CN116453696B - Respiratory tract disease infection risk prediction method based on personnel space-time distribution model - Google Patents
Respiratory tract disease infection risk prediction method based on personnel space-time distribution model Download PDFInfo
- Publication number
- CN116453696B CN116453696B CN202310700348.0A CN202310700348A CN116453696B CN 116453696 B CN116453696 B CN 116453696B CN 202310700348 A CN202310700348 A CN 202310700348A CN 116453696 B CN116453696 B CN 116453696B
- Authority
- CN
- China
- Prior art keywords
- time
- space
- diagnosis
- respiratory tract
- personnel
- 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.)
- Active
Links
- 208000015181 infectious disease Diseases 0.000 title claims abstract description 74
- 238000000034 method Methods 0.000 title claims abstract description 57
- 208000018569 Respiratory Tract disease Diseases 0.000 title claims abstract description 36
- 238000003745 diagnosis Methods 0.000 claims abstract description 70
- 238000010801 machine learning Methods 0.000 claims abstract description 17
- 230000008859 change Effects 0.000 claims abstract description 14
- 238000012544 monitoring process Methods 0.000 claims description 31
- 206010057190 Respiratory tract infections Diseases 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 6
- HEMHJVSKTPXQMS-UHFFFAOYSA-M Sodium hydroxide Chemical compound [OH-].[Na+] HEMHJVSKTPXQMS-UHFFFAOYSA-M 0.000 claims description 4
- 238000010219 correlation analysis Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 4
- 230000006698 induction Effects 0.000 claims description 4
- 230000029058 respiratory gaseous exchange Effects 0.000 claims description 4
- 238000009423 ventilation Methods 0.000 claims description 4
- 238000010832 independent-sample T-test Methods 0.000 claims description 3
- 238000001543 one-way ANOVA Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000012502 risk assessment Methods 0.000 abstract description 9
- 238000012423 maintenance Methods 0.000 abstract description 2
- 238000011156 evaluation Methods 0.000 abstract 1
- 238000004364 calculation method Methods 0.000 description 12
- 208000023504 respiratory system disease Diseases 0.000 description 10
- 238000013527 convolutional neural network Methods 0.000 description 8
- 230000033228 biological regulation Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 2
- 230000010485 coping Effects 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000004134 energy conservation Methods 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000011835 investigation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000010187 selection method Methods 0.000 description 2
- 206010011409 Cross infection Diseases 0.000 description 1
- 206010029803 Nosocomial infection Diseases 0.000 description 1
- 238000004378 air conditioning Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002902 bimodal effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000013468 resource allocation Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
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
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
- G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion
-
- 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
-
- 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
Abstract
The application discloses a respiratory tract disease infection risk prediction method based on a personnel space-time distribution model, which comprises the following steps: acquiring real-time diagnosis information and diagnosis space data; inputting the diagnosis information and the diagnosis space data into a trained prediction model based on the space-time distribution of the staff established by machine learning to obtain the real-time number of people and the residence time of the staff; and according to the obtained residence time of the real-time number of people and the doctor, adopting a preset evaluation formula to evaluate the infection risk of the respiratory tract diseases in the waiting area. The application aims to solve the problem of uncertainty of infection risk assessment caused by complex personnel change in a waiting area of a hospital so as to improve the flexibility and accuracy of the infection risk assessment, thereby guiding reasonable allocation of diagnosis and treatment resources of the hospital and accurate operation and maintenance of an air conditioner fresh air system in the waiting area.
Description
Technical Field
The application relates to the technical field of disease risk prediction based on medical big data, in particular to a respiratory tract disease infection risk prediction method based on a personnel space-time distribution model.
Background
Hospitals are the forefront departments of respiratory disease discovery and reporting, but as the central treatment department, hospitals are also the primary sites of cross-infection and aggregate morbidity. Particularly for waiting areas, the high density, frequent flow and long exposure of personnel distribution features increases the risk of airborne infections within them and at the same time exacerbates the difficulty of risk assessment. At present, the existing researches are mainly based on fixed personnel distribution given by specifications or real-time personnel number development infection risk assessment based on monitoring videos and sensor monitoring, influence of time fluctuation and changes of the number of patients and waiting time caused by natural characteristic differences of departments is ignored, seasonal fluctuation characteristics of respiratory diseases are considered, and the existing researches are not considered.
Existing patents currently approaching the present application are as follows:
1. the application relates to an intelligent air quantity control system and a control method (CN 115789904A) for inhibiting the spreading risk of new crown infection, which are used for obtaining the number of indoor personnel in real time through an image acquisition device so as to realize fine control of fresh air unit equipment according to the minimum required fresh air quantity. However, the method ignores the influence of the residence time of the personnel in the room on the infection risk calculation, and the risk calculation based on the real-time personnel number monitoring cannot predict the prepositive risk, namely, the control of fresh air has hysteresis.
2. The application relates to an air treatment system (CN 115789885A) for a large range, which is used for acquiring indoor personnel distribution conditions by combining image recognition and an infrared sensor and realizing the regulation and control of air treatment equipment such as a fresh air handling unit. However, the processing module of the indoor personnel information needs to rely on introducing an additional sensor to process the image information, so that the building input cost is additionally increased, and the regulation and control principle of the method also depends on feedback data of the sensor and the video, namely the possibility of excessively high indoor infection risk caused by the fact that feedforward control cannot be realized.
3. The application relates to a respiratory tract infection risk prediction system (CN 114360741A) for a hospital waiting area, which takes the aspect ratio, the spatial layout, the air supply form, the air exchange rate and the number of real-time people in a typical waiting area as input parameters, performs infection risk assessment by constructing a prediction model, and guides the regulation of the air supply end. Although the method realizes infection risk calculation based on real-time people number prediction, the method also ignores the influence of the residence time on the infection risk calculation, especially for places with strong personnel time variability in a waiting area of a hospital. Moreover, relying on real-time head-count predictions of video information can seriously threaten the privacy of hospital patients.
In summary, the application results of the method for predicting the infection risk of the waiting area of the hospital, which takes account of the number of people in real time and the residence time, are not found at present, and the research on the infection risk prediction by the existing method still depends on the method for monitoring videos or adding additional sensors for real-time monitoring, so that the influence of residence time is ignored, and the construction investment cost and the privacy safety of patients are greatly influenced. Under the background, the application provides a real-time dynamic prediction method for respiratory tract disease infection risk in a waiting area based on treatment information.
Disclosure of Invention
Therefore, the application aims to provide a respiratory tract disease infection risk prediction method based on a personnel space-time distribution model, which solves the problem of uncertainty of infection risk assessment caused by complex personnel change in a waiting area of a hospital, so as to improve flexibility and accuracy of the infection risk assessment, thereby guiding reasonable allocation of diagnosis and treatment resources of the hospital and accurate operation and maintenance of an air conditioner fresh air system in the waiting area.
In order to achieve the above object, the method for predicting the infection risk of the respiratory tract disease based on the personnel space-time distribution model comprises the following steps:
s1, acquiring real-time diagnosis information and diagnosis space data;
s2, inputting the diagnosis information and the diagnosis space data into a trained prediction model based on the space-time distribution of the staff established by machine learning to obtain the residence time of the real-time number of people and the diagnosis staff;
s3, according to the obtained real-time number of people and residence time of the doctor, evaluating the infection risk of the respiratory tract diseases in the waiting area by adopting the following formula:
；
wherein ,is the number of people in real time indoor; />Probability of infection for airborne induction;qquantum generation rate (quata/h) per unit of infected person;plung respiration amount (m) 3 /h)；/>Exposure time (h); />For the total area (m) 2 )；BIs the initial infection rate; />Is the effectiveness of air distribution; />For minimum ventilation times (h -1 )；/>For the total volume (m) 3 )。
Further preferably, in S1, the diagnosis information includes time variable, department information and current state of respiratory disease; the visit spatial data includes a visit area plane parameter.
Further preferably, in S2, the prediction model of the spatial-temporal distribution of the person is trained by the following steps:
s201, acquiring historical diagnosis information, diagnosis space data and diagnosis personnel change data in a monitoring video corresponding to the historical diagnosis time;
s202, extracting first key features from historical visit information and visit space data, and extracting second key features from change data of visit personnel of a monitoring video; taking the first key feature as an input variable and the second key feature as an output variable; training a prediction model of the space-time distribution of the personnel established based on machine learning; obtaining the trained model parameters.
Further preferably, the extracting the second key feature includes the following steps:
collecting images of consultants in a consultation area; dividing a data set by taking time as an index for the acquired images of the patients according to the diagnosis time in the diagnosis information acquired in the step S1;
counting the number of people in the images of the consultants in the consultation area by using the trained CNN model, and predicting the number of people in real time;
and randomly selecting patients from the monitoring video images for tracking by taking fixed time intervals as a period, respectively recording specific moments of tracking the patients entering a waiting area and entering a consulting room, and taking the two moment intervals as the stay time of the corresponding patients.
Further preferably, when the first key feature and the second key feature are extracted, the method further comprises selecting feature variables affecting real-time people number and residence time of a treatment area in treatment information through differential analysis and correlation analysis, wherein the selection method comprises any one of the following steps: independent sample T-test, one-way anova and pearson correlation test.
The application also provides a respiratory tract disease infection risk prediction system based on the personnel space-time distribution model, which comprises the following steps: the system comprises a diagnosis information acquisition module, a prediction model of personnel space-time distribution and a respiratory tract disease infection risk module in a waiting area;
the diagnosis information acquisition module is used for acquiring real-time diagnosis information and diagnosis space data;
the prediction model of the personnel space-time distribution is used for predicting the residence time of the number of people in real time and the personnel in the doctor by taking the acquired doctor information and the doctor space data as input variables and utilizing the trained network parameters;
the respiratory tract disease infection risk module in the waiting area is used for dynamically evaluating the respiratory tract disease infection risk in the waiting area in real time according to the real-time number and the residence time of the visiting staff obtained by the prediction model of the space-time distribution of the staff.
Further preferably, in the diagnosis information acquisition module, the diagnosis information includes time variable, department information and current state of respiratory disease, and the diagnosis space data includes a diagnosis area plane parameter.
Further preferably, the system further comprises an image acquisition module, wherein the image acquisition module is used for acquiring the change data of the doctor in the monitoring video corresponding to the historical doctor information and the doctor space data.
Further preferably, the training of the prediction model of the personnel space-time distribution includes:
extracting first key features from historical visit information and visit space data, and extracting second key features from visit personnel change data of the monitoring video; taking the first key feature as an input variable and the second key feature as an output variable; training a prediction model of the space-time distribution of the personnel established based on machine learning; obtaining trained model parameters;
the second key feature extraction method comprises the following steps:
collecting images of consultants in a consultation area; dividing a data set by taking time as an index for the acquired images of the patients according to the acquired times of the patients in the information;
counting the number of people in the images of the consultants in the consultation area by using the trained CNN model, and predicting the number of people in real time;
and randomly selecting patients from the monitoring video images for tracking by taking fixed time intervals as a period, respectively recording specific moments of tracking the patients entering a waiting area and entering a consulting room, and taking the two moment intervals as the stay time of the corresponding patients.
The application also provides a computer storage medium, wherein the computer storage medium is stored with a computer program, and the computer program realizes the steps of the respiratory tract disease infection risk prediction method based on the personnel space-time distribution model when being executed by a processor.
Compared with the prior art, the respiratory tract disease infection risk prediction method based on the personnel space-time distribution model provided by the application has at least the following advantages:
the real-time infection risk calculation and prediction are realized by adopting the treatment data in the existing treatment information system of the hospital, and the indoor respiratory tract disease infection risk under the double influences of the number of real-time people and the residence time is comprehensively considered. Compared with the method for acquiring personnel characteristic data to perform infection risk calculation by adding a specific sensor, the method can realize prepositive prediction of respiratory disease incidence rules of different periods throughout the day and even different seasons throughout the year of each department while reducing the initial investment cost of a building, thereby guiding early warning and coping strategy formulation of a hospital and promoting construction of a good hospital health diagnosis and treatment environment.
Compared with the method for predicting the personnel characteristics in real time through the monitoring video, which is widely used at present, the personnel distribution prediction based on the treatment data does not involve a large amount of patient privacy information leakage caused by the processing of the monitoring video, and is particularly sensitive in hospital places. Although the method also needs to verify the accuracy of the prediction model based on the monitoring video data in the early stage, the video information can be separated from the video information in the use stage in the later stage, and the infection risk prediction can be directly completed only through the information system data of the doctor, so that the safety of the data in the use process is ensured.
The application can guide the establishment of a hospital respiratory disease coping mechanism based on dynamic prediction data aiming at infection risk. On one hand, hospitals can make medical resource allocation strategies in advance according to different levels of predicted infection risks, so that reasonable allocation of medical resources is realized; on the other hand, the prediction of the real-time infection risk can further guide the fine regulation and control of the indoor air conditioning system, and the real-time control of the infection risk is realized by providing the indoor fresh air quantity which dynamically changes, so that the safety and the health of the indoor environment are more accurately ensured, and the energy consumption waste caused by the excessive use of the fresh air is avoided.
Drawings
FIG. 1 is a workflow diagram of a respiratory tract disease infection risk prediction method based on a human spatiotemporal distribution model of the present application.
Fig. 2 is a graph showing a comparison between a real-time number of persons predicted based on the machine learning model and a predicted value in example 2.
Fig. 3 is a graph showing a comparison between the actual value and the predicted value of the residence time predicted based on the machine learning model in example 2.
Fig. 4 is a plot of a waiting area respiratory tract infection risk prediction based on the visit information.
FIG. 5 is a graph of the results of the accuracy of the prediction of method I.
FIG. 6 is a graph of the results of the prediction accuracy of method II.
Fig. 7 is a graph of the result of the accuracy of the prediction of method iii.
Fig. 8 is a result diagram of the prediction accuracy of the present method.
Detailed Description
The application is described in further detail below with reference to the drawings and the detailed description.
As shown in fig. 1, a respiratory tract disease infection risk prediction method based on a human space-time distribution model according to an embodiment of the present application includes the following steps:
s1, acquiring diagnosis information and diagnosis space data; in this embodiment, the visit information includes time variable, department information, and current status of respiratory disease. Specifically, the time variable includes, but is not limited to, date, time of day, and the like. The department information comprises department categories and open clinic numbers. The current state of respiratory disease, i.e. whether it is currently the outbreak. The visit spatial data includes a visit area plane parameter.
S2, inputting the diagnosis information and the diagnosis space data into a trained prediction model based on the space-time distribution of the staff established by machine learning to obtain the residence time of the real-time number of people and the diagnosis staff;
it should be noted that, the prediction model of the personnel space-time distribution is trained by adopting the following processes:
s201, acquiring historical diagnosis information, diagnosis space data and diagnosis personnel change data in a monitoring video corresponding to the historical diagnosis time;
s202, extracting first key features from historical visit information and visit space data, and extracting second key features from change data of visit personnel of a monitoring video; taking the first key feature as an input variable and the second key feature as an output variable; training a prediction model of the space-time distribution of the personnel established based on machine learning; obtaining the trained model parameters.
The second key feature extraction process comprises the following steps:
collecting images of consultants in a consultation area; and (3) dividing the acquired image of the doctor by taking the time as an index according to the doctor-making time in the doctor-making information acquired in the step (S1).
Counting the number of people in the images of the consultants in the consultation area by using the trained CNN model, and predicting the number of people in real time; when the CNN model is trained, people in the images need to be selected in the data set of the corresponding time period to be marked, and the marked images are converted into density images; calculating the number of people in real time by using the density map; the CNN model can automatically identify the number of people in the image after training.
And randomly selecting patients from the monitoring video images for tracking by taking fixed time intervals as a period, respectively recording specific moments of tracking the patients entering a waiting area and entering a consulting room, and taking the two moment intervals as the stay time of the corresponding patients.
The method also comprises the step of selecting characteristic variables affecting the real-time number and the residence time of the treatment area in the treatment information through differential analysis and correlation analysis, wherein the selection method comprises any one of the following steps: independent sample T-test, one-way anova and pearson correlation test.
According to the data, a machine learning method is used for establishing a personnel distribution prediction model, input variables of the personnel distribution prediction model comprise time variables, department information and current states of respiratory diseases, real-time people and stay time are set as output variables, the real-time people count marked people in the density map by using a trained CNN model through converting marked images into the density map, and the predicted real-time people are obtained; the obtained residence time of the real-time number of people and the doctor-seeing person; and inputting the parameters into a personnel distribution prediction model, and performing iterative learning according to machine learning to obtain trained neural network parameters.
S3, according to the obtained real-time number of people and residence time of the doctor, evaluating the infection risk of the respiratory tract diseases in the waiting area by adopting the following formula:
the infection risk assessment method is based on obtaining dynamic personnel distribution prediction results, and correcting real-time indoor real-time people in a Wells-Riley correction model) And residence time (+)>) And key parameters are adopted, so that the dynamic prediction of respiratory tract disease infection risk in a waiting area based on the treatment information is realized. The calculation formula of the Wells-Riley correction model is as follows:
(equation 1)
wherein ,is the number of people in real time indoor; />Probability of infection for airborne induction;qquantum generation rate (quata/h) per unit of infected person;plung respiration amount (m) 3 /h)；/>Exposure time (h); />For the total area (m) 2 )；BIs the initial infection rate; />Is the effectiveness of air distribution; />For minimum ventilation times (h -1 )；/>Waiting for each departmentTotal volume of zones (m) 3 )。
The application also provides a respiratory tract disease infection risk prediction system based on the personnel space-time distribution model, which comprises the following steps: the system comprises a diagnosis information acquisition module, a prediction model of personnel space-time distribution and a respiratory tract disease infection risk module in a waiting area;
the diagnosis information acquisition module is used for acquiring real-time diagnosis information and diagnosis space data;
the prediction model of the personnel space-time distribution is used for predicting the residence time of the real-time number of people and the personnel to be treated by taking the acquired information and the space data to be treated as input variables and utilizing the trained network parameters;
the respiratory tract disease infection risk module in the waiting area is used for dynamically evaluating the respiratory tract disease infection risk in the waiting area in real time according to the residence time of the real-time number and the visiting personnel obtained by the prediction model of the space-time distribution of the personnel.
Further preferably, in the diagnosis information acquisition module, the diagnosis information includes time variable, department information and current state of respiratory disease, and the diagnosis space data includes a diagnosis area plane parameter. Plane parameters include dimensions, plane layout, etc.
The system further preferably comprises an image acquisition module, wherein the image acquisition module is used for acquiring the change data of the doctor in the monitoring video corresponding to the historical doctor information and the doctor space data; extracting first key features from historical visit information and visit space data, and extracting second key features from visit personnel change data of the monitoring video; taking the first key feature as an input variable and the second key feature as an output variable; training a prediction model of the space-time distribution of the personnel established based on machine learning; obtaining trained model parameters;
the second key feature extraction method comprises the following steps:
collecting images of consultants in a consultation area; dividing a data set by taking time as an index for the acquired images of the patients according to the acquired times of the patients in the information;
selecting characters in the images in the data sets of the corresponding time periods to mark, and converting the marked images into density maps;
and randomly selecting patients from the monitoring video images for tracking by taking fixed time intervals as a period, respectively recording specific moments of tracking the patients entering a waiting area and entering a consulting room, and taking the two moment intervals as the stay time of the corresponding patients.
Further preferably, a dynamic personnel distribution prediction result is obtained according to a personnel space-time distribution prediction model; the risk of respiratory tract disease infection in the waiting area is assessed by the following formula:
(equation 1)
wherein ,is the number of people in real time indoor; />Probability of infection for airborne induction;qquantum generation rate (quata/h) per unit of infected person;plung respiration amount (m) 3 /h)；/>Exposure time (h); />For the total area (m) 2 )；BIs the initial infection rate; />Is the effectiveness of air distribution; />For minimum ventilation times (h -1 )；/>For the total volume (m) 3 )。
The application also provides a computer storage medium, wherein the computer storage medium is stored with a computer program, and the computer program realizes the steps of the respiratory tract disease infection risk prediction method based on the personnel space-time distribution model when being executed by a processor.
When the scheme is actually implemented and verified, 10 department waiting areas in a comprehensive hospital outpatient building such as a third class A in a cold area are selected to explain the embodiment of the application. As shown in Table 1, the building information characteristics of 10 investigated departments and their waiting areas are listed, the whole waiting area is composed of waiting area and waiting corridor, and the total area is 89 m 2 To 141 m 2 The clear heights are 3.4 and m.
The waiting area is a space where patients and accompanying persons mainly concentrate on waiting for the treatment, and is provided with service facilities such as a seat, a nurse station and the like, so that the patients and accompanying persons can sit for rest and problem consultation during the waiting period; waiting lobbies are traffic spaces shared by doctors and patients going to the office.
The details are shown in table 1 below:
table 1: specific information summary table of 10 departments of selected hospitals in example
The entire investigation was started on day 1 of 12 of 2021 for a total of two months. The hospital visit information system resource platform is used for collecting data information, after the registered patient data which are not subjected to the diagnosis are removed, 113393 pieces of registration and diagnosis record data in the whole investigation period are obtained, and the time variable, the department category and the current state variable information of the hospital during the diagnosis are counted for each piece of diagnosis data on the basis.
When training a prediction model of personnel space-time distribution, selecting 17 video monitoring points covering the whole waiting area in 10 departments to collect video data information, intercepting monitoring photos at time intervals of 1 minute, establishing a CNN model, and identifying real-time people in the waiting area in the intercepted images by using the trained CNN model; and marking the time when the person enters the waiting area for the first time and the time when the person leaves the area and enters the consulting room to obtain the residence time of the person in the waiting area. In order to protect privacy of patients and medical staff in hospitals, the monitoring video data are collected and processed in a monitoring center of the hospitals according to the requirements of the hospital. Finally, according to the monitoring video information of the 10 waiting areas collected during the monitoring period, 37820 crowd counting results and the stay time results of 23326 patients are accumulated and collected.
And summarizing the obtained data, and respectively carrying out difference and correlation analysis on the time variable, the department category and the current state variable of the hospital. According to the analysis result, for the present embodiment, the 9 specific variables are key variables affecting the real-time number of people in the waiting area and the waiting time. As shown in table 2 below;
TABLE 2 Key characteristic variables affecting indoor real-time population and residence time
Further, the time period, holiday, week, time period, time before starting diagnosis, time, department, open consulting room number and waiting patient number are used as input variables, the real-time number and residence time of the waiting area are used as output parameters, machine learning model establishment is carried out, and the optimal prediction model is obtained through algorithm selection and super parameter adjustment.
Fig. 2 and 3 show the comparison of the actual values of the real-time population and residence time of 10 department waiting areas and the machine learning predicted values, taking the predicted results of a certain day as an example. As can be seen from the graph, the machine learning predicted values and the real values of the real-time population and the residence time of the 10 department waiting areas show similar variation trends, and the average accuracy of the real-time population and the residence time of the established model in predicting the waiting areas of the departments is 73.31% and 72.66% respectively.
Finally, based on the real-time number and the residence time of the waiting areas of the 10 departments obtained by prediction in the fourth step, the infection risk of each department at different moments in the day is calculated, and the calculation formula of the infection risk is shown in the formula 1. As shown in fig. 4, the infection probability of 10 departments on the same day was kept fluctuating within 5% from the overall point of view, and the fluctuation characteristics of infection risk exhibited a bimodal change pattern. The degree of fluctuation and the probability of infection for different time periods show a significant difference for each department.
On the basis of the results shown in fig. 5-8, the calculation results of the present application are compared with other existing common methods for calculating risk of infection, and the settings of the real-time number of people and the residence time in each method are shown in table 3. The actual value represents the real-time number of people obtained based on the video information and the infection risk assessment result carried out by the residence time, and the comparison process uses the infection risk calculated by the method as a reference to check the accuracy of other methods. Method I-method III is simplified for real-time population and residence time, respectively, according to specifications and residence time provided empirically by hospitals in research, and method IV is the result of calculation according to the method of the application.
Table 3:
summary of 4 methods for predicting infection risk and real-time number of people and residence time
Data sources GB 50189-2015 of public building energy conservation design standard and GB 55015-2021 of general building energy conservation and renewable energy utilization standard.
Fig. 5 to 8 show the relationship between the four risk calculation results and the true value calculation results of 10 departments based on the actual value. Wherein, the point on the real line indicates that the calculated result is consistent with the real value, and the point between the two dotted lines indicates that the absolute error of the calculated result is within + -1%. It can be seen that the method I has the greatest error and the method IV has the least error, and the conventional assignment of the real-time number of people and the residence time according to the relevant standard and experience is unreliable, so that the dynamic prediction of the real-time number of people and the residence time is beneficial to improving the accuracy of indoor infection risk. Among them, the method used in the present application, method IV, has significant advantages over other prediction methods in terms of accuracy.
It is apparent that the above examples are given by way of illustration only and are not limiting of the embodiments. Other variations or modifications of the above teachings will be apparent to those of ordinary skill in the art. It is not necessary here nor is it exhaustive of all embodiments. While still being apparent from variations or modifications that may be made by those skilled in the art are within the scope of the application.
Claims (10)
1. A respiratory tract disease infection risk prediction method based on a personnel space-time distribution model is characterized by comprising the following steps:
s1, acquiring real-time diagnosis information and diagnosis space data;
s2, inputting the diagnosis information and the diagnosis space data into a trained prediction model based on the space-time distribution of the staff established by machine learning to obtain the residence time of the real-time number of people and the diagnosis staff;
s3, according to the obtained real-time number of people and residence time of the doctor, evaluating the infection risk of the respiratory tract diseases in the waiting area by adopting the following formula:
；
wherein ,is the number of people in real time indoor; />Probability of infection for airborne induction;qquantum generation rate (quata/h) per unit of infected person;plung respiration amount (m) 3 /h)；/>Exposure time (h); />For the total area (m) 2 )；BIs the initial infection rate; />Is the effectiveness of air distribution; />For minimum ventilation times (h -1 )；For the total volume (m) 3 )。
2. The method for predicting risk of respiratory tract infection based on a human spatiotemporal distribution model of claim 1, wherein in S1, the visit information includes time variable, department information, and current state of respiratory tract disease; the visit spatial data includes a visit area plane parameter.
3. The respiratory tract infection risk prediction method based on the human space-time distribution model according to claim 2, wherein in S2, the human space-time distribution prediction model is trained by the following steps:
s201, acquiring historical diagnosis information, diagnosis space data and diagnosis personnel change data in a monitoring video corresponding to the historical diagnosis time;
s202, extracting first key features from historical visit information and visit space data, and extracting second key features from change data of visit personnel of a monitoring video; taking the first key feature as an input variable and the second key feature as an output variable; training a prediction model of the space-time distribution of the personnel established based on machine learning; obtaining the trained model parameters.
4. A method for predicting risk of respiratory tract infection based on a human spatiotemporal distribution model according to claim 3, characterized in that said extracting the second key features comprises the steps of:
collecting images of consultants in a consultation area; dividing a data set by taking time as an index for the acquired images of the patients according to the diagnosis time in the diagnosis information acquired in the step S1;
counting the number of people in the images of the consultants in the consultation area by using the trained CNN model, and predicting the number of people in real time;
and randomly selecting patients from the monitoring video images for tracking by taking fixed time intervals as a period, respectively recording specific moments of tracking the patients entering a waiting area and entering a consulting room, and taking the two moment intervals as the stay time of the corresponding patients.
5. The method for predicting risk of respiratory tract infection based on a human space-time distribution model according to claim 3, wherein the step of extracting the first key feature and the second key feature further comprises selecting feature variables affecting the number of people in real time and the residence time in the visit information by differential analysis and correlation analysis, wherein the selection adopts any one of the following methods: independent sample T-test, one-way anova and pearson correlation test.
6. A respiratory tract disease infection risk prediction system based on a human space-time distribution model, which is characterized by being used for the respiratory tract disease infection risk prediction method based on the human space-time distribution model according to any one of claims 1-5 in real time; comprising the following steps: the system comprises a diagnosis information acquisition module, a prediction model of personnel space-time distribution and a respiratory tract disease infection risk module in a waiting area;
the diagnosis information acquisition module is used for acquiring real-time diagnosis information and diagnosis space data;
the prediction model of the personnel space-time distribution is used for predicting the residence time of the real-time number of people and the personnel to be treated by taking the acquired information and the space data to be treated as input variables and utilizing the trained network parameters;
the respiratory tract disease infection risk module in the waiting area is used for dynamically evaluating the respiratory tract disease infection risk in the waiting area in real time according to the residence time of the real-time number and the visiting personnel obtained by the prediction model of the space-time distribution of the personnel.
7. The respiratory tract infection risk prediction system based on the human space-time distribution model according to claim 6, wherein in the diagnosis information acquisition module, the diagnosis information comprises time variable, department information and current respiratory tract disease state, and the diagnosis space data comprises diagnosis area plane parameters.
8. The respiratory tract infection risk prediction system based on the personnel space-time distribution model according to claim 6, further comprising an image acquisition module, wherein the image acquisition module is used for acquiring the variation data of the personnel in the monitoring video corresponding to the historical diagnosis information and the diagnosis space data.
9. The respiratory tract infection risk prediction system based on the human space-time distribution model according to claim 8, wherein the human space-time distribution prediction model comprises, when trained:
extracting first key features from historical visit information and visit space data, and extracting second key features from visit personnel change data of the monitoring video; taking the first key feature as an input variable and the second key feature as an output variable; training a prediction model of the space-time distribution of the personnel established based on machine learning; obtaining trained model parameters;
the second key feature extraction method comprises the following steps:
collecting images of consultants in a consultation area; dividing a data set by taking time as an index for the acquired images of the patients according to the acquired times of the patients in the information;
counting the number of people in the images of the consultants in the consultation area by using the trained CNN model, and predicting the number of people in real time;
and randomly selecting patients from the monitoring video images for tracking by taking fixed time intervals as a period, respectively recording specific moments of tracking the patients entering a waiting area and entering a consulting room, and taking the two moment intervals as the stay time of the corresponding patients.
10. A computer storage medium, wherein a computer program is stored on the computer storage medium, and when executed by a processor, the computer program implements the steps of the respiratory tract disease infection risk prediction method based on the human spatiotemporal distribution model according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310700348.0A CN116453696B (en) | 2023-06-14 | 2023-06-14 | Respiratory tract disease infection risk prediction method based on personnel space-time distribution model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310700348.0A CN116453696B (en) | 2023-06-14 | 2023-06-14 | Respiratory tract disease infection risk prediction method based on personnel space-time distribution model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116453696A CN116453696A (en) | 2023-07-18 |
| CN116453696B true CN116453696B (en) | 2023-08-15 |
Family
ID=87127689
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310700348.0A Active CN116453696B (en) | 2023-06-14 | 2023-06-14 | Respiratory tract disease infection risk prediction method based on personnel space-time distribution model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116453696B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107220482A (en) * | 2017-05-09 | 2017-09-29 | 清华大学 | Respiratory infectious disease risk evaluating system and appraisal procedure |
| EP3557222A1 (en) * | 2017-12-26 | 2019-10-23 | Chongqing Research Institute Co., Ltd. Of China Coal Technology & Engineering Group | Calculation method and testing system for cumulative dust exposure to respirable dust of mine workers |
| CN110782070A (en) * | 2019-09-25 | 2020-02-11 | 北京市交通信息中心 | Urban rail transit emergency passenger flow space-time distribution prediction method |
| CN111652446A (en) * | 2020-06-15 | 2020-09-11 | 深圳前海微众银行股份有限公司 | Method, apparatus and storage medium for predicting risk of infection of infectious disease |
| CN112669971A (en) * | 2020-12-30 | 2021-04-16 | 华南师范大学 | Personal space-time collaborative prediction method based on big data deep learning and robot |
| CN114360741A (en) * | 2022-01-29 | 2022-04-15 | 天津大学 | Respiratory infectious disease infection risk prediction system in waiting area of hospital |
| WO2022167086A1 (en) * | 2021-02-05 | 2022-08-11 | Cbmed Gmbh Center For Biomarker Research In Medicine | Representing a biological image as a grid data-set |
-
2023
- 2023-06-14 CN CN202310700348.0A patent/CN116453696B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107220482A (en) * | 2017-05-09 | 2017-09-29 | 清华大学 | Respiratory infectious disease risk evaluating system and appraisal procedure |
| EP3557222A1 (en) * | 2017-12-26 | 2019-10-23 | Chongqing Research Institute Co., Ltd. Of China Coal Technology & Engineering Group | Calculation method and testing system for cumulative dust exposure to respirable dust of mine workers |
| CN110782070A (en) * | 2019-09-25 | 2020-02-11 | 北京市交通信息中心 | Urban rail transit emergency passenger flow space-time distribution prediction method |
| CN111652446A (en) * | 2020-06-15 | 2020-09-11 | 深圳前海微众银行股份有限公司 | Method, apparatus and storage medium for predicting risk of infection of infectious disease |
| CN112669971A (en) * | 2020-12-30 | 2021-04-16 | 华南师范大学 | Personal space-time collaborative prediction method based on big data deep learning and robot |
| WO2022167086A1 (en) * | 2021-02-05 | 2022-08-11 | Cbmed Gmbh Center For Biomarker Research In Medicine | Representing a biological image as a grid data-set |
| CN114360741A (en) * | 2022-01-29 | 2022-04-15 | 天津大学 | Respiratory infectious disease infection risk prediction system in waiting area of hospital |
Non-Patent Citations (1)
| Title |
|---|
| An hour-ahead predictive control strategy for maximizing natural ventilation in passive buildings based on weather forecasting;Yibo Chen 等;《ELSEVIER》;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116453696A (en) | 2023-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Awada et al. | Ten questions concerning occupant health in buildings during normal operations and extreme events including the COVID-19 pandemic | |
| Pazhoohesh et al. | A satisfaction-range approach for achieving thermal comfort level in a shared office | |
| CN105975740B (en) | A kind of medical system with intelligent diagnostics | |
| Luo | The dynamics and mechanism of human thermal adaptation in building environment: A glimpse to adaptive thermal comfort in buildings | |
| CN113344274A (en) | Hospital emergency spatial layout optimization method and device, computer equipment and storage medium | |
| CN114360741A (en) | Respiratory infectious disease infection risk prediction system in waiting area of hospital | |
| Fan et al. | Research on risk scorecard of sick building syndrome based on machine learning | |
| Zhuang et al. | Probabilistic occupancy forecasting for risk-aware optimal ventilation through autoencoder Bayesian deep neural networks | |
| Aroua et al. | Forecast emergency room visits–a major diagnostic categories based approach | |
| CN116453696B (en) | Respiratory tract disease infection risk prediction method based on personnel space-time distribution model | |
| Verma et al. | Tropical climates and the interplay between IEQ and energy consumption in buildings: A review | |
| JP2008299392A (en) | System for determining number of sickbed in medical institution and method for evaluating medical center using the same system | |
| CN112685823A (en) | Method and device for measuring and calculating area of comprehensive hospital with thousands of beds and terminal equipment | |
| CN107192098A (en) | Hospital ventilation regulates and controls method and regulator control system | |
| CN116862283A (en) | Method and system for evaluating and controlling environment of aged people | |
| Zhao et al. | Thermal sensation and occupancy-based cooperative control method for multi-zone VAV air-conditioning systems | |
| CN114254249B (en) | Method and system for determining health index of spatial region | |
| CN113606755A (en) | Air conditioner management method based on demand response | |
| Wei et al. | A framework for airborne infection risk assessment based on real-time occupant distribution prediction in outpatient waiting space | |
| Murniati | Sick building syndrome in Indonesia and Singapore: A comparative study | |
| Jovanović et al. | Assessing the sustainability of Serbian school buildings by analyse and syntesis parameters under information deficiency method | |
| CN112086170A (en) | Evaluation system of traditional Chinese medicine big data cloud service platform | |
| CN117196349B (en) | Inpatient bed information intelligent management system | |
| Horne et al. | A public health approach to health needs assessment at the interface of primary care and community development: findings from an action research study | |
| Loftness et al. | Sustainable and Healthy Built Environment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |