CN109409266B - A kind of security incident identification reporting system and security incident identify report method - Google Patents
A kind of security incident identification reporting system and security incident identify report method Download PDFInfo
- Publication number
- CN109409266B CN109409266B CN201811197737.1A CN201811197737A CN109409266B CN 109409266 B CN109409266 B CN 109409266B CN 201811197737 A CN201811197737 A CN 201811197737A CN 109409266 B CN109409266 B CN 109409266B
- Authority
- CN
- China
- Prior art keywords
- image
- security incident
- central server
- patient
- frequency identification
- 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
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K17/00—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations
-
- 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
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Multimedia (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a kind of security incidents to identify reporting system, including multiple cameras, central server, multiple processing terminals and multiple radio-frequency identification readers, central server is connected by internal network with multiple cameras, multiple processing terminals, multiple radio-frequency identification readers.Positive negative sample is established in several ways, and it is preferable to use the sample images of sufferers themselves to carry out classifier training, improve the accuracy of classifier；Meanwhile when for sample rareness, analog sample image can be generated, training data is increased, improves the training effect of classifier.
Description
Technical field
The present invention relates to field of image processings, and in particular to a kind of security incident reporting system and security incident identification report
Method.
Background technique
With internet+medical treatment development, the artificial intelligence process in medical treatment is more and more, but at existing medical events
The overwhelming majority also relies on artificial treatment in when reason, and artificial treatment efficiency is slow, and dangerous；And it is auxiliary even if there is artificial intelligence to carry out
Processing is helped, but it is generally only simple automation, correctness is not high, easily causes wrong report.
And the application automatically analyzes corresponding security incident, automatic collection safety by utilizing image analyzing and processing technology
Event simultaneously reports automatically, facilitates the quick processing of medical safety event, and provide basis for retrospect later.
Summary of the invention
(1) the technical issues of solving
For automated image analysis is carried out in existing medical safety event, internet+medical treatment automation is improved, and lead to
The foundation for crossing sample database improves the accuracy of automatic analysis and judgment.
(2) technical solution and beneficial effect
The object above present invention is achieved by the following technical programs:
A kind of security incident identifies reporting system, including multiple cameras, central server, multiple processing terminals and multiple
Radio-frequency identification reader, central server are read by internal network and multiple cameras, multiple processing terminals, multiple radio frequency identifications
Device is read to be connected；
Multiple cameras can remotely be controlled by central server, and multiple cameras are arranged in hospital bed, ward doorway, examine
Room, operating room and corridor position；Multiple cameras can receive the acquisition instructions of central server, and be carried out according to acquisition instructions
Image Acquisition, and send the image of acquisition to central server；
Multiple radio-frequency identification readers are connect with central server by internal network, and multiple radio frequency identifications are read
Device is mounted on ward, consulting room, operating room and corridor；In the patient with RFID label tag in multiple radio-frequency identification readers
One of them when, one of them in multiple radio-frequency identification readers can identify the RFID label tag, and the RFID of identification is marked
Information is signed by transmission of network to central server, wherein the RFID label tag includes patient information；
Multiple processing terminals are connected with printer, when doctor by one of multiple processing terminals come for patient's Admission
When formality, by the printing of doctor, it includes patient information that printer printing is controlled according to patient information and generates one
RFID label tag, and the RFID label tag is pasted on patient's clothes or bracelet；
Central server, it includes sample database, security incidents to judge system and medical information management system, center clothes
Business device, which can receive the acquisition image of multiple camera transmission and be saved into sample database or send security incident to, to be sentenced
Disconnected system carries out security incident judgement；The FRID label information of multiple radio-frequency identification reader transmission can also be received, and is recorded
By FRID label information obtain patient information and identify the FRID label information radio-frequency identification reader ID, and in
Centre server has recorded the ID of each radio-frequency identification reader and the incidence relation of its installation site in advance；
Security incident judges that system can judge whether there is security incident in image to its received image, and is sentencing
When disconnected generation security incident, image, patient information and the security incident type by security incident warning message, currently judged is sent
To medical information management system；
Medical information management system receives and record security event, and people corresponding to the corresponding patient of query safe event
Member, is then sent to corresponding personnel for security incident, patient information, while security incident is saved and being counted；
Wherein, sample database includes positive sample image and negative sample image, and wherein positive sample image includes that fitting obtains
Multiple simulation positive sample images, wherein fitting obtains multiple simulation positive sample images and is fitted in the following manner:
A) it establishes confrontation and generates network, it includes discriminator and generator which, which generates network, and generator is depth convolution
Neural network comprising three convolutional layers and an output layer, input are random vector and a standard input picture, export and are
One analog image, Plays input picture are generated by calculator simulation；Discriminator is anti-neural network comprising five
Layer, three first layers are convolutional layer, and the 4th layer is full articulamentum, and the last layer is output layer, and input is the simulation of generator output
Image, it is true that analog image is exported by reference to true picture, or simulate；
B) alternative optimization discriminator and generator finally reach balance；
C) network is generated using the confrontation of b) optimization and one standard input picture is generated into a secondary analog image, repeat
The step obtains multiple analog images as multiple simulation positive sample images.
Preferred: security incident judges that system includes following device:
Training module selects positive and negative samples image from sample database, uses the negative sample image and positive sample of selection
The training of this image obtains a classifier；
Received image is inputted the classifier of above-mentioned training, output category result, and is in classification results by judgment module
There are when security incident, send warning message to medical information management system.
Preferred: sample database is established in the following way:
Establish negative sample image: the patient with RFID label tag is after hospital bed, the radio-frequency identification reader of arranging hospital beds
It identifies the RFID label tag, and the FRID label information of identification is passed through into transmission of network to central server；Central server identification
The FRID label information includes out patient information and identify the FRID label information radio-frequency identification reader ID, go forward side by side one
Step ground obtains according to the ID of each radio-frequency identification reader recorded in advance and the incidence relation of its installation site and identifies the RFID
The installation site of the radio-frequency identification reader of label information, then central server judge the patient information identified whether be
It is identified for the first time and whether the installation site of the radio-frequency identification reader is hospital bed, if the patient information identified is for the first time
Identified and the radio-frequency identification reader installation site is hospital bed, then central server sends image acquisition commands and knows to radio frequency
The corresponding camera of installation site of other reader, camera start to acquire image and send the image of acquisition to central service
The image of acquisition is saved as the corresponding negative sample image of patient information to sample database by device, central server；
Establish positive sample image: the image when security incident judges that security incident occurs for system judgement is saved in sample number
According in library and as positive sample image；The positive sample image that security incident occurs is obtained by other image data bases；It is fitted
It is saved in sample database to multiple simulation positive sample images.
A kind of security incident identification report method using security incident identification reporting system, wherein security incident identification
Reporting system includes multiple cameras, central server, multiple processing terminals and multiple radio-frequency identification readers, central server
It is connected by internal network with multiple cameras, multiple processing terminals, multiple radio-frequency identification readers；It is characterized in that, institute
The method of stating includes:
S1, sample database is established:
Establish negative sample image: the patient with RFID label tag is after hospital bed, the radio-frequency identification reader of arranging hospital beds
It identifies the RFID label tag, and the FRID label information of identification is passed through into transmission of network to central server；Central server identification
The FRID label information includes out patient information and identify the FRID label information radio-frequency identification reader ID, go forward side by side one
Step ground obtains according to the ID of each radio-frequency identification reader recorded in advance and the incidence relation of its installation site and identifies the RFID
The installation site of the radio-frequency identification reader of label information, then central server judge the patient information identified whether be
It is identified for the first time and whether the installation site of the radio-frequency identification reader is hospital bed, if the patient information identified is for the first time
Identified and the radio-frequency identification reader installation site is hospital bed, then central server sends image acquisition commands and knows to radio frequency
The corresponding camera of installation site of other reader, camera start to acquire image and send the image of acquisition to central service
The image of acquisition is saved as the corresponding negative sample image of patient information to sample database by device, central server；
Establish positive sample image: the image when security incident judges that security incident occurs for system judgement is saved in sample number
According in library and as positive sample image；The positive sample image that security incident occurs is obtained by other image data bases；It is fitted
It is saved in sample database to multiple simulation positive sample images；
Fitting obtain multiple simulation positive sample images the following steps are included:
A) it establishes confrontation and generates network, it includes discriminator and generator which, which generates network, and generator is depth convolution
Neural network comprising three convolutional layers and an output layer, input are random vector and a standard input picture, export and are
One analog image, Plays input picture are generated by calculator simulation；Discriminator is anti-neural network comprising five
Layer, three first layers are convolutional layer, and the 4th layer is full articulamentum, and the last layer is output layer, and input is the simulation of generator output
Image, it is true that analog image is exported by reference to true picture, or simulate；
B) alternative optimization discriminator and generator finally reach balance；
C) network is generated using the confrontation of b) optimization and one standard input picture is generated into a secondary analog image, repeat
The step obtains multiple analog images as multiple simulation positive sample images；
S2, acquisition patient's realtime graphic:
It is real that security incident judges that system is sent automatically by central server to the camera of the corresponding hospital bed of all patients
When image capture instruction, the security incident that collected realtime graphic is sent to central server judges system by camera, pacifies
Total event judges that system for each secondary realtime graphic, is performed both by following S3 training step and S4 judgment step；
S3, training step:
S31 selects sample image from sample database:
The negative sample image and second predetermined quantity of the first predetermined quantity are selected from the sample database established in S1
Positive sample image: according to the realtime graphic obtained in step S2, the corresponding current patents' information of the realtime graphic is obtained, according to this
Current patents' information selects the corresponding negative sample image of the current patient information and positive sample image from sample database；
S32 uses the step S31 negative sample image selected and positive sample image training classifier；
S4, judgment step:
The realtime graphic obtained in step S2 is input to the classifier that step S3 training obtains and carries out security incident judgement,
When classifier output belongs to security incident, security incident judges that system is issued to the medical information management system of central server
Alarm, while the realtime graphic of acquisition, patient information and security incident type also being sent together, medical information management system is protected
The security incident is deposited, and finds attending physician and the Night shift of patient according to patient information, to attending physician and supervisor's shield
The mobile phone of scholar sends warning information, while sending realtime graphic, patient information and security incident type, attending physician and master together
It manages and protects scholar to judge by realtime graphic, and sends response message to medical information management system.
Preferably, in step S4, if medical information management system does not receive attending physician and supervisor's shield within a certain period of time
The response message of scholar, then by warning information and realtime graphic, patient information and security incident type be sent to the second doctor and
Second nurse also can be transmitted to higher level doctor and higher level nurse.
Preferably, when judging security incident in S4 step, by the corresponding realtime graphic of security incident, patient information,
Security incident type is sent to sample database, and sample database is saved as positive sample image；Meanwhile sentencing when in S4 step
It is disconnected when there is no security incident, realtime graphic, patient information are sent to sample database, sample database is saved as bearing
Sample image.
Preferably, when the quantity of the negative sample image of patient a certain in sample database or positive sample image is more than scheduled
When quantity, the realtime graphic of a certain patient of acquisition is no longer updated to sample database, subsequent simultaneously for a certain patient
The realtime graphic of acquisition is directly classified using the classifier for the patient that training obtained in the past without step S3 is no longer executed,
Directly execute step S4.
Preferably, sample image is selected to specifically include in the step S31:
According to the realtime graphic obtained in step S2, the corresponding current patents' information of the realtime graphic is obtained, according to deserving
Preceding patient information selects the negative sample image and of corresponding first predetermined quantity of the current patient information from sample database
The positive sample image of two predetermined quantities, the negative sample image of the current patents obtained from sample database first and current patents
Positive sample image, if the first quantity of the negative sample image of the acquisition less than the first predetermined quantity, then obtains other patients
The second quantity negative sample quantity so that the first quantity and the second quantity and be equal to the first predetermined quantity；If obtaining just
The third quantity of sample image then obtains from common image database again and simulates positive sample image less than the second predetermined quantity
The positive sample quantity of the 4th quantity of middle random selection so that third quantity and the 4th quantity and be equal to the second predetermined quantity.
Through the above scheme, the application is enabled to reach following technical effect:
1) by radio frequency identification automatic collection realtime graphic, camera intelligent trigger is operated, and can remotely controls
System；
2) sample database is established, and establishes positive negative sample in several ways, and it is preferable to use the samples of sufferers themselves
Image carries out classifier training, improves the accuracy of classifier；Meanwhile when for sample rareness, analog sample can be generated
Image increases training data, improves the training effect of classifier.
3) security incident is automatically analyzed by image analysis processing, and sends warning message automatically to corresponding personnel, thus
So that corresponding personnel are quickly handled, treatment effeciency is improved, safety is also improved.
Detailed description of the invention
Fig. 1 is that security incident of the invention identifies reporting system；
Fig. 2 is that security incident of the invention identifies report method；
Fig. 3 is manual entry security incident interface of the invention；
Fig. 4 is security incident query interface of the invention；
Fig. 5 is that security incident of the invention counts interface.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
As shown in Figure 1, security incident reporting system includes multiple cameras, central server, multiple processing terminals, wherein
Central server is connected by internal network with multiple cameras, multiple processing terminals, and multiple cameras can be central
Server remote control, multiple cameras are located at the positions such as hospital bed, ward doorway, consulting room, operating room and corridor.
Medical intelligent system further includes multiple radio-frequency identification readers, multiple radio-frequency identification reader and central server
It is connected by internal network, and multiple radio-frequency identification readers are mounted on the positions such as ward, consulting room, operating room and corridor.It is multiple
Processing terminal is also connected with printer, when doctor by one of multiple processing terminals come for patient's Admission formality when, lead to
The printing for crossing doctor controls printer printing according to patient information and generates a RFID label tag comprising patient information, and
The RFID label tag is pasted on patient's clothes or bracelet.It is described to have the patient of RFID label tag close to multiple radio frequency identifications
When one of them in reader, one of them in multiple radio-frequency identification readers can identify the RFID label tag, the RFID
Label includes patient information, thus by the FRID label information of identification by transmission of network to central server, central server
Record the ID of the patient information obtained by FRID label information and the radio-frequency identification reader for identifying the FRID label information.
And central server has recorded the ID of each radio-frequency identification reader and the incidence relation of its installation site in advance.
Central server judges system and medical information management system comprising sample database, security incident.
As shown in Fig. 2, security incident judges that system executes security incident judgment method, this method includes establishing sample data
Library, acquisition patient's realtime graphic, training step and judgment step.
S1, sample database is established.
Sample image includes positive sample image and negative sample image, and positive sample image includes the figure that medical safety event occurs
Picture, the medical safety event include that abnormal patient respiratory, patient's revival, patient's falling from bed, catheter manipulation, patient take medicine, have been infused
At etc.；Negative sample image is the image that security incident does not occur.In sample database other than saving sample image, also save
Sample image belongs to positive negative sample {+1, _ 1 }, corresponding patient information, security incident type, and wherein security incident type includes
Patient respiratory is abnormal, patient's revival, patient's falling from bed, catheter manipulation, patient takes medicine, infusion is completed.
When carrying out sample image acquisition, there are two types of sources for the acquisition of negative sample image:
The first, be after patient handles into-hospital procedures, with RFID label tag patient after hospital bed, arranging hospital beds
Radio-frequency identification reader identify the RFID label tag, and by the FRID label information of identification by transmission of network to central service
Device；Central server identifies the patient information that the FRID label information includes and identifies the radio frequency identification of the FRID label information
The ID of reader, and pass is associated with according further to the ID of each radio-frequency identification reader recorded in advance and its installation site
System obtains the installation site for identifying the radio-frequency identification reader of the FRID label information, and then central server judges the identification
Whether patient information out is whether identified for the first time and the radio-frequency identification reader installation site is hospital bed, if the identification
Patient information out is that identified for the first time and the radio-frequency identification reader installation site is hospital bed, then central server sends figure
As the corresponding camera of the installation site of acquisition to radio-frequency identification reader, camera starts to acquire image and by acquisition
Image sends central server to, and the image of acquisition is saved as the corresponding negative sample image of patient information to sample by central server
Database.Preferably, camera image of taken at regular intervals and sends central server at regular intervals, central service
Device saves the images of all acquisitions to sample database as negative sample image, and the image of acquisition is related to patient information
Connection.
This kind of mode directly can also execute negative sample collecting work by doctor or nurse, when specific implementation, handle in patient
After complete staying-in-hospital procedures, one in doctor or the multiple processing terminals of nurse operation sends negative sample image to central server and adopts
Collection instruction, then central server sends image acquisition commands to the camera of disease bed position, and camera is by the figure of acquisition
As being sent to central server, the image of acquisition, patient information are saved in sample database as negative sample by central server
Image, wherein multi collect instruction can be transmitted in doctor or nurse, to acquire the image of patient's different angle, different gestures.
Second, the negative sample of required amount of other patients saved is obtained by the sample database of central server
This image, and the negative sample image of other patients is to be obtained in other patient's Admission formalities according to above-mentioned first way
Negative sample image.
There are three types of sources for the acquisition of positive sample image:
The first, is when the security incident of central server judges that security incident occurs for system judgement, used acquisition
Image, that is to say, that when the image of acquisition is input in security incident judgement system, which judges that system judgement occurs
When security incident, this image acquired is saved in sample database and as positive sample image, while record security event
Type, the type includes that patient respiratory is abnormal, patient's revival, patient's falling from bed, catheter manipulation, patient takes medicine, infusion is completed, together
When the also corresponding patient information of record acquisition image.
Second, due to it is aforementioned the first occur probability it is smaller, thus it is collected occur security incident image compared with
It is few, particularly, judge system initially in use, the positive sample image of above-mentioned first way acquisition is not deposited also in security incident
It is needing to obtain some positive sample images that security incident occurs, and handmarking by public sample database first at this time
The type of patient information and security incident once can also obtain the multiple images for obtaining and security incident occurring by other means,
As positive sample image.
The third, the positive sample image obtained due to above-mentioned second is relatively difficult, and quantity is few, can reduce safe thing
Part judges the accuracy of system, thus we to need on the basis of the positive sample image acquired at second fitting to obtain more
A simulation positive sample image, approximating method are completed using confrontation generation network, method particularly includes:
A) it establishes confrontation and generates network, it includes discriminator and generator which, which generates network, and generator is depth convolution
Neural network comprising three convolutional layers and an output layer, input are random vector and a standard input picture, export and are
One analog image, Plays input picture are generated by calculator simulation；Discriminator is anti-neural network comprising five
Layer, three first layers are convolutional layer, and the 4th layer is full articulamentum, and the last layer is output layer, and input is the simulation of generator output
Image, it is true that analog image is exported by reference to true picture, or simulate.
B) alternative optimization discriminator and generator finally reach balance.
C) network then being generated using the confrontation of optimization and one standard input picture being generated into a secondary analog image, repetition is held
The row step obtains multiple analog images as multiple simulation positive sample images.
S2, acquisition patient's realtime graphic stage
Security incident judges that system is constantly in working condition, corresponding to all patients automatically by central server
The camera of hospital bed sends real time image collection instruction, specifically can polling mode processing, that is to say, that one disease of acquisition every time
The realtime graphic of the patient of bed, successively acquires the realtime graphic of the patient of all hospital beds, then fixed time intervals are successively adopted again
Collect the realtime graphic of all hospital beds, collected realtime graphic is sent to the security incident judgement system of central server by camera
System, security incident judge that system for each secondary realtime graphic, is performed both by following training step and judgment step.
S3, in the training stage comprising following steps:
S31 selects sample image from sample database.
The negative sample image and second predetermined quantity of the first predetermined quantity are selected from the sample database established in S1
Positive sample image:
According to the realtime graphic obtained in step S2, the corresponding current patents' information of the realtime graphic is obtained, according to deserving
Preceding patient information selects the corresponding negative sample image of the current patient information and positive sample image from sample database, i.e., excellent
It first selects the negative sample image of current patents obtained in the first manner in sample database and obtains in the first manner
The positive sample image of current patents.If the first quantity of the negative sample image obtained in the first manner is less than the first predetermined number
Amount, then obtain the negative sample quantity of the second quantity in the second again so that the first quantity and the second quantity and be equal to the
One predetermined quantity；If the third quantity of the positive sample image obtained in the first manner is less than the second predetermined quantity, then with
Two kinds of modes and the third mode obtain the positive sample quantity of the 4th quantity at random, so that third quantity and the sum of the 4th quantity wait
In the second predetermined quantity.
S32, using the step S31 negative sample image selected and positive sample image training classifier, which can be with
It is carried out by the way of deep neural network.
Preferably, respective type is respectively trained according to negative sample image and the corresponding security incident type of positive sample image
Security event classification device.
S4, judgment step:
The realtime graphic obtained in step S2 is input to the classifier that step S3 training obtains and carries out security incident judgement,
When classifier output belongs to security incident, security incident judges that system is issued to the medical information management system of central server
Alarm, while the realtime graphic of acquisition, patient information and security incident type also being sent together, medical information management system is protected
The security incident is deposited, and finds attending physician and the Night shift of patient according to patient information, to attending physician and supervisor's shield
The mobile phone of scholar sends warning information, while sending realtime graphic, patient information and security incident type, attending physician and master together
It manages and protects scholar to judge by realtime graphic, and sends response message to medical information management system, if managing medical information system
System does not receive the response message of attending physician and Night shift within a certain period of time, then by warning information and realtime graphic, trouble
Person's information and security incident type are sent to the second doctor and the second nurse, also can be transmitted to higher level doctor and higher level nurse.
Preferably, when judging security incident in S4 step, by the corresponding realtime graphic of security incident, patient information,
Security incident type etc. is sent to sample database, and sample database is saved as positive sample image；Meanwhile when in S4 step
When there is no security incident in judgement, realtime graphic, patient information etc. can also be sent to sample database, sample database by its
Save as negative sample image.
Preferably, when the quantity of the negative sample image of patient a certain in sample database or positive sample image is sufficiently large,
Such as when more than scheduled quantity, the realtime graphic of a certain patient of acquisition is no longer updated to sample database, simultaneously for
Without being trained, i.e. step S3 is no longer executed the realtime graphic of a certain patient's subsequent acquisition, directly using former trained
To the classifier of the patient classify, i.e., directly execute step S4.
Preferably, step S32 is trained with neural network, can also have the SVM of supervision to carry out.
As supplement, when security incident judges that system does not judge security incident, but doctor or nurse pass through scene judgement
Out when security incident, doctor or nurse can directly pass through medical information management system typing security incident, can be as shown in Figure 3.
Medical information management system can provide the inquiry and statistics that interface carries out security incident for healthcare givers, specific visible
Shown in Figure 4 and 5.
Preferably, medical information management system also receives the FRID label information of radio-frequency identification reader identification, and judges
The identification position of the radio-frequency identification reader, if identification position is located at hospital doorway or patient is not allowed to the position entered
It sets, then send warning information to attending physician or cures mainly nurse.
Claims (7)
1. a kind of security incident identifies reporting system, including multiple cameras, central server, multiple processing terminals and multiple penetrate
Frequency identification reader, central server are read by internal network and multiple cameras, multiple processing terminals, multiple radio frequency identifications
Device is connected；
Multiple cameras can remotely be controlled by central server, and multiple cameras are arranged in hospital bed, ward doorway, consulting room, hand
Art room and corridor position；Multiple cameras can receive the acquisition instructions of central server, and carry out image according to acquisition instructions
Acquisition, and send the image of acquisition to central server；
Multiple radio-frequency identification readers are connect with central server by internal network, and multiple radio-frequency identification readers are pacified
Mounted in ward, consulting room, operating room and corridor；With RFID label tag patient in multiple radio-frequency identification readers wherein
At one, one of them in multiple radio-frequency identification readers can identify the RFID label tag, and the RFID label tag of identification is believed
Breath is by transmission of network to central server, and wherein the RFID label tag includes patient information；
Multiple processing terminals are connected with printer, when doctor by one of multiple processing terminals come for patient's Admission formality
When, by the printing of doctor, printer printing is controlled according to patient information and generates a RFID mark comprising patient information
Label, and the RFID label tag is pasted on patient's clothes or bracelet；
Central server, it includes sample database, security incidents to judge system and medical information management system, central server
The acquisition image of multiple camera transmission can be received and be saved into sample database or send security incident judgement system to
System carries out security incident judgement；The FRID label information of multiple radio-frequency identification reader transmission can also be received, and records and passes through
FRID label information obtain patient information and identify the FRID label information radio-frequency identification reader ID, and center clothes
Business device has recorded the ID of each radio-frequency identification reader and the incidence relation of its installation site in advance；
Security incident judges that system can judge whether there is security incident in image to its received image, and sends out in judgement
When raw security incident, image, patient information and the security incident type by security incident warning message, currently judged is sent to doctor
Treat information management system；
Medical information management system receives and record security event, and personnel corresponding to the corresponding patient of query safe event,
Then security incident, patient information are sent to corresponding personnel, while security incident is saved and counted；
Wherein, sample database includes positive sample image and negative sample image, and wherein positive sample image obtains more comprising fitting
A simulation positive sample image, wherein fitting obtains multiple simulation positive sample images and is fitted in the following manner:
A) it establishes confrontation and generates network, it includes discriminator and generator which, which generates network, and generator is depth convolutional Neural
Network comprising three convolutional layers and an output layer, input are random vector and a standard input picture, are exported as a mould
Quasi- image, Plays input picture are generated by calculator simulation；Discriminator is anti-neural network comprising five layers, preceding
Three layers are convolutional layer, and the 4th layer is full articulamentum, and the last layer is output layer, and input is the analog image of generator output,
It is true that it, which exports analog image by reference to true picture, or simulate；
B) alternative optimization discriminator and generator finally reach balance；
C) network is generated using the confrontation of b) optimization and one standard input picture is generated into a secondary analog image, repeat the step
Suddenly multiple analog images are obtained as multiple simulation positive sample images；
Sample database is established in the following way:
Establish negative sample image: after hospital bed, the radio-frequency identification reader of arranging hospital beds identifies the patient with RFID label tag
The RFID label tag, and the FRID label information of identification is passed through into transmission of network to central server；Central server identifies this
The ID of the patient information that FRID label information includes and the radio-frequency identification reader for identifying the FRID label information, and further
According to the incidence relation of the ID of each radio-frequency identification reader recorded in advance and its installation site, obtains and identify the RFID label tag
The installation site of the radio-frequency identification reader of information, then central server judges whether the patient information identified is for the first time
Whether identified and the radio-frequency identification reader installation site is hospital bed, if the patient information identified is to be known for the first time
Not and the installation site of the radio-frequency identification reader is hospital bed, then central server sends image acquisition commands and reads to radio frequency identification
The corresponding camera of installation site of device is read, camera starts to acquire image and sends the image of acquisition to central server,
The image of acquisition is saved as the corresponding negative sample image of patient information to sample database by central server；
Establish positive sample image: the image when security incident judges that security incident occurs for system judgement is saved in sample database
In and as positive sample image；The positive sample image that security incident occurs is obtained by other image data bases；Fitting obtains more
A simulation positive sample image is saved in sample database.
2. system according to claim 1, it is characterised in that: security incident judges that system includes following device:
Training module selects positive and negative samples image from sample database, uses the negative sample image and positive sample figure of selection
As training obtains a classifier；
Received image is inputted the classifier of above-mentioned training, output category result, and is to exist in classification results by judgment module
When security incident, warning message is sent to medical information management system.
3. a kind of security incident using security incident identification reporting system identifies report method, wherein in security incident identification
Reporting system includes multiple cameras, central server, multiple processing terminals and multiple radio-frequency identification readers, and central server is logical
Internal network is crossed to be connected with multiple cameras, multiple processing terminals, multiple radio-frequency identification readers；It is characterized in that, described
Method includes:
S1, sample database is established:
Establish negative sample image: after hospital bed, the radio-frequency identification reader of arranging hospital beds identifies the patient with RFID label tag
The RFID label tag, and the FRID label information of identification is passed through into transmission of network to central server；Central server identifies this
The ID of the patient information that FRID label information includes and the radio-frequency identification reader for identifying the FRID label information, and further
According to the incidence relation of the ID of each radio-frequency identification reader recorded in advance and its installation site, obtains and identify the RFID label tag
The installation site of the radio-frequency identification reader of information, then central server judges whether the patient information identified is for the first time
Whether identified and the radio-frequency identification reader installation site is hospital bed, if the patient information identified is to be known for the first time
Not and the installation site of the radio-frequency identification reader is hospital bed, then central server sends image acquisition commands and reads to radio frequency identification
The corresponding camera of installation site of device is read, camera starts to acquire image and sends the image of acquisition to central server,
The image of acquisition is saved as the corresponding negative sample image of patient information to sample database by central server；
Establish positive sample image: the image when security incident judges that security incident occurs for system judgement is saved in sample database
In and as positive sample image；The positive sample image that security incident occurs is obtained by other image data bases；Fitting obtains more
A simulation positive sample image is saved in sample database；
Fitting obtain multiple simulation positive sample images the following steps are included:
A) it establishes confrontation and generates network, it includes discriminator and generator which, which generates network, and discriminator is depth convolutional Neural
Network comprising three convolutional layers and an output layer, input are random vector and a standard input picture, are exported as a mould
Quasi- image, Plays input picture are generated by calculator simulation；Discriminator is anti-neural network comprising five layers, preceding
Three layers are convolutional layer, and the 4th layer is full articulamentum, and the last layer is output layer, and input is the analog image of generator output,
It is true that it, which exports analog image by reference to true picture, or simulate；
B) alternative optimization discriminator and generator finally reach balance；
C) network is generated using the confrontation of b) optimization and one standard input picture is generated into a secondary analog image, repeat the step
Suddenly multiple analog images are obtained as multiple simulation positive sample images；
S2, acquisition patient's realtime graphic:
Security incident judges that system sends figure in real time to the camera of the corresponding hospital bed of all patients automatically by central server
As acquisition instructions, the security incident that collected realtime graphic is sent to central server is judged system, safe thing by camera
Part judges that system for each secondary realtime graphic, is performed both by following S3 training step and S4 judgment step；
S3, training step:
S31 selects sample image from sample database:
The negative sample image of the first predetermined quantity and the positive sample of the second predetermined quantity are selected from the sample database established in S1
This image: according to the realtime graphic obtained in step S2, obtaining the corresponding current patents' information of the realtime graphic, current according to this
Patient information selects the corresponding negative sample image of the current patient information and positive sample image from sample database；
S32 uses the step S31 negative sample image selected and positive sample image training classifier；
S4, judgment step:
The realtime graphic obtained in step S2 is input to the obtained classifier of step S3 training and carries out security incident judgement, when point
When the output of class device belongs to security incident, security incident judges that system is issued to the medical information management system of central server and warns
Report, while the realtime graphic of acquisition, patient information and security incident type also being sent together, medical information management system saves
The security incident, and attending physician and the Night shift of patient are found according to patient information, to attending physician and Night shift
Mobile phone send warning information, while sending realtime graphic, patient information and security incident type, attending physician and supervisor together
Nurse is judged by realtime graphic, and sends response message to medical information management system.
4. according to the method described in claim 3, it is characterized in that,
In step S4, if medical information management system does not receive the response letter of attending physician and Night shift within a certain period of time
Breath, then be sent to the second doctor and the second nurse for warning information and realtime graphic, patient information and security incident type,
It can be transmitted to higher level doctor and higher level nurse.
5. according to the method described in claim 3, it is characterized in that, when judging security incident in S4 step, by safe thing
The corresponding realtime graphic of part, patient information, security incident type are sent to sample database, and sample database is saved as just
Sample image；Meanwhile when judging not there is no security incident in S4 step, realtime graphic, patient information are sent to sample number
According to library, sample database is saved as negative sample image.
6. according to the method described in claim 5, it is characterized in that, in the sample database negative sample image of a certain patient or
When the quantity of positive sample image is more than scheduled quantity, the realtime graphic of a certain patient of acquisition is no longer updated to sample data
Library, simultaneously for a certain patient's subsequent acquisition realtime graphic without no longer execute step S3, directly using in the past training obtain
The classifier of the patient classify, directly execution step S4.
7. according to the method described in claim 3, it is characterized in that, selecting sample image to specifically include in the step S31:
According to the realtime graphic obtained in step S2, the corresponding current patents' information of the realtime graphic is obtained, according to the current trouble
Person's information selects the negative sample image of corresponding first predetermined quantity of the current patient information and second pre- from sample database
The positive sample image of fixed number amount, the negative sample image of the current patents obtained from sample database first and current patents are just
Sample image, if the first quantity of the negative sample image of the acquisition less than the first predetermined quantity, then obtains the of other patients
The negative sample quantity of two quantity so that the first quantity and the second quantity and be equal to the first predetermined quantity；If the positive sample obtained
The third quantity of image less than the second predetermined quantity, then obtain again from common image database and simulation positive sample image in
Machine selects the positive sample quantity of the 4th quantity so that third quantity and the 4th quantity and be equal to the second predetermined quantity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811197737.1A CN109409266B (en) | 2018-10-15 | 2018-10-15 | A kind of security incident identification reporting system and security incident identify report method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811197737.1A CN109409266B (en) | 2018-10-15 | 2018-10-15 | A kind of security incident identification reporting system and security incident identify report method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109409266A CN109409266A (en) | 2019-03-01 |
| CN109409266B true CN109409266B (en) | 2019-09-06 |
Family
ID=65467984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811197737.1A Active CN109409266B (en) | 2018-10-15 | 2018-10-15 | A kind of security incident identification reporting system and security incident identify report method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109409266B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101866160A (en) * | 2010-06-21 | 2010-10-20 | 昆明理工大学 | Intelligent control system of building safety |
| EP2641577A2 (en) * | 2012-03-21 | 2013-09-25 | Hill-Rom Services, Inc. | Patient support apparatus with redundant identity verification |
| CN103390209A (en) * | 2013-05-12 | 2013-11-13 | 安徽工程大学 | Integrated management system for mental hospital |
| CN103729730A (en) * | 2013-12-13 | 2014-04-16 | 武汉方南科技有限公司 | Visual monitoring traceability system and method for animal husbandry products |
| WO2014074585A1 (en) * | 2012-11-06 | 2014-05-15 | Johnson Raymond C | Virtual hand-washing coach |
| CN205750807U (en) * | 2016-04-26 | 2016-11-30 | 北京森科赛德科技有限公司 | A kind of kindergarten ensuring pick-up safety system based on iris identification |
| CN107978044A (en) * | 2017-11-29 | 2018-05-01 | 南京甄视智能科技有限公司 | Based on face recognition technology and the high-precision identifying system of RFID technique and recognition methods |
| CN207319260U (en) * | 2017-10-18 | 2018-05-04 | 浙江钧普科技股份有限公司 | A kind of Managerial System of Medical Equipment based on RFID technique |
| CN109035454A (en) * | 2018-05-30 | 2018-12-18 | 广州爱易学智能信息科技有限公司 | A kind of multi-functional pre-payment badge system and its application method |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070132597A1 (en) * | 2005-12-09 | 2007-06-14 | Valence Broadband, Inc. | Methods and systems for monitoring patient support exiting and initiating response |
| CN105976291A (en) * | 2016-05-30 | 2016-09-28 | 镇江市第人民医院 | Hospital adverse event reporting processing system and hospital adverse event reporting processing method |
| CN206075353U (en) * | 2016-05-30 | 2017-04-05 | 镇江市第一人民医院 | A kind of hospital's adverse events report processing system |
| CN107194418B (en) * | 2017-05-10 | 2021-09-28 | 中国科学院合肥物质科学研究院 | Rice aphid detection method based on antagonistic characteristic learning |
| CN107633220A (en) * | 2017-09-13 | 2018-01-26 | 吉林大学 | A kind of vehicle front target identification method based on convolutional neural networks |
| CN107730681A (en) * | 2017-10-16 | 2018-02-23 | 长春北电软件有限公司 | A kind of Campus Security ensures intelligent supervision early warning system |
| CN107909621A (en) * | 2017-11-16 | 2018-04-13 | 深圳市唯特视科技有限公司 | It is a kind of based on it is twin into confrontation network medical image synthetic method |
-
2018
- 2018-10-15 CN CN201811197737.1A patent/CN109409266B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101866160A (en) * | 2010-06-21 | 2010-10-20 | 昆明理工大学 | Intelligent control system of building safety |
| EP2641577A2 (en) * | 2012-03-21 | 2013-09-25 | Hill-Rom Services, Inc. | Patient support apparatus with redundant identity verification |
| WO2014074585A1 (en) * | 2012-11-06 | 2014-05-15 | Johnson Raymond C | Virtual hand-washing coach |
| CN103390209A (en) * | 2013-05-12 | 2013-11-13 | 安徽工程大学 | Integrated management system for mental hospital |
| CN103729730A (en) * | 2013-12-13 | 2014-04-16 | 武汉方南科技有限公司 | Visual monitoring traceability system and method for animal husbandry products |
| CN205750807U (en) * | 2016-04-26 | 2016-11-30 | 北京森科赛德科技有限公司 | A kind of kindergarten ensuring pick-up safety system based on iris identification |
| CN207319260U (en) * | 2017-10-18 | 2018-05-04 | 浙江钧普科技股份有限公司 | A kind of Managerial System of Medical Equipment based on RFID technique |
| CN107978044A (en) * | 2017-11-29 | 2018-05-01 | 南京甄视智能科技有限公司 | Based on face recognition technology and the high-precision identifying system of RFID technique and recognition methods |
| CN109035454A (en) * | 2018-05-30 | 2018-12-18 | 广州爱易学智能信息科技有限公司 | A kind of multi-functional pre-payment badge system and its application method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109409266A (en) | 2019-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104615852B (en) | The method for order and the raising source service efficiency of registering for guarantee online booking | |
| CN109255731A (en) | A kind of architectural engineering labor service personnel management system and management method | |
| CN110321809A (en) | A kind of substation's operation field monitoring method and device based on deep learning | |
| CN106373063A (en) | Automatic triage treatment control method based on interactive communication and system | |
| CN107818536B (en) | Monitoring system and method for realizing community drug rehabilitation and rehabilitation based on intelligent mobile terminal | |
| CN106777929A (en) | Eeg monitoring system based on smart cloud | |
| CN110516568B (en) | College multi-scene data management method and system based on face recognition | |
| CN110942822A (en) | Intelligent management system for medical equipment | |
| CN109460744A (en) | A kind of video monitoring system based on deep learning | |
| CN107566790A (en) | With reference to REID and the real time monitoring apparatus and method of Video Supervision Technique | |
| CN113050478A (en) | Method for monitoring and early warning operation state of transformer substation constructor | |
| CN112911017A (en) | Operating personnel safety state risk management and control system based on intelligent Internet | |
| CN111352349A (en) | System and method for collecting and adjusting information of old people living environment | |
| CN115424340A (en) | Method for identifying protective clothing based on YOLOV5 neural network technology | |
| CN110084935A (en) | Enter and leave method for early warning and device | |
| CN114343595B (en) | Intelligent community endowment service system | |
| CN109376635B (en) | A kind of nursing quality checking system and security incident report method | |
| CN114974531A (en) | Intelligent medical care system based on big data | |
| CN114420302A (en) | Intelligent epidemic prevention control system for enterprises and public institutions | |
| CN109409266B (en) | A kind of security incident identification reporting system and security incident identify report method | |
| CN102419890A (en) | Monitor alarming system and method | |
| CN106778051A (en) | A kind of post health monitoring systems based on domestic CPU | |
| CN109949927A (en) | A kind of intelligent diagnosing method and its system based on deep neural network | |
| CN114822793A (en) | Isolation ward human resource management system | |
| CN113593706A (en) | Health monitoring and service management system based on artificial intelligence technology |
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 |