CN117059276A

CN117059276A - Patient vital sign monitoring method and system based on artificial intelligence

Info

Publication number: CN117059276A
Application number: CN202311067150.XA
Authority: CN
Inventors: 黄小平; 胡译丹; 吕琳
Original assignee: Chongqing Traditional Chinese Medicine Hospital
Current assignee: Chongqing Traditional Chinese Medicine Hospital
Priority date: 2023-08-23
Filing date: 2023-08-23
Publication date: 2023-11-14

Abstract

According to the patient vital sign monitoring method and system based on artificial intelligence, based on patient vital sign monitoring data to be analyzed, the to-be-determined data core is subjected to optimization processing, and an optimized data core corresponding to the to-be-determined data core and a compressed vital sign data set corresponding to the optimized data core are obtained. The correlation between the data analysis result and the vital sign data can be comprehensively considered. According to the data analysis method, through optimization processing and comprehensive consideration of the relevance between the data analysis result and vital sign data, the positioning accuracy index of the constraint range where the important indication is located is highly valued, so that the debugging window capable of indicating the important indication of the vital sign monitoring data of the patient to be analyzed more accurately is obtained, the accuracy of data analysis is further effectively improved, and the accuracy and reliability of the vital sign monitoring result of the patient can be guaranteed.

Description

Patient vital sign monitoring method and system based on artificial intelligence

Technical Field

The application relates to the technical field of data monitoring, in particular to a patient vital sign monitoring method and system based on artificial intelligence.

Background

Vital signs are used to determine whether a patient is critically ill or critically ill. Mainly heart rate, pulse, blood pressure, respiration, pain, blood oxygen, pupil and cornea reflex changes, etc. In a normal state, the pulse is 60 to 100 times/minute (generally 70 to 80 times/minute). Heart rate and pulse are significantly accelerated when drugs such as cardiac insufficiency, shock, hyperthermia, severe anemia and pain, thyroid crisis, myocarditis, and atropine are poisoned. When intracranial pressure increases, complete atrioventricular block, the pulse slows down. In general, the heart rate is consistent with the pulse, but in the case of arrhythmias such as atrial fibrillation, frequent premature beat, the pulse is less than the heart rate, known as a short pulse.

At present, manual monitoring or monitoring by a single monitoring device is generally adopted for monitoring vital signs, and all monitoring devices cannot be integrated, so that a situation that the monitoring is not in place may exist, the vital signs of a patient cannot be accurately known, and therefore, a technical scheme is needed to improve the technical problems.

Disclosure of Invention

In order to improve the technical problems in the related art, the application provides a patient vital sign monitoring method and system based on artificial intelligence.

In a first aspect, there is provided an artificial intelligence based patient vital sign monitoring method, the method comprising: when vital sign monitoring data of a patient to be analyzed are obtained, carrying out data analysis on important indications of the vital sign monitoring data of the patient to be analyzed to obtain a to-be-determined data core; performing optimization processing on the undetermined data core according to the vital sign monitoring data of the patient to be analyzed to obtain an optimized data core corresponding to the undetermined data core and a compressed vital sign data set corresponding to the optimized data core; the compressed vital sign dataset includes a compressed vital sign data; the compressed vital sign data are obtained by compressing the vital sign monitoring data of the patient to be analyzed according to an optimization window in the optimization data core; a is an integer greater than 0; classifying each compressed vital sign data in the A compressed vital sign data to obtain a classification result matched with the important indication; the classification result comprises C vital sign monitoring categories; c is not more than A and C is an integer greater than 0; and in the optimized data core, respectively debugging the optimized windows corresponding to the C vital sign monitoring types to obtain C debugging windows used for representing the important indication in the vital sign monitoring data of the patient to be analyzed, and obtaining the vital sign monitoring result of the patient through the debugging windows of the important indication.

It can be understood that when the vital sign monitoring data of the patient to be analyzed is obtained, the vital sign monitoring data of the patient to be analyzed can be subjected to data analysis to obtain the undetermined data core, wherein the vital sign monitoring data of the patient to be analyzed is carried with the data analysis function. Wherein the pending data core is formed by the constraint range where the originally identified important indicator is located. In order to effectively ensure the accuracy of the constraint range of the important indication, after the pending data core is obtained, secondary optimization processing (including de-optimization processing and debugging processing) can be performed on the pending data core. Firstly, the method can perform optimization processing on the to-be-determined data core based on the to-be-analyzed patient vital sign monitoring data to obtain an optimized data core corresponding to the to-be-determined data core and a compressed vital sign data set corresponding to the optimized data core. Wherein the compressed vital sign dataset herein comprises a compressed vital sign data; the compressed vital sign data is obtained by compressing the vital sign monitoring data of the patient to be analyzed based on an optimization window in the optimization data core; a is an integer greater than 0. Then, the correlation between the data analysis result and the vital sign data can be comprehensively considered, the undetermined data core (i.e. the optimized data core) after the optimization processing is performed again, namely, the classification processing is performed on each compressed vital sign data in the A compressed vital sign data respectively to obtain a classification result matched with the important indication, and further, in the optimized data core, the debugging processing can be performed on optimization windows corresponding to C vital sign monitoring types included in the classification result respectively, so as to accurately obtain C constraint ranges (i.e. debugging windows) for representing the important indication in the vital sign monitoring data of the patient to be analyzed. Wherein C is not more than A and C is an integer of more than 0. Therefore, according to the data analysis method provided by the embodiment of the application, through optimization processing and comprehensive consideration of the correlation between the data analysis result and vital sign data, the positioning accuracy index of the constraint range where the important indication is located is highly emphasized, so that the debugging window capable of more accurately indicating the important indication of the vital sign monitoring data of the patient to be analyzed is obtained, the accuracy of data analysis is further effectively improved, and the accuracy and reliability of the vital sign monitoring result of the patient can be ensured.

In an independently implemented embodiment, before data parsing the vital sign monitoring data of the patient to be parsed to obtain the pending data core, the method further comprises: obtaining first example vital sign data for configuring a first Ab resolution thread, and a first example catalog for representing a current category of the first example vital sign data; the first example vital sign data is obtained by preprocessing an important indication of the original example vital sign data; invoking the first Ab analysis thread to perform data analysis on the first example vital sign data to obtain a prediction analysis type credible coefficient of the first example vital sign data aiming at the important indication; and configuring the first Ab analysis thread according to the prediction analysis type credible coefficient of the first example vital sign data and the current type of the first example vital sign data to obtain a data analysis thread for carrying out data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed.

It can be understood that the vital sign monitoring data of the patient to be analyzed is required to be analyzed before the important indication of the vital sign monitoring data of the patient to be analyzed is subjected to data analysis to obtain the pending data core, so that the accuracy and the reliability of the vital sign monitoring data of the patient to be analyzed can be ensured.

In an independent embodiment, the configuring the first Ab analysis thread according to the prediction analysis type trusted coefficient of the first example vital sign data and the current type of the first example vital sign data to obtain a data analysis thread for performing data analysis on the important indication of the patient vital sign monitoring data to be analyzed includes: performing evaluation index calculation processing on the prediction analysis type credible coefficient of the first example vital sign data and the current type of the first example vital sign data, and determining a thread evaluation index of the first Ab analysis thread; configuring the first Ab analysis thread according to the thread evaluation index of the first Ab analysis thread to obtain a first thread configuration result; if the first thread configuration result indicates that the configured first Ab analysis thread meets the first thread Cheng Shoulian requirement, taking the first Ab analysis thread meeting the first thread convergence requirement as a second Ab analysis thread; and carrying out architecture analysis on the thread architecture of the second Ab analysis thread, and determining a data analysis thread for carrying out data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed.

It can be appreciated that, when the first Ab analysis thread is configured according to the prediction analysis type trusted coefficient of the first example vital sign data and the current type of the first example vital sign data, the problem that the prediction analysis type trusted coefficient and the current type are unclear is improved, and the cluster can accurately obtain the data analysis thread for performing data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed.

In an independently implemented embodiment, the performing architecture analysis on the thread architecture of the second Ab analysis thread, determining a data analysis thread for performing data analysis on the vital sign monitoring data of the patient to be analyzed, includes: performing architecture analysis on the thread architecture of the second Ab analysis thread to obtain an analysis result; if the analysis result indicates that a merging unit exists in the thread architecture of the second Ab analysis thread, switching the merging unit into a first feature extraction unit configured with a sliding step length; assigning a value to the first feature extraction unit according to the thread coefficient of the sum unit to obtain a second feature extraction unit; and taking a second Ab analysis thread comprising the second feature extraction unit as a data analysis thread for carrying out data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed.

It can be appreciated that when the architecture analysis is performed on the thread architecture of the second Ab analysis thread, the problem of inaccurate analysis is improved, so that the data analysis thread for performing data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed can be accurately determined.

In an independent embodiment, the performing, according to the vital sign monitoring data of the patient to be analyzed, a de-optimization process on the pending data core to obtain an optimized data core corresponding to the pending data core and a compressed vital sign data set corresponding to the optimized data core includes: performing de-duplication processing on the undetermined data core based on a repeated data correction network to obtain a data core to be processed; the data core to be processed comprises a window Ab to be determined; b is an integer not greater than M greater than 0; m is used for representing the total number of pending windows in the data core to be processed; invoking a vital sign data optimization thread, and optimizing the undetermined window Ab through the vital sign monitoring data of the patient to be analyzed to obtain an optimization window Cb and compressed vital sign data corresponding to the optimization window Cb; when M optimizing windows are obtained, carrying out de-duplication treatment on the M optimizing windows according to a de-duplication mode; and taking the optimized window after the de-duplication processing as an optimized data core corresponding to the undetermined data core, and taking the compressed vital sign data corresponding to the optimized window after the de-duplication processing as a compressed vital sign data set corresponding to the optimized data core.

It can be understood that when the to-be-analyzed patient vital sign monitoring data is subjected to the de-optimization processing, the problem of inaccurate processing is solved, so that an optimized data core corresponding to the to-be-analyzed data core and a compressed vital sign data set corresponding to the optimized data core can be accurately obtained.

In an independent embodiment, the pending data core is obtained when invoking a data parsing thread to parse the important indication of the patient vital sign monitoring data to be parsed; the pending data core comprises N pending windows; n is an integer greater than 0; the data analysis thread is further used for determining the prediction analysis type credibility coefficient corresponding to each of the N undetermined windows; the duplicate data correction network is used for carrying out duplicate removal processing on the undetermined data core to obtain the undetermined data core, and the duplicate data correction network comprises the following steps: based on the vital sign data compression rate specified by the repeated data correction network, respectively compressing each of the N undetermined windows to obtain N compressed windows; based on N prediction analysis type credible coefficients, sorting the N compression windows to obtain a distribution condition; taking a compression window carrying the maximum prediction analysis type credible coefficient in the distribution situation as a first compression window, and taking (N-1) compression windows except the first compression window in the distribution situation as a first to-be-cleaned set; performing de-duplication processing on the N compression windows according to the repetition rate between the first compression window and each compression window in the first set to be cleaned to obtain a reserved data core; and respectively carrying out vital sign data derivatization processing on each compression window in the reserved data core according to the vital sign data compression rate to obtain a data core to be processed.

It can be appreciated that when the duplicate data correction network is used to perform the duplicate processing on the pending data core, the problem of not in place processing is improved, so that the pending data core can be accurately obtained.

In an independent embodiment, the performing, according to the repetition rate between the first compression window and each compression window in the first to-be-cleaned set, deduplication processing on the N compression windows to obtain a reserved data core includes: determining repetition rates between the first compression window and each compression window in the first set to be cleaned respectively; if the first to-be-cleaned set has a repeated compression window with the repetition rate larger than the target value of the repetition rate, reserving the first compression window, and cleaning the repeated compression window in the first to-be-cleaned set; taking a compression window carrying the maximum prediction analysis type credible coefficient as a second compression window in the first set to be cleaned after cleaning, and taking compression windows except the second compression window as a second set to be cleaned; and reserving the second compression window, performing de-duplication treatment on the repeated compression windows in the second to-be-cleaned set according to the repetition rate between the second compression window and each compression window in the second to-be-cleaned set until the second to-be-cleaned set subjected to de-duplication treatment is empty, and taking the reserved first compression window and the reserved second compression window as reserved data cores.

It can be appreciated that, according to the repetition rate between the first compression window and each compression window in the first set to be cleaned, when the deduplication processing is performed on the N compression windows, the problem of inaccurate deduplication is improved, so that the reserved data core can be accurately obtained.

In an embodiment of the independent implementation, the invoking the vital sign data optimization thread performs optimization processing on the undetermined window Ab through the vital sign monitoring data of the patient to be analyzed to obtain an optimized window Cb and compressed vital sign data corresponding to the optimized window Cb, and includes: invoking the vital sign data optimization thread, and performing error quantity prediction analysis on the undetermined window Ab through the vital sign monitoring data of the patient to be analyzed to obtain a first prediction coefficient; optimizing the undetermined window Ab according to the first prediction coefficient to obtain a first optimized window corresponding to the undetermined window Ab; if the first prediction coefficient belongs to the target value range of the prediction coefficient, taking the first optimization window corresponding to the undetermined window Ab as an optimization window Cb corresponding to the undetermined window Ab; and carrying out compression processing on the vital sign monitoring data of the patient to be analyzed according to the AI vector positioning of the optimization window Cb to obtain compressed vital sign data corresponding to the optimization window Cb.

It can be understood that when the vital sign data optimization thread is invoked and the to-be-analyzed patient vital sign monitoring data is used for optimizing the to-be-determined window Ab, the problem of inaccurate optimization is solved, so that the optimization window Cb and the compressed vital sign data corresponding to the optimization window Cb can be accurately obtained.

In an independently implemented embodiment, after obtaining the first optimized window corresponding to the pending window Ab, the method further includes: if the first prediction coefficient does not belong to the target value range of the prediction coefficient, calling the vital sign data optimization thread, and performing error quantity prediction analysis on a first optimization window corresponding to the undetermined window Ab through vital sign monitoring data of the patient to be analyzed to obtain a second prediction coefficient; according to the second prediction coefficient, performing optimization processing on a first optimization window corresponding to the undetermined window Ab to obtain a second optimization window corresponding to the undetermined window Ab; and taking the second optimization window corresponding to the undetermined window Ab as the optimization window Cb corresponding to the undetermined window Ab until the second prediction coefficient belongs to the target value range of the prediction coefficient.

It can be appreciated that the reliability of the second optimization window can be improved by error amount prediction analysis.

In an independently implemented embodiment, before invoking the vital sign data optimization thread, the method further comprises: obtaining second example vital sign data for configuring the original optimization thread, and a second example catalog for representing a current AI vector location where an important indication of the second example vital sign data is located; invoking the original optimization thread, and performing offset prediction analysis processing on the second example vital sign data to obtain a prediction analysis prediction coefficient of the second example vital sign data; optimizing the AI vector positioning of the second example vital sign data according to the prediction analysis prediction coefficient to obtain the prediction analysis AI vector positioning of the second example vital sign data; configuring the original optimization thread according to the prediction analysis AI vector positioning and the current AI vector positioning to obtain a second thread configuration result; and if the second thread configuration result indicates that the configured original optimization thread meets the second thread convergence requirement, taking the original optimization thread meeting the second thread convergence requirement as the vital sign data optimization thread.

It can be appreciated that by optimizing AI vector positioning, the accuracy of the vital sign data optimization thread can be improved.

In an independent embodiment, the classifying the compressed vital sign data of the a compressed vital sign data to obtain a classification result matched with the important indication includes: determining vital sign data to be processed from the a compressed vital sign data; invoking a vital sign data classification thread, and reading and processing the vital sign data to be processed according to the vital sign data classification type in the vital sign data classification thread to obtain a prediction analysis type credible coefficient corresponding to the vital sign data to be processed; deleting the non-important indication of the vital sign data to be processed through the vital sign data segmentation type in the vital sign data classification thread to obtain a vital sign monitoring type corresponding to the vital sign data to be processed; and if the credible coefficient of the predictive analysis type corresponding to the vital sign data to be processed is larger than the credible coefficient target value, taking the vital sign monitoring type corresponding to the vital sign data to be processed as a classification result matched with the important indication.

It can be understood that when classifying each of the a compressed vital sign data, the problem of inaccurate classification is improved, so that a classification result matching the important indication can be accurately obtained.

In an independently implemented embodiment, before invoking the vital sign data classification thread, the method further comprises: obtaining example data for configuring an original classification thread and an example catalog corresponding to the example data; configuring a loss thread and a vital sign data segmentation type in the original classification thread according to third example vital sign data in the example data and a third example catalog used for representing current vital sign shielding data of the third example vital sign data to obtain a first classification thread; according to the fourth example vital sign data in the example data and a fourth example catalog used for representing the current type of the fourth example vital sign data, configuring a loss thread and a vital sign data classification type in a first reading thread to obtain a second classification thread; configuring a loss thread and vital sign data segmentation type in the second classification thread according to the third example vital sign data and a third example catalog to obtain a third classification thread; and locking the loss thread and the vital sign data segmentation type in the third classification thread, and configuring the vital sign data classification type in the locked third classification thread according to the fourth example vital sign data and the fourth example catalog to obtain the vital sign data classification thread.

It can be appreciated that by configuring the loss thread and the vital sign data classification type in the first read thread, the reliability of the vital sign data classification thread can be ensured.

In an independent embodiment, the configuring the loss thread and the vital sign data partition type in the second classification thread according to the third exemplary vital sign data and the third exemplary directory to obtain a third classification thread includes: invoking the second classification thread to delete the non-important indication of the third example vital sign data to obtain prediction analysis vital sign shielding data corresponding to the third example vital sign data; traversing the data of the third example vital sign data, and taking the traversed data as data to be processed; taking the credible coefficient of the to-be-processed data in the current vital sign shielding data indicated by the third example catalog as a first credible coefficient, and taking the credible coefficient of the to-be-processed data in the predictive analysis vital sign shielding data in the to-be-processed data as a second credible coefficient; performing evaluation index calculation processing on the first trusted coefficient and the second trusted coefficient, and determining a thread evaluation index of the second classification thread; configuring a loss thread and vital sign data segmentation type in the second classification thread according to the thread evaluation index of the second classification thread to obtain a third thread configuration result; and if the third thread configuration result indicates that the configured second classification thread meets the first type convergence requirement in the third thread Cheng Shoulian requirements, taking the second classification thread meeting the first type convergence requirement as a third classification thread.

It can be appreciated that, when the loss thread and the vital sign data segmentation type in the second classification thread are configured according to the third exemplary vital sign data and the third exemplary directory, the problem of data abnormality is improved, so that the third classification thread can be accurately obtained.

In an independent embodiment, the locking the loss thread and the vital sign data segmentation type in the third classification thread configures the vital sign data classification type in the locked third classification thread according to the fourth example vital sign data and the fourth example catalog to obtain a vital sign data classification thread, which includes: locking a loss thread and vital sign data segmentation type in the third classification thread, and taking the locked third classification thread as a fourth classification thread; invoking the fourth classification thread, reading the fourth example vital sign data, and determining an example prediction analysis trusted coefficient corresponding to the fourth example vital sign data; performing evaluation index calculation processing on the current category indicated by the fourth catalog and the sample prediction analysis credibility coefficient, and determining a thread evaluation index of the fourth classification thread; configuring vital sign data classification types in the fourth classification thread according to the thread evaluation index of the fourth classification thread to obtain a fourth thread configuration result; and if the fourth thread configuration result indicates that the configured fourth classification thread meets the second type convergence requirement in the third line Cheng Shoulian requirements, taking the fourth classification thread meeting the second type convergence requirement as a vital sign data classification thread.

It can be understood that the loss thread and the vital sign data segmentation type in the third classification thread are locked, and when the vital sign data classification type in the locked third classification thread is configured according to the fourth example vital sign data and the fourth example catalog, the problem of inaccurate segmentation is solved, so that the vital sign data classification thread can be accurately obtained.

In a second aspect, an artificial intelligence based patient vital sign monitoring system is provided comprising a processor and a memory in communication with each other, the processor being adapted to read a computer program from the memory and execute the computer program to implement the method as described above.

According to the patient vital sign monitoring method and system based on artificial intelligence, which are provided by the embodiment of the application, when the patient vital sign monitoring data to be analyzed is obtained, the important indication of the patient vital sign monitoring data to be analyzed can be subjected to data analysis, so that the undetermined data core is obtained. Wherein the pending data core is formed by the constraint range where the originally identified important indicator is located. In order to effectively ensure the accuracy of the constraint range of the important indication, after the pending data core is obtained, secondary optimization processing (including de-optimization processing and debugging processing) can be performed on the pending data core. Firstly, the method can perform optimization processing on the to-be-determined data core based on the to-be-analyzed patient vital sign monitoring data to obtain an optimized data core corresponding to the to-be-determined data core and a compressed vital sign data set corresponding to the optimized data core. Wherein the compressed vital sign dataset herein comprises a compressed vital sign data; the compressed vital sign data is obtained by compressing the vital sign monitoring data of the patient to be analyzed based on an optimization window in the optimization data core; a is an integer greater than 0. Then, the correlation between the data analysis result and the vital sign data can be comprehensively considered, the undetermined data core (i.e. the optimized data core) after the optimization processing is performed again, namely, the classification processing is performed on each compressed vital sign data in the A compressed vital sign data respectively to obtain a classification result matched with the important indication, and further, in the optimized data core, the debugging processing can be performed on optimization windows corresponding to C vital sign monitoring types included in the classification result respectively, so as to accurately obtain C constraint ranges (i.e. debugging windows) for representing the important indication in the vital sign monitoring data of the patient to be analyzed. Wherein C is not more than A and C is an integer of more than 0. Therefore, according to the data analysis method provided by the embodiment of the application, through optimization processing and comprehensive consideration of the correlation between the data analysis result and vital sign data, the positioning accuracy index of the constraint range where the important indication is located is highly emphasized, so that the debugging window capable of more accurately indicating the important indication of the vital sign monitoring data of the patient to be analyzed is obtained, the accuracy of data analysis is further effectively improved, and the accuracy and reliability of the vital sign monitoring result of the patient can be ensured.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a flowchart of a patient vital sign monitoring method based on artificial intelligence according to an embodiment of the present application.

Detailed Description

In order to better understand the above technical solutions, the following detailed description of the technical solutions of the present application is made by using the accompanying drawings and specific embodiments, and it should be understood that the specific features of the embodiments and the embodiments of the present application are detailed descriptions of the technical solutions of the present application, and not limiting the technical solutions of the present application, and the technical features of the embodiments and the embodiments of the present application may be combined with each other without conflict.

Referring to fig. 1, a patient vital sign monitoring method based on artificial intelligence is shown, which may include the following technical solutions described in step S101-step S104.

Step S101, when the vital sign monitoring data of the patient to be analyzed is obtained, carrying out data analysis on important indications of the vital sign monitoring data of the patient to be analyzed to obtain a pending data core.

Illustratively, patient vital sign monitoring data to be parsed includes: data obtained by a heart monitor, data obtained by a need monitor, data obtained by a pulse monitor, data obtained by a ventilator, etc. These data can characterize the patient's physical condition (e.g., ICU equipped with bedside monitors, central monitors, multifunctional respiratory therapy machines, anesthesia machines, electrocardiographs, defibrillators, pacemakers, infusion pumps, microinjectors, tracheal intubation, emergency equipment for tracheotomy, CPM articulation therapy caregivers, etc.).

For example, an important indication may be understood as key information in the patient vital sign monitoring data to be parsed, which can indicate whether the patient vital condition is a threatening related data (e.g., obtaining monitoring data from one monitoring instrument that may not be threatening to the patient vital sign, but obtaining monitoring data from several monitoring instruments may be threatening to the patient vital sign).

Further, the pending data core may be understood as a data set composed after parsing the important indication.

The pending data core is obtained when a data analysis thread is called and important indication of vital sign monitoring data of a patient to be analyzed is subjected to data analysis, and the pending data core can comprise N pending windows; n is an integer greater than 0. In addition, the data analysis thread is further configured to determine a prediction analysis type reliability coefficient corresponding to each pending window in the pending data core, where the prediction analysis type reliability coefficient refers to a probability that a classification type corresponding to the pending window belongs to a class of important indication, and a higher prediction analysis type reliability coefficient means a higher probability that the pending window contains the important indication. The thread input of the data analysis thread is patient vital sign monitoring data to be analyzed, the thread output of the data analysis thread is whether important indication exists in the patient vital sign monitoring data to be analyzed, if so, the AI vector positioning of the window where the important indication exists is output, and the format can be [ A, C, M ].

It can be understood that the data analysis thread can also automatically compress the vital sign monitoring data of the patient to be analyzed based on the AI vector positioning of the window to be determined, so as to output specific compressed vital sign data comprising important instructions. In other words, the data analysis thread in the embodiment of the present application is mainly responsible for analyzing the vital sign monitoring data of the patient to be analyzed with an extremely high coverage rate, so as to determine whether a suspected important indication exists in the vital sign monitoring data of the patient to be analyzed, if so, the constraint range (i.e. the window to be determined) where the important indication exists is compressed, and the constraint range is stored in a certain format, so that the vital sign data file can be obtained.

The data analysis thread is used as the first step in the data analysis flow in the embodiment of the application, the operation speed of the data analysis thread determines the operation speed of the whole flow to a great extent, but at the same time, the expression force of the data analysis thread should be strong enough to complete the rough analysis task. The second thread architecture is a full convolution thread architecture compared to the first thread architecture.

Based on the above, if the thread architecture of the invoked data analysis thread is the first thread architecture, the deriving or compressing process is performed on the vital sign monitoring data of the patient to be analyzed, so that the vital sign data amount of the vital sign monitoring data of the patient to be analyzed is derived or compressed into the fixed vital sign data amount allowed to be received by the first thread architecture, and further the derived or compressed vital sign monitoring data of the patient to be analyzed can be directly loaded to the data analysis thread, and through the data analysis thread, the important indication of the derived or compressed vital sign monitoring data of the patient to be analyzed is subjected to data analysis, and then a plurality of undetermined windows obtained after the data analysis are used as undetermined data cores.

If the thread architecture of the called data analysis thread is the second thread architecture, a great amount of time is not required to be consumed, the vital sign monitoring data of the patient to be analyzed can be directly loaded to the data analysis thread, the important indication of the vital sign monitoring data of the patient to be analyzed is subjected to data analysis through the data analysis thread, and a plurality of undetermined windows obtained after the data analysis are used as undetermined data cores.

Step S102, based on patient vital sign monitoring data to be analyzed, performing optimization processing on the to-be-determined data core to obtain an optimized data core corresponding to the to-be-determined data core and a compressed vital sign data set corresponding to the optimized data core.

By way of example, the optimization process may include data cleansing and data compensation, among other operations.

Specifically, the method can perform deduplication processing on the pending data core based on the duplicate data correction network to obtain the pending data core. Wherein, the data core to be processed comprises a pending window Ab; b is an integer not greater than M greater than 0; m is used to represent the total number of pending windows in the data core to be processed. Further, the vital sign data optimizing thread can be called, and the undetermined window Ab is optimized through the vital sign monitoring data of the patient to be analyzed to obtain an optimized window Cb and compressed vital sign data corresponding to the optimized window Cb; when the M optimization windows are obtained, the method can continue to perform de-duplication processing on the M optimization windows based on a de-duplication mode, the de-duplication processed optimization windows are used as optimization data cores corresponding to undetermined data cores, and compressed vital sign data corresponding to the de-duplication processed optimization windows are used as compressed vital sign data sets corresponding to the optimization data cores. Wherein the compressed vital sign dataset comprises a compressed vital sign data; the compressed vital sign data is obtained by compressing the vital sign monitoring data of the patient to be analyzed based on an optimization window in the optimization data core; a is an integer greater than 0.

It can be understood that the data analysis thread can effectively analyze the suspected important indication range in the vital sign monitoring data of the patient to be analyzed, but is limited by the performance of the thread and the configuration data amount, and the data analysis thread has the situation that some areas similar to the important indication are mistakenly analyzed as important indications, and meanwhile, due to the adoption of a method of carrying out data analysis by a moving window, the problem that the critical range is repeated inevitably occurs, and based on the problem, the data analysis thread needs to further carry out de-duplication processing on the undetermined data core based on a repeated data correction network.

If the pending data cores include N pending data cores, the method may correct the vital sign data compression rate specified by the network based on the repeated data, and perform compression processing on each of the N pending windows to obtain N compression windows. Further, the data analysis thread is further configured to determine the prediction analysis type trusted coefficients corresponding to each of the N pending windows, so that the N compression windows may be sequenced based on the N prediction analysis type trusted coefficients, thereby obtaining a distribution situation. The compression window carrying the maximum prediction analysis type trusted coefficient in the distribution case can be used as a first compression window, and (N-1) compression windows except the first compression window in the distribution case are used as a first set to be cleaned.

Then, the N compression windows can be subjected to de-duplication processing based on the repetition rate between the first compression window and each compression window in the first set to be cleaned, so as to obtain a reserved data core. Wherein it is understood that the repetition rate between the first compression window and each compression window in the first set to be cleaned can be determined separately. If the first to-be-cleaned set has a repeated compression window with the repetition rate larger than the target value of the repetition rate, the first compression window can be reserved, the repeated compression window is cleaned in the first to-be-cleaned set, then the compression window with the maximum prediction analysis type credible coefficient is taken as a second compression window in the cleaned first to-be-cleaned set, and the compression windows except the second compression window are taken as a second to-be-cleaned set. And then, the second compression window can be reserved, the repeated compression windows in the second to-be-cleaned set are subjected to de-duplication processing based on the repetition rate between the second compression window and each compression window in the second to-be-cleaned set until the second to-be-cleaned set after the de-duplication processing is empty, and the reserved first compression window and the reserved second compression window are used as reserved data cores.

Further, the vital sign data optimization thread needs to be called to perform offset prediction analysis on each pending window in the data core to be processed, so that the positions of each pending window are corrected iteratively. The thread body architecture of the vital sign data optimization thread can be based on convolution nerve threads or attention threads, and the like, and the thread body architecture of the vital sign data optimization thread is not limited herein.

The thread input of the vital sign data optimizing thread is AI vector positioning of the region coordinates of each undetermined window in the data core to be processed (namely AI vector positioning of the undetermined window) and the vital sign monitoring data of the patient to be analyzed, and the thread output of the vital sign data optimizing thread is the undetermined window (namely the optimizing window) which is subjected to optimizing processing. It can be appreciated that the vital sign data optimization thread can output the specific region coordinates where the optimization window is located (i.e., the AI vector location of the pending window), where the AI vector location format of the optimization window is [ a, C, M ]. For example, the vital sign data optimizing thread can be invoked, so that error quantity prediction analysis can be performed on the window Ab to be determined through patient vital sign monitoring data to be analyzed, and a first prediction coefficient can be obtained. Then, the to-be-determined window Ab may be subjected to optimization processing based on the first prediction coefficient, so as to obtain a first optimization window corresponding to the to-be-determined window Ab, and further whether the optimization is ended may be determined based on a relationship between the first prediction coefficient and the target value range of the prediction coefficient. The target value range of the prediction coefficient can be dynamically optimized according to the current requirement, and the target value range of the prediction coefficient is not limited herein.

If the first prediction coefficient belongs to the target value range of the prediction coefficient, the optimization can be determined to be ended, and then the first optimization window corresponding to the undetermined window Ab can be directly used as the optimization window Cb corresponding to the undetermined window Ab. Optionally, if the first prediction coefficient does not belong to the target value range of the prediction coefficient, the method may determine that the optimization processing needs to be further performed on the first optimization window, and may call the vital sign data optimization thread, perform error prediction analysis on the first optimization window corresponding to the to-be-determined window Ab through the vital sign monitoring data of the patient to be analyzed to obtain a second prediction coefficient, and further perform optimization processing on the first optimization window corresponding to the to-be-determined window Ab based on the second prediction coefficient to obtain a second optimization window corresponding to the to-be-determined window Ab until the second prediction coefficient belongs to the target value range of the prediction coefficient, and use the second optimization window corresponding to the to-be-determined window Ab as the optimization window Cb corresponding to the to-be-determined window Ab, and further perform compression processing on the vital sign monitoring data of the patient to be analyzed based on AI vector positioning of the optimization window Cb to obtain compressed vital sign data corresponding to the optimization window Cb.

It can be understood that, in order to effectively ensure the accuracy of the constraint range where the important indication is located, when the data cores are subjected to deduplication (i.e. first deduplication), the coefficient adopted is relatively loose (for example, the target value of the repetition rate is specified relatively large), that is, there may still be two undetermined windows overlapping to a certain extent in M undetermined windows in the data cores to be processed obtained after the first deduplication, after the vital sign data optimization thread performs optimization processing, the window to be determined where the important indication of two different objects is located is optimized farther, the window to be determined where the important indication of two identical objects is located is optimized closer, which means that there may still be a repeated optimization window in the optimized data cores to be processed, in order to effectively improve the efficiency of subsequent analysis, when M optimization windows are obtained, the deduplication needs to be performed again, and then the optimized window after the deduplication is used as the optimized data core corresponding to the undetermined data, the compressed vital sign data corresponding to the optimized window after the deduplication is used as the compressed vital sign data corresponding to the optimized data core, and the compressed vital sign data corresponding to the optimized data core is used as the compressed vital sign data set (for example, the vital sign data is input to the thread and classified).

Step S103, classifying each compressed vital sign data in the A compressed vital sign data to obtain a classification result matched with the important instruction.

By way of example, the compressed vital sign data are classified, so that the data can be effectively regulated, and the data can be stored or extracted more quickly and conveniently. And the data processing efficiency is improved.

Specifically, the vital sign data to be processed can be determined from the A compressed vital sign data, and then the vital sign data classification thread can be called, the vital sign data to be processed is read and processed according to the vital sign data classification type in the vital sign data classification thread to obtain the prediction analysis type credibility coefficient corresponding to the vital sign data to be processed, and then the vital sign monitoring type corresponding to the vital sign data to be processed is obtained by deleting the non-important indication of the vital sign data to be processed according to the vital sign data segmentation type in the vital sign data classification thread. If the credibility coefficient of the prediction analysis type corresponding to the vital sign data to be processed is larger than the credibility coefficient target value, the vital sign data to be processed is considered to belong to the classification type of the important indication, and at the moment, the vital sign monitoring type corresponding to the vital sign data to be processed can be used as the classification result matched with the important indication. Wherein the classification result comprises C vital sign monitoring categories; c is not more than A and C is an integer greater than 0.

For ease of understanding, the compressed vital sign data set may further include a compressed vital sign data, a is an integer greater than 0, and one compressed vital sign data is obtained by compressing patient vital sign monitoring data to be analyzed based on an optimization window in the optimization data core.

In order to learn more detailed information and distinguish the constraint range of the analysis which cannot be correctly performed in the first two steps, the vital sign data classification thread called in the embodiment of the application consists of a loss thread and a multi-task type. The loss thread may include a plurality of feature extraction units and a plurality of residual bottleneck architectures, and the pooling layer in the vital sign data classification thread may be an average pooling layer.

Similarly, the compressed vital sign data may also be loaded into a vital sign data classification thread. And reading and processing the compressed vital sign data through the vital sign data classification types to obtain a prediction analysis type credible coefficient corresponding to the compressed vital sign data. Meanwhile, deleting the non-important indication of the compressed vital sign data through the vital sign data segmentation type to obtain the vital sign monitoring type corresponding to the compressed vital sign data.

Step S104, in the optimized data core, debugging processing is respectively carried out on the optimized windows corresponding to the C vital sign monitoring types, C debugging windows used for representing the important indications in the vital sign monitoring data of the patient to be analyzed are obtained, and the vital sign monitoring result of the patient is obtained through the debugging windows of the important indications.

Specifically, when obtaining the C vital sign monitoring types in the classification result, it is necessary to determine, in the optimization data core, optimization windows corresponding to the C vital sign monitoring types, respectively, so that the AI vector positioning of the determined optimization windows may be debugged based on the AI vector positioning of the vital sign monitoring types, thereby obtaining C debug windows for representing important indications in the vital sign monitoring data of the patient to be analyzed.

In the embodiment of the application, the cascade thread group formed by connecting links of the data analysis thread, the repeated data correction network, the vital sign data optimization thread, the vital sign data classification thread and the like in series is used for completing the data analysis and the end-to-end closed loop design. According to the data analysis method, through optimization processing and comprehensive consideration of the relevance between the data analysis result and vital sign data, the positioning accuracy index of the constraint range where the important indication is located is highly valued, so that the debugging window capable of indicating the important indication of the vital sign monitoring data of the patient to be analyzed more accurately is obtained, the accuracy of data analysis is further effectively improved, and the accuracy and reliability of the vital sign monitoring result of the patient can be guaranteed.

Because the pending data core may have repeated pending windows, the iterative offset correction positioning frame stage can utilize the network based on repeated data correction to perform intelligent deduplication on the pending data core so as to reserve a plurality of most effective pending windows, thereby reducing subsequent calculation amount. Further, the undetermined data core (i.e. the to-be-processed data core) after the deduplication processing and the patient vital sign monitoring data to be analyzed can be loaded to the vital sign data optimization thread together, and the M undetermined windows in the to-be-processed data core are subjected to position offset fine tuning (i.e. optimization processing) iteratively until M constraint ranges (i.e. optimization windows) for representing optimal positioning of each important indication are obtained; m is an integer greater than 0. In order to improve the subsequent analysis efficiency, the de-duplication processing can be performed on the M undetermined windows again, so as to obtain an optimized data core and a compressed vital sign data set corresponding to the optimized data core. The compressed vital sign data in the compressed vital sign data set are obtained after compression processing is performed on the vital sign monitoring data of the patient to be analyzed based on an optimization window in an optimization data core.

For example, the a compressed vital sign data in the compressed vital sign data set may be loaded to the vital sign data classification thread, so that the vital sign monitoring category including the category belonging to the important indication may be determined through the vital sign data classification type and the vital sign data segmentation type in the vital sign data classification thread, and a classification result matched with the important indication may be obtained. Wherein the classification result herein may include C vital sign monitoring categories; c is not more than A and C is an integer greater than 0; a is an integer not greater than M and greater than 0.

Finally, in the optimized data core, debugging processing can be performed on the optimization windows corresponding to the C vital sign monitoring types, so that the debugged data core can be obtained. Wherein the debug data core here may comprise C debug windows for representing important indications in the patient vital sign monitoring data to be parsed.

It will be appreciated that the input to the analysis system may be patient vital sign monitoring data to be analyzed that is required to be analyzed, and the output of the system may be an important indication of whether or not the patient vital sign monitoring data currently to be analyzed is present. If the data exists, outputting the constraint range (namely the debugging window) where the important indication exists and the foreground vital sign data of the data level corresponding to each debugging window.

Further, the method may be executed by a server carrying a thread configuration function, which may be a terminal device or a thread configuration function, and is not limited herein. The method may include at least the following description of step S201-step S203.

Step S201 obtains first example vital sign data for configuring the first Ab resolution thread, and a first example catalog for representing a current category of the first example vital sign data.

Step S202, a first Ab analysis thread is called to conduct data analysis on the first example vital sign data, and a prediction analysis type credible coefficient of the first example vital sign data aiming at important instructions is obtained.

Step S203, configuring the first Ab analysis thread based on the prediction analysis type trusted coefficient of the first example vital sign data and the current type of the first example vital sign data, to obtain a data analysis thread for performing data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed.

Specifically, the method may perform evaluation index calculation processing on the prediction analysis type trusted coefficient of the first example vital sign data and the current type of the first example vital sign data, determine a thread evaluation index of the first Ab analysis thread, and further configure the first Ab analysis thread based on the thread evaluation index of the first Ab analysis thread, to obtain a first thread configuration result. If the first thread configuration result indicates that the configured first Ab analysis thread meets the first thread Cheng Shoulian requirement, the first Ab analysis thread meeting the first thread convergence requirement can be used as a second Ab analysis thread, so that the thread architecture of the second Ab analysis thread can be subjected to architecture analysis, and a data analysis thread for carrying out data analysis on important instructions of vital sign monitoring data of a patient to be analyzed is determined.

If the thread architecture of the first Ab analysis thread designed in the embodiment of the present application is a first thread architecture, the first Ab analysis thread may be configured, and when the first Ab analysis thread (i.e., the second Ab analysis thread) meeting the thread convergence requirement is obtained, the second Ab analysis thread may be directly used as a final data analysis thread.

If the first thread configuration result indicates that the configured first Ab analysis thread meets the first thread Cheng Shoulian requirement, the first Ab analysis thread meeting the first thread convergence requirement can be used as a second Ab analysis thread, and then the second Ab analysis thread can be directly used as a data analysis thread for performing data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed.

In the embodiment of the application, a scheme of a data analysis thread based on a full convolution thread is provided, and the scheme can carry out moving window analysis on vital sign data under any data volume and detect important instructions from the vital sign data as much as possible with extremely high coverage rate. The design is in accordance with task characteristics, important indications of any data quantity in vital sign data of any data quantity can be detected in a very short time, good pre-selection is provided for subsequent procedures, the overall operation efficiency of the system is improved, and a large number of invalid calculations are avoided.

Further, the method may be executed by a server carrying a thread configuration function, which may be a terminal device or a thread configuration function, and is not limited herein. The method may include at least the following description of step S301-step S305.

Step S301 obtains second example vital sign data for configuring the original optimization thread and a second example catalog of current AI vector locations for representing important indications of the second example vital sign data.

In order to obtain enough offset pattern classification capability and reduce the computational complexity as much as possible, the embodiment of the application can analyze the thread architecture of the thread in data.

Step S302, calling an original optimization thread, and performing offset prediction analysis processing on the second example vital sign data to obtain a prediction analysis prediction coefficient of the second example vital sign data.

Step S303, based on the prediction analysis prediction coefficient, the AI vector positioning of the second example vital sign data is optimized to obtain the prediction analysis AI vector positioning of the second example vital sign data.

Specifically, based on each item of prediction analysis error amount in the prediction analysis prediction coefficient of the second example vital sign data, the corresponding coordinates in the AI vector positioning of the second example vital sign data are respectively subjected to optimization processing, and the prediction analysis AI vector positioning of the second example vital sign data is determined.

And step S304, configuring an original optimization thread based on the prediction analysis AI vector positioning and the current AI vector positioning to obtain a second thread configuration result.

Specifically, the thread evaluation index of the original optimization thread may be determined based on the prediction analysis AI vector location and the current AI vector location indicated by the second exemplary directory, and then the original optimization thread may be configured based on the thread evaluation index of the original optimization thread, so as to obtain a second thread configuration result.

Wherein, for the region where the important indication of the second example vital sign data is located, the original optimization thread needs to predictively resolve the error amount between it and the nearest constraint range (e.g., the prediction resolution error amount corresponding to the upper left corner of the constraint range, the prediction resolution error amount corresponding to the height, and the prediction resolution error amount corresponding to the width). The learning objective is formulated as a predictive question and the thread evaluation index of the original optimization thread may be determined based on the example evaluation index of each of the examples in the second example vital sign data.

Wherein b is used to represent the b-th example in the second example vital sign data, and is used to represent the predictive interpretation AI vector positioning of the b-th example determined based on the original optimization thread; a predictive interpretation AI vector location for representing a b-th instance determined based on the second instance catalog.

In step S305, if the second thread configuration result indicates that the configured original optimization thread meets the second thread convergence requirement, the original optimization thread meeting the second thread convergence requirement is used as the vital sign data optimization thread.

The second thread convergence requirement here means that the current thread evaluation index reaches a specified target value or less or the number of configuration steps reaches a specified target value or more. It can be understood that the thread architecture of the vital sign data optimization thread can be a position correction thread of an important indication area, and the thread is used for correcting vital sign monitoring data of a patient to be analyzed through prediction coefficients of a constraint range where important indications are located after feature extraction is performed on the vital sign monitoring data.

Further, the method may be executed by a server carrying a thread configuration function, which may be a terminal device or a thread configuration function, and is not limited herein. The method may include at least the following description of step S401 to step S405.

In step S401, the sample data for configuring the original classification thread and the sample directory corresponding to the sample data are obtained.

However, the configuration set of the two tasks is incompatible, because for the vital sign data splitting task, if the vital sign data of the non-important indication is introduced, the splitting mask of the vital sign data of the non-important indication will be set to the full-view non-important indication, in which case the evaluation index function in the configuration process will increase several times, which will greatly affect the correct splitting of the important indication. In view of the above, the embodiment of the present application performs these two tasks in a phased configuration manner, that is, the example data herein may include third example vital sign data for configuring the vital sign data partition type, and fourth example vital sign data for configuring the vital sign data classification type, and the example catalog herein may include a third example catalog (i.e., for representing current vital sign occlusion data) corresponding to the third example vital sign data and a fourth example catalog (i.e., for representing current category) corresponding to the fourth example vital sign data.

Step S402, configuring the loss thread and the vital sign data segmentation type in the original classification thread based on the third example vital sign data in the example data and the third example catalog for representing the current vital sign occlusion data of the third example vital sign data, to obtain the first classification thread.

Step S403, configuring the loss thread and the vital sign data classification type in the first read thread based on the fourth example vital sign data in the example data and the fourth example catalog for representing the current category of the fourth example vital sign data, to obtain the second classification thread.

Step S404, configuring the loss thread and the vital sign data partition type in the second classification thread based on the third exemplary vital sign data and the third exemplary directory, to obtain a third classification thread.

Specifically, after step S403 is performed, the loss thread and the vital sign data segmentation type in the second reading thread are repeatedly configured with a smaller learning rate based on the third exemplary vital sign data and the third exemplary directory until the evaluation index is stable.

For example, the vital sign data segmentation type of the second classification thread can be called, the non-important indication of the vital sign data of the third example is deleted, the predicted analysis vital sign shielding data corresponding to the vital sign data of the third example is obtained, the data of the vital sign data of the third example can be traversed, and the traversed data is used as data to be processed. At this time, the trusted coefficient of the to-be-processed data in the current vital sign shielding data indicated by the third exemplary directory may be used as a first trusted coefficient, the trusted coefficient of the to-be-processed data in the predicted and parsed vital sign shielding data in the to-be-processed data may be used as a second trusted coefficient, and further, the thread evaluation index of the second classification thread may be determined based on the above formula (11) and performing evaluation index calculation processing on the first trusted coefficient and the second trusted coefficient. Then, the method can configure the loss thread and vital sign data segmentation type in the second classification thread based on the thread evaluation index of the second classification thread to obtain a third thread configuration result, and if the third thread configuration result indicates that the configured second classification thread meets the first type convergence requirement in the third thread Cheng Shoulian requirement, the second classification thread meeting the first type convergence requirement is used as a third classification thread. The first type of convergence requirement here may refer to that the thread evaluation index reaches a stability.

Step S405 locks the loss thread and the vital sign data segmentation type in the third classification thread, and configures the vital sign data classification type in the locked third classification thread based on the fourth example vital sign data and the fourth example directory, to obtain the vital sign data classification thread.

Specifically, the loss thread and vital sign data segmentation type in the third classification thread can be locked, and the locked third classification thread can be used as a fourth classification thread. At this time, the fourth classification thread may be called, and the fourth example vital sign data is read according to the vital sign data classification type in the fourth classification thread, so as to determine an example prediction analysis trusted coefficient corresponding to the fourth example vital sign data, and the current category indicated by the fourth directory and the example prediction analysis trusted coefficient are subjected to evaluation index calculation so as to determine a thread evaluation index of the fourth classification thread. At this time, the vital sign data classification type in the fourth classification thread can be configured based on the thread evaluation index of the fourth classification thread to obtain a fourth thread configuration result; and if the fourth thread configuration result indicates that the configured fourth classification thread meets the second type convergence requirement in the third line Cheng Shoulian requirements, taking the fourth classification thread meeting the second type convergence requirement as the vital sign data classification thread. The second type of convergence requirement here may also mean that the thread evaluation index reaches stability.

According to the embodiment of the application, through the staged configuration method, the loss thread in the finally obtained vital sign data classification thread can effectively express the characteristics of the input vital sign data for both tasks, and after the characteristics are transmitted to the respective types, the respective tasks can be well executed. Based on this, embodiments of the present application provide for combining these two tasks in one thread, and phased joint configuration in configuration, which enables the thread to learn important indication features that compromise both tasks. In the current application process of the threads, the vital sign data segmentation type in the vital sign data classification threads can utilize vital sign data information in compressed vital sign data to clean the background or other non-main important indications in the compressed vital sign data, so that the important indications in vital sign monitoring types are clearer and purer, vital sign data information of a data level can be output, in addition, the vital sign data classification type in the vital sign data classification threads can also conduct data analysis again so as to predict and analyze classification types of an optimization window corresponding to the compressed vital sign data, and further accuracy of data analysis can be improved.

On the basis of the above, there is provided an artificial intelligence based patient vital sign monitoring device, the device comprising:

the data core obtaining module is used for carrying out data analysis on important indications of vital sign monitoring data of a patient to be analyzed when the vital sign monitoring data of the patient to be analyzed is obtained, so as to obtain a pending data core;

the data compression module is used for carrying out optimization processing on the undetermined data core according to the vital sign monitoring data of the patient to be analyzed to obtain an optimized data core corresponding to the undetermined data core and a compressed vital sign data set corresponding to the optimized data core; the compressed vital sign dataset includes a compressed vital sign data; the compressed vital sign data are obtained by compressing the vital sign monitoring data of the patient to be analyzed according to an optimization window in the optimization data core; a is an integer greater than 0;

the result classification module is used for respectively classifying each compressed vital sign data in the A compressed vital sign data to obtain a classification result matched with the important indication; the classification result comprises C vital sign monitoring categories; c is not more than A and C is an integer greater than 0;

And the result monitoring module is used for respectively debugging the optimization windows corresponding to the C vital sign monitoring types in the optimization data core to obtain C debugging windows used for representing the important indications in the vital sign monitoring data of the patient to be analyzed, and obtaining the vital sign monitoring result of the patient through the debugging windows of the important indications.

On the above basis, an artificial intelligence based patient vital sign monitoring system is shown comprising a processor and a memory in communication with each other, the processor being adapted to read a computer program from the memory and execute it for carrying out the method as described above.

On the basis of the above, there is also provided a computer readable storage medium on which a computer program stored which, when run, implements the above method.

In summary, based on the above scheme, when the vital sign monitoring data of the patient to be analyzed is obtained, carrying the data analysis function, the important indication of the vital sign monitoring data of the patient to be analyzed can be subjected to data analysis, so as to obtain the undetermined data core. Wherein the pending data core is formed by the constraint range where the originally identified important indicator is located. In order to effectively ensure the accuracy of the constraint range of the important indication, after the pending data core is obtained, secondary optimization processing (including de-optimization processing and debugging processing) can be performed on the pending data core. Firstly, the method can perform optimization processing on the to-be-determined data core based on the to-be-analyzed patient vital sign monitoring data to obtain an optimized data core corresponding to the to-be-determined data core and a compressed vital sign data set corresponding to the optimized data core. Wherein the compressed vital sign dataset herein comprises a compressed vital sign data; the compressed vital sign data is obtained by compressing the vital sign monitoring data of the patient to be analyzed based on an optimization window in the optimization data core; a is an integer greater than 0. Then, the correlation between the data analysis result and the vital sign data can be comprehensively considered, the undetermined data core (i.e. the optimized data core) after the optimization processing is performed again, namely, the classification processing is performed on each compressed vital sign data in the A compressed vital sign data respectively to obtain a classification result matched with the important indication, and further, in the optimized data core, the debugging processing can be performed on optimization windows corresponding to C vital sign monitoring types included in the classification result respectively, so as to accurately obtain C constraint ranges (i.e. debugging windows) for representing the important indication in the vital sign monitoring data of the patient to be analyzed. Wherein C is not more than A and C is an integer of more than 0. Therefore, according to the data analysis method provided by the embodiment of the application, through optimization processing and comprehensive consideration of the correlation between the data analysis result and vital sign data, the positioning accuracy index of the constraint range where the important indication is located is highly emphasized, so that the debugging window capable of more accurately indicating the important indication of the vital sign monitoring data of the patient to be analyzed is obtained, the accuracy of data analysis is further effectively improved, and the accuracy and reliability of the vital sign monitoring result of the patient can be ensured.

It should be appreciated that the systems and modules thereof shown above may be implemented in a variety of ways. For example, in some embodiments, the system and its modules may be implemented in hardware, software, or a combination of software and hardware. Wherein the hardware portion may be implemented using dedicated logic; the software portions may then be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or special purpose design hardware. Those skilled in the art will appreciate that the methods and systems described above may be implemented using computer executable instructions and/or embodied in processor control code, such as provided on a carrier medium such as a magnetic disk, CD or DVD-ROM, a programmable memory such as read only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The system of the present application and its modules may be implemented not only with hardware circuitry such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but also with software executed by various types of processors, for example, and with a combination of the above hardware circuitry and software (e.g., firmware).

It should be noted that, the advantages that may be generated by different embodiments may be different, and in different embodiments, the advantages that may be generated may be any one or a combination of several of the above, or any other possible advantages that may be obtained.

While the basic concepts have been described above, it will be apparent to those skilled in the art that the foregoing detailed disclosure is by way of example only and is not intended to be limiting. Although not explicitly described herein, various modifications, improvements and adaptations of the application may occur to one skilled in the art. Such modifications, improvements, and modifications are intended to be suggested within the present disclosure, and therefore, such modifications, improvements, and adaptations are intended to be within the spirit and scope of the exemplary embodiments of the present disclosure.

Meanwhile, the present application uses specific words to describe embodiments of the present application. Reference to "one embodiment," "an embodiment," and/or "some embodiments" means that a particular feature, structure, or characteristic is associated with at least one embodiment of the application. Thus, it should be emphasized and should be appreciated that two or more references to "an embodiment" or "one embodiment" or "an alternative embodiment" in various positions in this specification are not necessarily referring to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of the application may be combined as suitable.

Furthermore, those skilled in the art will appreciate that the various aspects of the application are illustrated and described in the context of a number of patentable categories or circumstances, including any novel and useful procedures, machines, products, or materials, or any novel and useful modifications thereof. Accordingly, aspects of the application may be performed entirely by hardware, entirely by software (including firmware, resident software, micro-code, etc.) or by a combination of hardware and software. The above hardware or software may be referred to as a "data block," module, "" engine, "" unit, "" component, "or" system. Furthermore, aspects of the application may take the form of a computer product, comprising computer-readable program code, embodied in one or more computer-readable media.

The computer storage medium may contain a propagated data signal with the computer program code embodied therein, for example, on a baseband or as part of a carrier wave. The propagated signal may take on a variety of forms, including electro-magnetic, optical, etc., or any suitable combination thereof. A computer storage medium may be any computer readable medium that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code located on a computer storage medium may be propagated through any suitable medium, including radio, cable, fiber optic cable, RF, or the like, or a combination of any of the foregoing.

The computer program code necessary for operation of portions of the present application may be written in any one or more programming languages, including an object oriented programming language such as Java, scala, smalltalk, eiffel, JADE, emerald, C ++, c#, vb net, python, etc., a conventional programming language such as C language, visual Basic, fortran 2003, perl, COBOL 2002, PHP, ABAP, dynamic programming languages such as Python, ruby and Groovy, or other programming languages, etc. The program code may execute entirely on the user's computer or as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any form of network, such as a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet), or the use of services such as software as a service (SaaS) in a cloud computing environment.

Furthermore, the order in which the elements and sequences are presented, the use of numerical letters, or other designations are used in the application is not intended to limit the sequence of the processes and methods unless specifically recited in the claims. While certain presently useful inventive embodiments have been discussed in the foregoing disclosure, by way of example, it is to be understood that such details are merely illustrative and that the appended claims are not limited to the disclosed embodiments, but, on the contrary, are intended to cover all modifications and equivalent arrangements included within the spirit and scope of the embodiments of the application. For example, while the system components described above may be implemented by hardware devices, they may also be implemented solely by software solutions, such as installing the described system on an existing server or mobile device.

Similarly, it should be noted that in order to simplify the description of the present disclosure and thereby aid in understanding one or more inventive embodiments, various features are sometimes grouped together in a single embodiment, figure, or description thereof. This method of disclosure, however, is not intended to imply that more features than are required by the subject application. Indeed, less than all of the features of a single embodiment disclosed above.

In some embodiments, numbers describing the components, number of attributes are used, it being understood that such numbers being used in the description of embodiments are modified in some examples by the modifier "about," approximately, "or" substantially. Unless otherwise indicated, "about," "approximately," or "substantially" indicate that the numbers allow for adaptive variation. Accordingly, in some embodiments, numerical parameters set forth in the specification and claims are approximations that may vary depending upon the desired properties sought to be obtained by the individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and employ a method for preserving the general number of digits. Although the numerical ranges and parameters set forth herein are approximations in some embodiments for use in determining the breadth of the range, in particular embodiments, the numerical values set forth herein are as precisely as possible.

Each patent, patent application publication, and other material, such as articles, books, specifications, publications, documents, etc., cited herein is hereby incorporated by reference in its entirety. Except for the application history file that is inconsistent or conflicting with this disclosure, the file (currently or later attached to this disclosure) that limits the broadest scope of the claims of this disclosure is also excluded. It is noted that the description, definition, and/or use of the term in the appended claims controls the description, definition, and/or use of the term in this application if there is a discrepancy or conflict between the description, definition, and/or use of the term in the appended claims.

Finally, it should be understood that the embodiments described herein are merely illustrative of the principles of the embodiments of the present application. Other variations are also possible within the scope of the application. Thus, by way of example, and not limitation, alternative configurations of embodiments of the application may be considered in keeping with the teachings of the application. Accordingly, the embodiments of the present application are not limited to the embodiments explicitly described and depicted herein.

The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and variations of the present application will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the application are to be included in the scope of the claims of the present application.

Claims

1. A method for monitoring vital signs of a patient based on artificial intelligence, the method comprising:

when vital sign monitoring data of a patient to be analyzed are obtained, carrying out data analysis on important indications of the vital sign monitoring data of the patient to be analyzed to obtain a to-be-determined data core;

performing optimization processing on the undetermined data core according to the vital sign monitoring data of the patient to be analyzed to obtain an optimized data core corresponding to the undetermined data core and a compressed vital sign data set corresponding to the optimized data core; the compressed vital sign dataset includes a compressed vital sign data; the compressed vital sign data are obtained by compressing the vital sign monitoring data of the patient to be analyzed according to an optimization window in the optimization data core; a is an integer greater than 0;

classifying each compressed vital sign data in the A compressed vital sign data to obtain a classification result matched with the important indication; the classification result comprises C vital sign monitoring categories; c is not more than A and C is an integer greater than 0;

and in the optimized data core, respectively debugging the optimized windows corresponding to the C vital sign monitoring types to obtain C debugging windows used for representing the important indication in the vital sign monitoring data of the patient to be analyzed, and obtaining the vital sign monitoring result of the patient through the debugging windows of the important indication.

2. The method of claim 1, wherein prior to data parsing the vital sign monitoring data of the patient to be parsed to obtain a pending data core, the method further comprises:

obtaining first example vital sign data for configuring a first Ab resolution thread, and a first example catalog for representing a current category of the first example vital sign data; the first example vital sign data is obtained by preprocessing an important indication of the original example vital sign data;

invoking the first Ab analysis thread to perform data analysis on the first example vital sign data to obtain a prediction analysis type credible coefficient of the first example vital sign data aiming at the important indication;

and configuring the first Ab analysis thread according to the prediction analysis type credible coefficient of the first example vital sign data and the current type of the first example vital sign data to obtain a data analysis thread for carrying out data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed.

3. The method according to claim 2, wherein configuring the first Ab analysis thread according to the predicted analysis type confidence coefficient of the first example vital sign data and the current type of the first example vital sign data to obtain a data analysis thread for performing data analysis on the vital sign monitoring data of the patient to be analyzed includes:

Performing evaluation index calculation processing on the prediction analysis type credible coefficient of the first example vital sign data and the current type of the first example vital sign data, and determining a thread evaluation index of the first Ab analysis thread; configuring the first Ab analysis thread according to the thread evaluation index of the first Ab analysis thread to obtain a first thread configuration result;

if the first thread configuration result indicates that the configured first Ab analysis thread meets the first thread Cheng Shoulian requirement, taking the first Ab analysis thread meeting the first thread convergence requirement as a second Ab analysis thread;

and carrying out architecture analysis on the thread architecture of the second Ab analysis thread, and determining a data analysis thread for carrying out data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed.

4. A method according to claim 3, wherein said performing architectural resolution on the thread architecture of the second Ab resolution thread, determining a data resolution thread for data resolution of the vital sign monitoring data of the patient to be resolved, comprises:

performing architecture analysis on the thread architecture of the second Ab analysis thread to obtain an analysis result;

If the analysis result indicates that a merging unit exists in the thread architecture of the second Ab analysis thread, switching the merging unit into a first feature extraction unit configured with a sliding step length;

assigning a value to the first feature extraction unit according to the thread coefficient of the sum unit to obtain a second feature extraction unit;

and taking a second Ab analysis thread comprising the second feature extraction unit as a data analysis thread for carrying out data analysis on the important indication of the vital sign monitoring data of the patient to be analyzed.

5. The method according to claim 1, wherein the performing, according to the patient vital sign monitoring data to be analyzed, a de-optimization process on the pending data core to obtain an optimized data core corresponding to the pending data core and a compressed vital sign data set corresponding to the optimized data core includes:

performing de-duplication processing on the undetermined data core based on a repeated data correction network to obtain a data core to be processed; the data core to be processed comprises a window Ab to be determined; b is an integer not greater than M greater than 0; m is used for representing the total number of pending windows in the data core to be processed;

Invoking a vital sign data optimization thread, and optimizing the undetermined window Ab through the vital sign monitoring data of the patient to be analyzed to obtain an optimization window Cb and compressed vital sign data corresponding to the optimization window Cb; when M optimizing windows are obtained, carrying out de-duplication treatment on the M optimizing windows according to a de-duplication mode;

and taking the optimized window after the de-duplication processing as an optimized data core corresponding to the undetermined data core, and taking the compressed vital sign data corresponding to the optimized window after the de-duplication processing as a compressed vital sign data set corresponding to the optimized data core.

6. The method of claim 5, wherein the pending data core is obtained when invoking a data resolution thread to perform data resolution on an important indication of vital sign monitoring data of the patient to be resolved; the pending data core comprises N pending windows; n is an integer greater than 0; the data analysis thread is further used for determining the prediction analysis type credibility coefficient corresponding to each of the N undetermined windows; the duplicate data correction network is used for carrying out duplicate removal processing on the undetermined data core to obtain the undetermined data core, and the duplicate data correction network comprises the following steps:

Based on the vital sign data compression rate specified by the repeated data correction network, respectively compressing each of the N undetermined windows to obtain N compressed windows; based on N prediction analysis type credible coefficients, sorting the N compression windows to obtain a distribution condition;

taking a compression window carrying the maximum prediction analysis type credible coefficient in the distribution situation as a first compression window, and taking (N-1) compression windows except the first compression window in the distribution situation as a first to-be-cleaned set;

performing de-duplication processing on the N compression windows according to the repetition rate between the first compression window and each compression window in the first set to be cleaned to obtain a reserved data core;

respectively carrying out vital sign data derivatization processing on each compression window in the reserved data core according to the vital sign data compression rate to obtain a data core to be processed;

the step of performing deduplication processing on the N compression windows according to the repetition rate between the first compression window and each compression window in the first set to be cleaned to obtain a reserved data core includes:

Determining repetition rates between the first compression window and each compression window in the first set to be cleaned respectively;

if the first to-be-cleaned set has a repeated compression window with the repetition rate larger than the target value of the repetition rate, reserving the first compression window, and cleaning the repeated compression window in the first to-be-cleaned set;

taking a compression window carrying the maximum prediction analysis type credible coefficient as a second compression window in the first set to be cleaned after cleaning, and taking compression windows except the second compression window as a second set to be cleaned;

and reserving the second compression window, performing de-duplication treatment on the repeated compression windows in the second to-be-cleaned set according to the repetition rate between the second compression window and each compression window in the second to-be-cleaned set until the second to-be-cleaned set subjected to de-duplication treatment is empty, and taking the reserved first compression window and the reserved second compression window as reserved data cores.

7. The method according to claim 5, wherein the invoking the vital sign data optimization thread to optimize the pending window Ab by the patient vital sign monitoring data to be parsed to obtain an optimized window Cb and compressed vital sign data corresponding to the optimized window Cb includes:

Invoking the vital sign data optimization thread, and performing error quantity prediction analysis on the undetermined window Ab through the vital sign monitoring data of the patient to be analyzed to obtain a first prediction coefficient;

optimizing the undetermined window Ab according to the first prediction coefficient to obtain a first optimized window corresponding to the undetermined window Ab;

if the first prediction coefficient belongs to the target value range of the prediction coefficient, taking the first optimization window corresponding to the undetermined window Ab as an optimization window Cb corresponding to the undetermined window Ab;

compressing the vital sign monitoring data of the patient to be analyzed according to the AI vector positioning of the optimization window Cb to obtain compressed vital sign data corresponding to the optimization window Cb;

after obtaining the first optimized window corresponding to the undetermined window Ab, the method further includes:

if the first prediction coefficient does not belong to the target value range of the prediction coefficient, calling the vital sign data optimization thread, and performing error quantity prediction analysis on a first optimization window corresponding to the undetermined window Ab through vital sign monitoring data of the patient to be analyzed to obtain a second prediction coefficient;

According to the second prediction coefficient, performing optimization processing on a first optimization window corresponding to the undetermined window Ab to obtain a second optimization window corresponding to the undetermined window Ab;

and taking the second optimization window corresponding to the undetermined window Ab as the optimization window Cb corresponding to the undetermined window Ab until the second prediction coefficient belongs to the target value range of the prediction coefficient.

8. The method of claim 5, wherein prior to invoking the vital sign data optimization thread, the method further comprises:

obtaining second example vital sign data for configuring the original optimization thread, and a second example catalog for representing a current AI vector location where an important indication of the second example vital sign data is located;

invoking the original optimization thread, and performing offset prediction analysis processing on the second example vital sign data to obtain a prediction analysis prediction coefficient of the second example vital sign data; optimizing the AI vector positioning of the second example vital sign data according to the prediction analysis prediction coefficient to obtain the prediction analysis AI vector positioning of the second example vital sign data;

Configuring the original optimization thread according to the prediction analysis AI vector positioning and the current AI vector positioning to obtain a second thread configuration result;

and if the second thread configuration result indicates that the configured original optimization thread meets the second thread convergence requirement, taking the original optimization thread meeting the second thread convergence requirement as the vital sign data optimization thread.

9. The method according to claim 1, wherein the classifying each of the a compressed vital sign data to obtain a classification result matching the importance indication comprises:

determining vital sign data to be processed from the a compressed vital sign data; invoking a vital sign data classification thread, and reading and processing the vital sign data to be processed according to the vital sign data classification type in the vital sign data classification thread to obtain a prediction analysis type credible coefficient corresponding to the vital sign data to be processed;

deleting the non-important indication of the vital sign data to be processed through the vital sign data segmentation type in the vital sign data classification thread to obtain a vital sign monitoring type corresponding to the vital sign data to be processed;

If the credibility coefficient of the predictive analysis type corresponding to the vital sign data to be processed is larger than the credibility coefficient target value, taking the vital sign monitoring type corresponding to the vital sign data to be processed as a classification result matched with the important indication;

wherein, before invoking the vital sign data classification thread, the method further comprises:

obtaining example data for configuring an original classification thread and an example catalog corresponding to the example data;

configuring a loss thread and a vital sign data segmentation type in the original classification thread according to third example vital sign data in the example data and a third example catalog used for representing current vital sign shielding data of the third example vital sign data to obtain a first classification thread;

according to the fourth example vital sign data in the example data and a fourth example catalog used for representing the current type of the fourth example vital sign data, configuring a loss thread and a vital sign data classification type in a first reading thread to obtain a second classification thread;

configuring a loss thread and vital sign data segmentation type in the second classification thread according to the third example vital sign data and a third example catalog to obtain a third classification thread;

Locking the loss thread and the vital sign data segmentation type in the third classification thread, and configuring the vital sign data classification type in the locked third classification thread according to the fourth example vital sign data and the fourth example catalog to obtain a vital sign data classification thread;

the configuring the loss thread and the vital sign data segmentation type in the second classification thread according to the third example vital sign data and the third example catalog to obtain a third classification thread includes:

invoking the second classification thread to delete the non-important indication of the third example vital sign data to obtain prediction analysis vital sign shielding data corresponding to the third example vital sign data;

traversing the data of the third example vital sign data, and taking the traversed data as data to be processed;

taking the credible coefficient of the to-be-processed data in the current vital sign shielding data indicated by the third example catalog as a first credible coefficient, and taking the credible coefficient of the to-be-processed data in the predictive analysis vital sign shielding data in the to-be-processed data as a second credible coefficient;

Performing evaluation index calculation processing on the first trusted coefficient and the second trusted coefficient, and determining a thread evaluation index of the second classification thread;

configuring a loss thread and vital sign data segmentation type in the second classification thread according to the thread evaluation index of the second classification thread to obtain a third thread configuration result;

if the third thread configuration result indicates that the configured second classification thread meets the first type convergence requirement in the third thread Cheng Shoulian requirements, taking the second classification thread meeting the first type convergence requirement as a third classification thread;

the locking the loss thread and the vital sign data segmentation type in the third classification thread, and configuring the vital sign data classification type in the locked third classification thread according to the fourth example vital sign data and the fourth example catalog to obtain a vital sign data classification thread, which comprises the following steps:

locking a loss thread and vital sign data segmentation type in the third classification thread, and taking the locked third classification thread as a fourth classification thread;

invoking the fourth classification thread, reading the fourth example vital sign data, and determining an example prediction analysis trusted coefficient corresponding to the fourth example vital sign data; performing evaluation index calculation processing on the current category indicated by the fourth catalog and the sample prediction analysis credibility coefficient, and determining a thread evaluation index of the fourth classification thread;

Configuring vital sign data classification types in the fourth classification thread according to the thread evaluation index of the fourth classification thread to obtain a fourth thread configuration result;

and if the fourth thread configuration result indicates that the configured fourth classification thread meets the second type convergence requirement in the third line Cheng Shoulian requirements, taking the fourth classification thread meeting the second type convergence requirement as a vital sign data classification thread.

10. An artificial intelligence based patient vital sign monitoring system comprising a processor and a memory in communication with each other, the processor being adapted to read a computer program from the memory and execute the computer program to implement the method of any of claims 1-9.