marking module (2) is used for analyzing and summarizing the standardization of the image characteristics of the stored data, and marking the image characteristics, wherein the marks comprise: the characteristics of a marking region and a focus of the novel coronavirus infection pneumonia;
step S104: preprocessing the marked medical image in the database module (1) by using a preprocessing module (3), processing an interference area outside the mark to obtain a processed medical image, and taking one part of the preprocessed medical image data as a training image in a training and learning module (4);
step S106: inputting the training image into a preset artificial intelligence network model (5), and training the preset artificial intelligence network model (5) by utilizing the preprocessed training image to obtain a novel coronavirus infection pneumonia artificial intelligence recognition preliminary model (6-1).
10. The method of forming a novel coronavirus infection pneumonia artificial intelligence rapid screening system of claim 9, comprising:
after step S106, CT raw data of a new predetermined number of patients is imported;
correcting the preliminary model (6-1) to obtain a novel coronavirus infection pneumonia artificial intelligent identification final version model, and preliminarily forming an artificial intelligent rapid screening system;
the novel coronavirus infection pneumonia artificial intelligent rapid screening system is connected with a clinical work CT scanning platform for verification, and is finally corrected and optimized by cooperation of a clinician;
forming a final version of the novel artificial intelligent rapid screening system for coronavirus infection pneumonia.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010083930.3A CN112349425A (en) | 2020-02-10 | 2020-02-10 | Novel artificial intelligent rapid screening system for coronavirus infection pneumonia |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010083930.3A CN112349425A (en) | 2020-02-10 | 2020-02-10 | Novel artificial intelligent rapid screening system for coronavirus infection pneumonia |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112349425A true CN112349425A (en) | 2021-02-09 |
Family
ID=74357148
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010083930.3A Pending CN112349425A (en) | 2020-02-10 | 2020-02-10 | Novel artificial intelligent rapid screening system for coronavirus infection pneumonia |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112349425A (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060218010A1 (en) * | 2004-10-18 | 2006-09-28 | Bioveris Corporation | Systems and methods for obtaining, storing, processing and utilizing immunologic information of individuals and populations |
| CN102460471A (en) * | 2009-06-19 | 2012-05-16 | 美国医软科技公司 | Systems for computer aided lung nodule detection in chest tomosynthesis imaging |
| CN107103187A (en) * | 2017-04-10 | 2017-08-29 | 四川省肿瘤医院 | The method and system of Lung neoplasm detection classification and management based on deep learning |
| CN107910061A (en) * | 2017-12-01 | 2018-04-13 | 中南大学 | A kind of medical data processing method and system |
| CN108520778A (en) * | 2018-04-20 | 2018-09-11 | 中日友好医院 | A kind of screening lung cancer system and method based on big data |
| CN108776962A (en) * | 2018-04-11 | 2018-11-09 | 浙江师范大学 | A method of the structure good pernicious prediction model of lung neoplasm |
| CN108898593A (en) * | 2018-06-25 | 2018-11-27 | 重庆知遨科技有限公司 | A kind of detection method based on the classification of abdominal CT Medical image fusion |
| CN108985345A (en) * | 2018-06-25 | 2018-12-11 | 重庆知遨科技有限公司 | A kind of detection device based on the classification of lung's Medical image fusion |
| CN109117713A (en) * | 2018-06-27 | 2019-01-01 | 淮阴工学院 | A kind of drawing printed page analysis of full convolutional neural networks and character recognition method |
| CN110504027A (en) * | 2019-08-20 | 2019-11-26 | 东北大学 | A kind of X-Ray rabat pneumonia intelligent diagnosis system and method based on deep learning |
| KR20190134586A (en) * | 2019-11-28 | 2019-12-04 | 연세대학교 산학협력단 | Method for predicting diseases and device for predicting diseases using the same |
| CN110619947A (en) * | 2019-09-19 | 2019-12-27 | 南京工程学院 | Lung CT auxiliary screening system and method based on lightweight deep learning |
| US20210042916A1 (en) * | 2018-02-07 | 2021-02-11 | Ai Technologies Inc. | Deep learning-based diagnosis and referral of diseases and disorders |
-
2020
- 2020-02-10 CN CN202010083930.3A patent/CN112349425A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060218010A1 (en) * | 2004-10-18 | 2006-09-28 | Bioveris Corporation | Systems and methods for obtaining, storing, processing and utilizing immunologic information of individuals and populations |
| CN102460471A (en) * | 2009-06-19 | 2012-05-16 | 美国医软科技公司 | Systems for computer aided lung nodule detection in chest tomosynthesis imaging |
| CN107103187A (en) * | 2017-04-10 | 2017-08-29 | 四川省肿瘤医院 | The method and system of Lung neoplasm detection classification and management based on deep learning |
| CN107910061A (en) * | 2017-12-01 | 2018-04-13 | 中南大学 | A kind of medical data processing method and system |
| US20210042916A1 (en) * | 2018-02-07 | 2021-02-11 | Ai Technologies Inc. | Deep learning-based diagnosis and referral of diseases and disorders |
| CN108776962A (en) * | 2018-04-11 | 2018-11-09 | 浙江师范大学 | A method of the structure good pernicious prediction model of lung neoplasm |
| CN108520778A (en) * | 2018-04-20 | 2018-09-11 | 中日友好医院 | A kind of screening lung cancer system and method based on big data |
| CN108898593A (en) * | 2018-06-25 | 2018-11-27 | 重庆知遨科技有限公司 | A kind of detection method based on the classification of abdominal CT Medical image fusion |
| CN108985345A (en) * | 2018-06-25 | 2018-12-11 | 重庆知遨科技有限公司 | A kind of detection device based on the classification of lung's Medical image fusion |
| CN109117713A (en) * | 2018-06-27 | 2019-01-01 | 淮阴工学院 | A kind of drawing printed page analysis of full convolutional neural networks and character recognition method |
| CN110504027A (en) * | 2019-08-20 | 2019-11-26 | 东北大学 | A kind of X-Ray rabat pneumonia intelligent diagnosis system and method based on deep learning |
| CN110619947A (en) * | 2019-09-19 | 2019-12-27 | 南京工程学院 | Lung CT auxiliary screening system and method based on lightweight deep learning |
| KR20190134586A (en) * | 2019-11-28 | 2019-12-04 | 연세대학교 산학협력단 | Method for predicting diseases and device for predicting diseases using the same |
Non-Patent Citations (1)
| Title |
|---|
| 管汉雄等: "新型冠状病毒肺炎（COVID-19）临床影像学特征", 《放射学实践》, vol. 35, no. 2, 3 February 2020 (2020-02-03), pages 125 - 130 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109670510B (en) | Deep learning-based gastroscope biopsy pathological data screening system | |
| CN112365464B (en) | GAN-based medical image lesion area weak supervision positioning method | |
| CN110991536B (en) | Training method of early warning model of primary liver cancer | |
| CN111340130A (en) | Urinary calculus detection and classification method based on deep learning and imaging omics | |
| CN109727260A (en) | A kind of three-dimensional lobe of the lung dividing method based on CT images | |
| CN110584618A (en) | Infectious disease machine recognition system based on artificial intelligence | |
| CN111862075A (en) | Lung image analysis system and method based on deep learning | |
| Shen et al. | Guest editorial: special issue on imaging-based diagnosis of COVID-19 | |
| CN115862819B (en) | Medical image management method based on image processing | |
| CN111833330B (en) | Intelligent lung cancer detection method and system based on fusion of image and machine olfaction | |
| CN111462201B (en) | Follow-up analysis system and method based on novel coronavirus pneumonia CT image | |
| CN109117871A (en) | A kind of system and method that the pernicious probability of pulmonary lesions is obtained by deep learning | |
| CN112116571A (en) | X-ray lung disease automatic positioning method based on weak supervised learning | |
| CN113160974A (en) | Mental disease biological type mining method based on hypergraph clustering | |
| CN113314215A (en) | Ultrasonic thyroid nodule sample abundance and benign and malignant automatic auxiliary identification system | |
| CN113011257A (en) | Breast cancer immunohistochemical artificial intelligence interpretation method | |
| CN113706518A (en) | Femoral head necrosis stage analysis method and system based on artificial intelligence | |
| CN115705928A (en) | TC (total cholesterol) disease risk prediction modeling method based on cell proliferation marker | |
| CN113033667A (en) | Ultrasound image two-stage deep learning breast tumor classification method and device | |
| CN112349425A (en) | Novel artificial intelligent rapid screening system for coronavirus infection pneumonia | |
| CN111507964B (en) | Integrated learning grading method for new coronary pneumonia, electronic equipment and storage medium | |
| CN112489022A (en) | COVID-19 rapid feature extraction system based on radiologics, rapid diagnosis system and disease course prediction system | |
| WO2017166778A1 (en) | Bacteria identification method and apparatus | |
| CN116072302A (en) | Medical unbalanced data classification method based on biased random forest model | |
| CN110334748A (en) | The cancer subtypes classification method of multiple groups data integration is carried out based on D-S evidence theory |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |