CN114983352A - Method and device for identifying new coronary pneumonia based on attention mechanism - Google Patents
Method and device for identifying new coronary pneumonia based on attention mechanism Download PDFInfo
- Publication number
- CN114983352A CN114983352A CN202110222978.2A CN202110222978A CN114983352A CN 114983352 A CN114983352 A CN 114983352A CN 202110222978 A CN202110222978 A CN 202110222978A CN 114983352 A CN114983352 A CN 114983352A
- Authority
- CN
- China
- Prior art keywords
- module
- data
- batch
- oxygen
- attention
- 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 32
- 238000000034 method Methods 0.000 title claims abstract description 25
- 230000007246 mechanism Effects 0.000 title claims abstract description 18
- 238000001514 detection method Methods 0.000 claims abstract description 34
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000007405 data analysis Methods 0.000 claims abstract description 11
- 238000012544 monitoring process Methods 0.000 claims abstract description 9
- 230000005540 biological transmission Effects 0.000 claims abstract description 8
- 230000002159 abnormal effect Effects 0.000 claims abstract description 5
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 24
- 229910052760 oxygen Inorganic materials 0.000 claims description 24
- 239000001301 oxygen Substances 0.000 claims description 24
- 230000036760 body temperature Effects 0.000 claims description 16
- 238000004458 analytical method Methods 0.000 claims description 10
- 230000029058 respiratory gaseous exchange Effects 0.000 claims description 9
- 206010037660 Pyrexia Diseases 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 7
- 108010064719 Oxyhemoglobins Proteins 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000007781 pre-processing Methods 0.000 claims description 5
- 239000013598 vector Substances 0.000 claims description 5
- 238000007418 data mining Methods 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 4
- 206010011224 Cough Diseases 0.000 claims description 3
- 239000008186 active pharmaceutical agent Substances 0.000 claims description 3
- 230000002457 bidirectional effect Effects 0.000 claims description 3
- 239000008280 blood Substances 0.000 claims description 3
- 210000004369 blood Anatomy 0.000 claims description 3
- UBAZGMLMVVQSCD-UHFFFAOYSA-N carbon dioxide;molecular oxygen Chemical compound O=O.O=C=O UBAZGMLMVVQSCD-UHFFFAOYSA-N 0.000 claims description 3
- 230000008030 elimination Effects 0.000 claims description 3
- 238000003379 elimination reaction Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 239000007789 gas Substances 0.000 claims description 3
- 230000036541 health Effects 0.000 claims description 3
- 238000012905 input function Methods 0.000 claims description 3
- 238000007726 management method Methods 0.000 claims description 3
- 230000036284 oxygen consumption Effects 0.000 claims description 3
- 238000006213 oxygenation reaction Methods 0.000 claims description 3
- 238000004171 remote diagnosis Methods 0.000 claims description 3
- 230000036387 respiratory rate Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000013144 data compression Methods 0.000 claims description 2
- 238000011176 pooling Methods 0.000 claims description 2
- 238000013480 data collection Methods 0.000 claims 1
- 230000007547 defect Effects 0.000 abstract description 4
- 238000013499 data model Methods 0.000 abstract description 3
- 230000005856 abnormality Effects 0.000 abstract description 2
- 238000005192 partition Methods 0.000 abstract description 2
- 238000012549 training Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 6
- 201000010099 disease Diseases 0.000 description 5
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 241000711573 Coronaviridae Species 0.000 description 2
- 241000700605 Viruses Species 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 210000004072 lung Anatomy 0.000 description 2
- 150000007523 nucleic acids Chemical class 0.000 description 2
- 102000039446 nucleic acids Human genes 0.000 description 2
- 108020004707 nucleic acids Proteins 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/0205—Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
- A61B5/02055—Simultaneously evaluating both cardiovascular condition and temperature
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/024—Detecting, measuring or recording pulse rate or heart rate
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/08—Detecting, measuring or recording devices for evaluating the respiratory organs
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/08—Detecting, measuring or recording devices for evaluating the respiratory organs
- A61B5/083—Measuring rate of metabolism by using breath test, e.g. measuring rate of oxygen consumption
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/145—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue
- A61B5/14542—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue for measuring blood gases
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7271—Specific aspects of physiological measurement analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Public Health (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Physiology (AREA)
- Surgery (AREA)
- Medical Informatics (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Artificial Intelligence (AREA)
- Cardiology (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Pulmonology (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Fuzzy Systems (AREA)
- Emergency Medicine (AREA)
- Obesity (AREA)
- Optics & Photonics (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention discloses a method and equipment for identifying new coronary pneumonia based on an attention mechanism, which utilize the strong real-time transmission capability of a 5G network to transmit data acquired by wearable equipment to a background computing center and utilize the strong computing capability of a computing server to establish a big data model, thereby avoiding the defects of single function and low precision caused by the fact that a microprocessor is used for training data analysis in a single machine offline mode of the traditional wearable equipment. The remote monitoring mode provided by the invention collects and uploads medical data to monitor, analyze, arrange and partition archive. And processing abnormal data, such as call return, knowing the reason of the abnormality, and timely prompting a doctor in a remote hospital to pay attention to the condition of the patient. The remote monitoring and pre-detection method can greatly reduce manual detection work and accelerate the patient detection speed.
Description
Technical Field
The invention relates to the technical field of medical detection, in particular to a method and equipment for identifying new coronary pneumonia based on an attention mechanism.
Background
At present, the method is used for the accurate diagnosis of the new coronavirus, and patients directly go to a hospital to carry out clinical body temperature detection, lung CT detection and nucleic acid detection. At present, the new coronavirus is not pre-judged by a medical record data mining method. Therefore, how to reduce the manual detection work and accelerate the patient detection speed, and the adoption of an artificial intelligence means to carry out the auxiliary medical treatment is imperative.
Disclosure of Invention
The invention provides a new coronary pneumonia identification method and equipment based on an attention mechanism, aiming at overcoming the defect that heavy manual detection workload exists in new coronary virus detection and aiming at old people and high-risk people with related past medical history, and aims to provide a method based on neural network deep learning, wherein the method is used for identifying remote monitoring data of a 5G network to diagnose whether a patient has new coronary virus signs. The automatic level of hospital detection is provided, and the hospital working efficiency is effectively improved.
In order to achieve the purpose, the invention provides the following technical scheme: the equipment for identifying the new coronary pneumonia based on the attention mechanism comprises a data acquisition module, a data analysis module and a result display module, wherein the data acquisition module, the data analysis module and the result display module are communicated through a 5G network;
the data acquisition module uses wearable check out test set to gather real-time data and to the collection to basic auxiliary information, and main body temperature characteristic, breathing characteristic and pulse characteristic of gathering specifically have the following index: body temperature, heart rate, respiratory rate, oxygen output and oxygen saturation;
the data analysis module is used for conjecturing other index values including inhalation oxygen pressure, arterial oxygen content, inhalation oxygen fraction, oxygen consumption, oxygen use fraction, oxygenation index, respiration index, arterial oxygen content, arterial oxyhemoglobin saturation, oxyhemoglobin saturation and cough frequency by combining data acquired by the data acquisition module with a related calculation formula, so that the risk that a user may suffer from pneumonia or the possibility that the user may suffer from pneumonia is analyzed, and high-risk people are controlled in time;
and the result display module is a use terminal for sending the final analysis result to the user and medical staff in an analysis report form.
Preferably, the data acquisition module comprises a basic module, a signal acquisition module and a signal processing module;
the basic module comprises a power supply module, a storage model and a transmission module;
the signal acquisition module comprises a body temperature detection module, a pulse detection module and a breath detection module;
the signal processing module comprises a filtering module, an amplifying module and an A/D module.
Preferably, the power supply module relates to a hardware battery, and provides energy for the operation of the whole equipment to ensure long-time endurance; the storage model relates to a hardware flash memory chip and caches the acquired data; the transmission module is used as a transmitting end to acquire signal data.
Preferably, the body temperature detection module uses a thermopile infrared temperature sensor to acquire a body temperature signal; the pulse detection module collects pulse wave signals by using a reflective photoelectric sensor; the respiration detection module relates to hardware, and comprises a mask, a sampling needle, a photodiode and a signal receiving device, and signals such as oxygen, carbon dioxide gas content, blood oxygen saturation and the like are collected in a non-invasive mode.
Preferably, the filtering module is a filtering circuit; the amplifying module is a signal amplifying circuit; the A/D module is an analog-to-digital converter which digitizes the received signal for supply to the signal transmission module and the storage module.
Preferably, the auxiliary information of the storage and computation center is: sex, age, height, weight, history of having or not having the pulmonary inflammation of the relative, time since last disease, last disease time since the relative, recent history of having the cold, recent history of having the fever, time since the cold, last fever time, recent route place, history of having or not having the contact of the patient and history of having or not having the indirect contact of the patient.
Preferably, the method for identifying the new coronary pneumonia based on the attention mechanism specifically comprises the following steps:
s1: providing a position tracking service through a 5G network, and tracking the position of user equipment through an LTaaS layer;
s2: the user wears an intelligent helmet with the data input function, the intelligent helmet is used as wearable remote diagnosis and treatment monitoring equipment, and self health data are collected at home timely and automatically uploaded;
s3: the intelligent helmet is wirelessly connected with the resident mobile phone through Bluetooth, and the acquired data are transmitted to a 5G network RAN cloud network resource management layer through a mobile phone API;
s4: the storage and calculation center of the medical institution end mines the uploaded data;
s5: and the final analysis result is sent to the user and the doctor terminal through the 5G network.
Preferably, the data mining process of the storage and computation center specifically includes the following steps:
s4.1: the process begins with auxiliary data, i.e., historical condition information and data collected by the equipment as inputs;
s4.2: preprocessing, including missing value processing, abnormal value elimination, continuous feature standardization and class feature Onehot coding, setting a batch _ masks matrix, wherein the value of the position with the feature is 1, the other positions are 0, the matrix is used for calculating Attention later, and the attentions of the rest positions without the feature are changed into 0;
s4.3: the method comprises the following steps that firstly, a model network structure carries out primary feature extraction on original features through an Encoder Encoder, and the specific operation is that preprocessed data are spliced and then transmitted into a CNNEncoder, wherein the CNNEncoder defines 3 convolution kernels, the sizes of the convolution kernels are [2, 10], [3, 10], [4, 10], namely, the output of;
s4.4: transmitting the send _ reps into an LSTMEncoder, wherein the LSTMEncoder is a 2-layer bidirectional LSTM structure, multiplying the LSTMEncoder by a mask according to positions, changing the position of the feature loss into 0, and recording the final output data as batch _ hiddens;
s4.5: the Attention is introduced into a model for pneumonia identification, and when a certain characteristic changes, the weights of other associated characteristics change: taking the batch _ hiddens and batch _ masks as the input of the Attention; in the Attention, firstly copying one part of batch _ hiddens to obtain key through linear change, keeping the dimension unchanged, and additionally copying two parts of batch _ hiddens to obtain query and value, wherein the query is used for calculating similarity with keys, and the value is used for weighting with the calculated attribute weight to obtain final output; multiplying the key and the query to obtain outputs which represent the weight distributed to each feature; next, performing softmax processing on the atttion, setting the weight without the characteristic to-1 e32 by using batch _ masks, and marking the obtained result as masked _ attn _ scores; finally, multiplying the masked _ attn _ scores with the value to obtain the batch _ outputs;
s4.6: obtaining a vector of classification probability through a full-connection FC layer;
s4.7: and predicting the new data through the trained model to obtain a judgment result.
Compared with the prior art, the invention has the beneficial effects that:
1) according to the attention mechanism-based new coronary pneumonia identification method and equipment, real-time physiological data are collected locally, transmitted to a remote computing center through a 5G network, combined with other basic information and past medical history of a user, effective information is further mined through artificial intelligence technologies such as a neural network and the like, whether the user has a risk of illness or the possibility of illness is judged, and the user with high risk is reminded to arrive at a hospital for rechecking as soon as possible.
2) According to the method and the device for identifying the new coronary pneumonia based on the attention mechanism, various data are collected through various physiological index collection sensors on the wearable device, and the characteristics of multi-user, multi-point, multi-antenna and multi-ingestion collaborative networking of a 5G network are utilized, so that the collected data can be rapidly transmitted to a remote computing end, and the defects that a traditional wearable device is high in cost due to the use of a simple computing module, poor in computing performance and the like are overcome.
3) The method for identifying the new coronary pneumonia based on the attention mechanism comprises the steps of signal preprocessing, feature engineering, neural networks, model debugging and the like, and the disease risk of a user is deeply mined by combining personalized features (basic information and medical history) of the user and real-time physiological index data. And informing the user judged to be at high risk to visit a nearby hospital for doctor face diagnosis and lung CT or nucleic acid detection.
According to the wearable data model building method, the strong real-time transmission capability of the 5G network is utilized, the data collected by the wearable equipment are transmitted to the background computing center, and the strong computing capability of the computing server is utilized to build a big data model, so that the defect that the function is single and the precision is low when a microprocessor is used for training data analysis in a single machine offline mode of the traditional wearable equipment is overcome. The remote monitoring mode provided by the invention collects and uploads medical data to carry out monitoring, analysis and arrangement and partition filing. And processing abnormal data, such as call return visit, knowing the cause of the abnormality and prompting a doctor in a remote hospital to pay attention to the condition of the patient in time. The remote monitoring and pre-detection method can greatly reduce the manual detection work and accelerate the patient detection speed.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for describing the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a system architecture diagram of the present invention;
FIG. 2 is a diagram of the hardware functional modules of the wearable device of the present invention;
FIG. 3 is a block diagram of an auxiliary diagnostic method of the present invention;
FIG. 4 is a diagram of a neural network topology of the present invention;
FIG. 5 is a diagram of the Encoder network structure of the present invention;
FIG. 6 is a logical structure diagram of the Attention mechanism of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention.
Referring to fig. 1 to 6, the present invention provides the following technical solutions:
example one
The equipment for identifying the new coronary pneumonia based on the attention mechanism comprises a data acquisition module, a data analysis module and a result display module, wherein the data acquisition module, the data analysis module and the result display module are communicated through a 5G network;
the data acquisition module uses wearable check out test set to gather real-time data and to the collection to basis auxiliary information, mainly gathers body temperature characteristic, breathing characteristic and pulse characteristic, specifically has following index: body temperature, heart rate, respiratory rate, oxygen output and oxygen saturation;
the data analysis module is used for conjecturing other index values including inhalation oxygen pressure, arterial oxygen content, inhalation oxygen fraction, oxygen consumption, oxygen use fraction, oxygenation index, respiration index, arterial oxygen content, arterial oxyhemoglobin saturation, oxyhemoglobin saturation and cough frequency by combining data acquired by the data acquisition module with a related calculation formula, so that the risk that a user may suffer from pneumonia or the possibility that the user may suffer from pneumonia is analyzed, and high-risk people are controlled in time;
and the result display module is a use terminal for sending the final analysis result to the user and medical staff in an analysis report form.
The data acquisition module comprises a basic module, a signal acquisition module and a signal processing module;
the basic module comprises a power supply module, a storage model and a transmission module;
the signal acquisition module comprises a body temperature detection module, a pulse detection module and a breath detection module;
the signal processing module comprises a filtering module, an amplifying module and an A/D module.
The power supply module relates to a hardware battery and provides energy for the operation of the whole equipment to ensure long-time endurance; the storage model relates to a hardware flash memory chip and caches the acquired data; the transmission module is used as a transmitting end to acquire signal data.
The body temperature detection module acquires a body temperature signal by using a thermopile infrared temperature sensor; the pulse detection module uses a reflective photoelectric sensor to collect pulse wave signals; the respiration detection module relates to hardware, and comprises a mask, a sampling needle, a photodiode and a signal receiving device, and signals such as oxygen, carbon dioxide gas content, blood oxygen saturation and the like are collected in a non-invasive mode.
The filtering module is a filtering circuit; the amplifying module is a signal amplifying circuit; the A/D module is an analog-to-digital converter which digitizes the received signal to provide to the signal sending module and the storage module.
The auxiliary information of the storage and calculation center comprises the following information: sex, age, height, weight, history of having or not having the pulmonary inflammation of the relative, time since last disease, last disease time since the relative, recent history of having the cold, recent history of having the fever, time since the cold, last fever time, recent route place, history of having or not having the contact of the patient and history of having or not having the indirect contact of the patient.
Example two
A new coronary pneumonia identification method based on an attention mechanism specifically comprises the following steps:
s1: providing a position tracking service through a 5G network, and tracking the position of user equipment through an LTaaS layer;
s2: the user wears an intelligent helmet with the data input function, the intelligent helmet is used as wearable remote diagnosis and treatment monitoring equipment, and self health data are collected at home timely and automatically uploaded;
s3: the intelligent helmet is wirelessly connected with the resident mobile phone through Bluetooth, and the acquired data are transmitted to a 5G network RAN cloud network resource management layer through a mobile phone API;
s4: the storage and calculation center of the medical institution end excavates the uploaded data;
s5: and the final analysis result is sent to the user and the doctor terminal through the 5G network.
The data mining process of the storage and calculation center specifically comprises the following steps:
s4.1: the process begins with auxiliary data, i.e., historical condition information and data collected by the equipment as inputs;
s4.2: preprocessing is carried out, wherein the preprocessing comprises missing value processing, abnormal value elimination, continuous feature standardization and class feature Onehot coding, a batch _ masks matrix is set, the value of a position with a feature is 1, the value of other positions is 0, the following calculation of Attention is carried out, and attentions of other positions without the feature are changed into 0;
s4.3: firstly, performing primary feature extraction on original features by using an Encoder Encoder, specifically splicing preprocessed data, and then transmitting the preprocessed data into a CNNEncoder, wherein the CNNEncoder defines 3 convolution kernels with the sizes of [2, 10], [3, 10], [4, 10], namely all output puthannels are 10 dimensions, performing data compression on a convolved result through a pooling layer, and finally splicing the obtained 3 vectors in a second dimension to obtain data which is recorded as sent _ reps;
s4.4: transmitting the send _ reps into an LSTMEncoder, wherein the LSTMEncoder is a 2-layer bidirectional LSTM structure, multiplying the LSTMEncoder by a mask according to positions, changing the position of the feature loss into 0, and recording the final output data as batch _ hiddens;
s4.5: the Attention is introduced into a model for pneumonia identification, and when a certain characteristic changes, the weights of other associated characteristics change: taking the batch _ hiddens and batch _ masks as the input of the Attention; in the Attention, firstly copying one part of batch _ hiddens to obtain key through linear change, keeping the dimension unchanged, and additionally copying two parts of batch _ hiddens to obtain query and value, wherein the query is used for calculating similarity with keys, and the value is used for weighting with the calculated attribute weight to obtain final output; multiplying the key and the query to obtain outputs which represent the weight distributed to each feature; next, softmax processing is carried out on the attribute, batch _ masks are used, the weight without features is set to be-1 e32, and the obtained result is marked as masked _ attn _ scores; finally, multiplying the masked _ attn _ scores with the value to obtain the batch _ outputs;
s4.6: obtaining a vector of classification probability through a full-connection FC layer;
s4.7: and predicting the new data by the trained model to obtain a judgment result.
Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium.
The preferred embodiments of the invention disclosed above are intended to be illustrative only. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
Claims (8)
1. The equipment for identifying the new coronary pneumonia based on the attention mechanism is characterized by comprising a data acquisition module, a data analysis module and a result display module, wherein the data acquisition module, the data analysis module and the result display module are communicated through a 5G network;
the data acquisition module uses wearable check out test set to gather real-time data and to the collection to basic auxiliary information, and main body temperature characteristic, breathing characteristic and pulse characteristic of gathering specifically have following index: body temperature, heart rate, respiratory rate, oxygen output and oxygen saturation;
the data analysis module is carried out through a storage and calculation center of a medical institution end, and other index values including inhalation oxygen pressure, arterial oxygen content, inhalation oxygen fraction, oxygen consumption, oxygen use fraction, oxygenation index, respiration index, arterial oxygen content, arterial oxyhemoglobin saturation, oxyhemoglobin saturation and cough frequency are conjectured from the data collected by the data collection module by combining a related calculation formula, so that the risk that a user may suffer from pneumonia or the possibility of suffering from pneumonia is analyzed, and the high risk group is controlled in time;
and the result display module is a use terminal for sending the final analysis result to the user and medical staff in an analysis report form.
2. The apparatus for identifying new coronary pneumonia according to claim 1, wherein the data acquisition module comprises a base module, a signal acquisition module and a signal processing module;
the basic module comprises a power supply module, a storage model and a transmission module;
the signal acquisition module comprises a body temperature detection module, a pulse detection module and a breath detection module;
the signal processing module comprises a filtering module, an amplifying module and an A/D module.
3. The apparatus for identifying new coronary pneumonia according to claim 2, wherein the power supply module is a hardware battery, and provides energy for the operation of the whole apparatus to ensure long-time cruising; the storage model relates to a hardware flash memory chip and caches the acquired data; the transmission module is used as a transmitting end to acquire signal data.
4. The apparatus for identifying new coronary pneumonia based on attention mechanism according to claim 2 is characterized in that the body temperature detection module uses a thermopile infrared temperature sensor to collect body temperature signals; the pulse detection module collects pulse wave signals by using a reflective photoelectric sensor; the respiration detection module relates to hardware, such as a mask, a sampling needle, a photodiode and a signal receiving device, and acquires signals such as oxygen, carbon dioxide gas content, blood oxygen saturation and the like in a non-invasive mode.
5. The apparatus for identifying new coronary pneumonia based on attention mechanism according to claim 2 is characterized in that the filter module is a filter circuit; the amplifying module is a signal amplifying circuit; the A/D module is an analog-to-digital converter which digitizes the received signal to provide to the signal sending module and the storage module.
6. The apparatus for identifying new coronary pneumonia according to claim 1, wherein the auxiliary information of the storage and calculation center is: gender, age, height, weight, history of whether or not the patient has suffered from the pneumonia, time since last illness, history of whether or not the patient has suffered from the flu recently, history of whether or not the patient has suffered from the fever recently, time since the cold last, time since the fever last, place of recent route, history of whether or not the patient has suffered from the fever last, and history of whether or not the patient has suffered from the indirect contact.
7. A new coronary pneumonia identification method based on an attention mechanism is characterized by specifically comprising the following steps:
s1: providing a position tracking service through a 5G network, and tracking the position of user equipment through an LTaaS layer;
s2: the user wears an intelligent helmet with the data input function, and the intelligent helmet is used as wearable remote diagnosis and treatment monitoring equipment, collects self health data at proper time at home and automatically uploads the data;
s3: the intelligent helmet is wirelessly connected with the resident mobile phone through Bluetooth, and the acquired data are transmitted to a 5G network RAN cloud network resource management layer through a mobile phone API;
s4: the storage and calculation center of the medical institution end mines the uploaded data;
s5: and the final analysis result is sent to the user and the doctor terminal through the 5G network.
8. The method for identifying new crown pneumonia based on attention mechanism as claimed in claim 7, wherein the storage and computation center data mining process specifically includes the following steps:
s4.1: the process begins with auxiliary data, i.e., historical condition information and data collected by the equipment as inputs;
s4.2: preprocessing, including missing value processing, abnormal value elimination, continuous feature standardization and class feature Onehot coding, setting a batch _ masks matrix, wherein the value of the position with the feature is 1, the values of other positions are 0, the later calculation of Attention is performed, and the attentions of the rest positions without the feature are changed into 0;
s4.3: firstly, performing primary feature extraction on original features by using an Encoder Encoder, specifically splicing preprocessed data, and then transmitting the preprocessed data into a CNNEncoder, wherein the CNNEncoder defines 3 convolution kernels with the sizes of [2, 10], [3, 10], [4, 10], namely all output puthannels are 10 dimensions, performing data compression on a convolved result through a pooling layer, and finally splicing the obtained 3 vectors in a second dimension to obtain data which is recorded as sent _ reps;
s4.4: transmitting the send _ reps into an LSTMEncoder, wherein the LSTMEncoder is a 2-layer bidirectional LSTM structure, multiplying the LSTMEncoder by a mask according to positions, changing the position of the characteristic loss into 0, and recording the final output data as batch _ hiddens;
s4.5: the Attention is introduced into a model for pneumonia identification, and when a certain characteristic changes, the weights of other associated characteristics change: taking the batch _ hiddens and batch _ masks as the input of the Attention; in the Attention, firstly copying one part of batch _ hiddens to obtain key through linear change, keeping the dimensionality unchanged, and additionally copying two parts of batch _ hiddens to obtain query and value, wherein the query is used for calculating similarity with keys, and the value is used for weighting with the calculated Attention weight to obtain final output; multiplying the key and the query to obtain outputs which represent the weight distributed to each feature; next, performing softmax processing on the atttion, setting the weight without the characteristic to-1 e32 by using batch _ masks, and marking the obtained result as masked _ attn _ scores; finally, multiplying the masked _ attn _ scores with the value to obtain the batch _ outputs;
s4.6: obtaining a vector of classification probability through a full-connection FC layer;
s4.7: and predicting the new data by the trained model to obtain a judgment result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110222978.2A CN114983352A (en) | 2021-03-01 | 2021-03-01 | Method and device for identifying new coronary pneumonia based on attention mechanism |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110222978.2A CN114983352A (en) | 2021-03-01 | 2021-03-01 | Method and device for identifying new coronary pneumonia based on attention mechanism |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114983352A true CN114983352A (en) | 2022-09-02 |
Family
ID=83018824
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110222978.2A Pending CN114983352A (en) | 2021-03-01 | 2021-03-01 | Method and device for identifying new coronary pneumonia based on attention mechanism |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114983352A (en) |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109480823A (en) * | 2018-11-27 | 2019-03-19 | 浙江远图互联科技股份有限公司 | A kind of electrocardiographicdata data acquisition system |
CN110085327A (en) * | 2019-04-01 | 2019-08-02 | 东莞理工学院 | Multichannel LSTM neural network Influenza epidemic situation prediction technique based on attention mechanism |
CN111248879A (en) * | 2020-02-20 | 2020-06-09 | 电子科技大学 | Hypertension old people activity analysis method based on multi-mode attention fusion |
CN111370084A (en) * | 2020-02-07 | 2020-07-03 | 山东师范大学 | Electronic health record representation learning method and system based on BilSTM |
CN111413492A (en) * | 2020-04-30 | 2020-07-14 | 山东大学 | Method and system for detecting novel coronavirus COVID-2019 pneumonia |
CN111603144A (en) * | 2020-04-08 | 2020-09-01 | 中彦医疗科技有限责任公司 | Novel rapid primary screening detector and detection method for coronavirus infected patient |
CN111695779A (en) * | 2020-05-14 | 2020-09-22 | 华南师范大学 | Knowledge tracking method, knowledge tracking device and storage medium |
CN111714117A (en) * | 2020-05-07 | 2020-09-29 | 北京中科千寻科技有限公司 | Electrocardiosignal abnormality detection method based on time sequence depth model |
CN112037239A (en) * | 2020-08-28 | 2020-12-04 | 大连理工大学 | Text guidance image segmentation method based on multi-level explicit relation selection |
CN112037925A (en) * | 2020-07-29 | 2020-12-04 | 郑州大学第一附属医院 | LSTM algorithm-based early warning method for newly-released major infectious diseases |
CN112163705A (en) * | 2020-09-29 | 2021-01-01 | 重庆邮电大学 | Epidemic situation prediction and early warning method based on CNN + SEIR and LSTM |
WO2021032219A2 (en) * | 2019-08-20 | 2021-02-25 | 山东众阳健康科技集团有限公司 | Method and system for disease classification coding based on deep learning, and device and medium |
CN112420176A (en) * | 2019-08-22 | 2021-02-26 | 浙江远图互联科技股份有限公司 | Hierarchical diagnosis guide system based on structured information base |
-
2021
- 2021-03-01 CN CN202110222978.2A patent/CN114983352A/en active Pending
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109480823A (en) * | 2018-11-27 | 2019-03-19 | 浙江远图互联科技股份有限公司 | A kind of electrocardiographicdata data acquisition system |
CN110085327A (en) * | 2019-04-01 | 2019-08-02 | 东莞理工学院 | Multichannel LSTM neural network Influenza epidemic situation prediction technique based on attention mechanism |
WO2021032219A2 (en) * | 2019-08-20 | 2021-02-25 | 山东众阳健康科技集团有限公司 | Method and system for disease classification coding based on deep learning, and device and medium |
CN112420176A (en) * | 2019-08-22 | 2021-02-26 | 浙江远图互联科技股份有限公司 | Hierarchical diagnosis guide system based on structured information base |
CN111370084A (en) * | 2020-02-07 | 2020-07-03 | 山东师范大学 | Electronic health record representation learning method and system based on BilSTM |
CN111248879A (en) * | 2020-02-20 | 2020-06-09 | 电子科技大学 | Hypertension old people activity analysis method based on multi-mode attention fusion |
CN111603144A (en) * | 2020-04-08 | 2020-09-01 | 中彦医疗科技有限责任公司 | Novel rapid primary screening detector and detection method for coronavirus infected patient |
CN111413492A (en) * | 2020-04-30 | 2020-07-14 | 山东大学 | Method and system for detecting novel coronavirus COVID-2019 pneumonia |
CN111714117A (en) * | 2020-05-07 | 2020-09-29 | 北京中科千寻科技有限公司 | Electrocardiosignal abnormality detection method based on time sequence depth model |
CN111695779A (en) * | 2020-05-14 | 2020-09-22 | 华南师范大学 | Knowledge tracking method, knowledge tracking device and storage medium |
CN112037925A (en) * | 2020-07-29 | 2020-12-04 | 郑州大学第一附属医院 | LSTM algorithm-based early warning method for newly-released major infectious diseases |
CN112037239A (en) * | 2020-08-28 | 2020-12-04 | 大连理工大学 | Text guidance image segmentation method based on multi-level explicit relation selection |
CN112163705A (en) * | 2020-09-29 | 2021-01-01 | 重庆邮电大学 | Epidemic situation prediction and early warning method based on CNN + SEIR and LSTM |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108717678B (en) | Wisdom endowment system | |
CN102908130B (en) | Device for monitoring human health | |
CN107103182A (en) | A kind of heart disease Warning System and method based on deep learning algorithm | |
CN208942118U (en) | A kind of COPD reconditioning intelligent monitor wearable system based on Internet of Things | |
CN102054118A (en) | Remote medical system | |
CN109949934A (en) | Calculation method for evaluating body health state by utilizing AI algorithm | |
CN117617921B (en) | Intelligent blood pressure monitoring system and method based on Internet of things | |
CN116098595B (en) | System and method for monitoring and preventing sudden cardiac death and sudden cerebral death | |
CN113012823A (en) | Cloud platform system for electrocardiograph | |
CN110141213A (en) | Electrocardio monitoring method, device and computer readable storage medium | |
Arunkumar et al. | Wearable Devices for Patient Monitoring System using IoT | |
CN107041740A (en) | Animal heart rate monitoring system and the Heart Rate States recognition methods based on neutral net | |
CN117017301A (en) | Wearable electrocardiograph monitoring alarm system and method | |
CN110811638A (en) | SVM classifier construction method, system and method for monitoring sleep | |
CN113393936B (en) | Disabled elderly health monitoring system based on end-edge-cloud architecture | |
CN110428893A (en) | A kind of tele-medicine vital sign data acquisition device | |
Elango et al. | Super artificial intelligence medical healthcare services and smart wearable system based on IoT for remote health monitoring | |
CN112037916A (en) | Shared multifunctional sudden death prevention physiological information detection system and method thereof | |
CN111671402A (en) | Health data and stroke positioning system | |
CN114983352A (en) | Method and device for identifying new coronary pneumonia based on attention mechanism | |
CN106709260A (en) | Understanding doctor type doctor-patient online interactive method | |
CN116741375A (en) | Health data management method and device for high-altitude hydropower station constructors | |
Zhao et al. | Intelligent Gateway Based Human Cardiopulmonary Health Monitoring System | |
CN216257073U (en) | Chest pain on-line monitoring system based on intelligent bracelet | |
US20210290180A1 (en) | Method and system for quickly detecting an abnormal concentration of potassium ions in blood from an electrocardiogram |
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 |