CN117954116B - Breathing severe patient state monitoring and early warning method based on artificial intelligence - Google Patents
Breathing severe patient state monitoring and early warning method based on artificial intelligence Download PDFInfo
- Publication number
- CN117954116B CN117954116B CN202410354888.2A CN202410354888A CN117954116B CN 117954116 B CN117954116 B CN 117954116B CN 202410354888 A CN202410354888 A CN 202410354888A CN 117954116 B CN117954116 B CN 117954116B
- Authority
- CN
- China
- Prior art keywords
- abnormal
- monitoring
- type
- interval
- suspected
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 320
- 238000000034 method Methods 0.000 title claims abstract description 61
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 22
- 230000029058 respiratory gaseous exchange Effects 0.000 title description 3
- 230000002159 abnormal effect Effects 0.000 claims abstract description 467
- 238000005070 sampling Methods 0.000 claims abstract description 49
- 230000005856 abnormality Effects 0.000 claims abstract description 34
- 230000035772 mutation Effects 0.000 claims abstract description 24
- 230000000241 respiratory effect Effects 0.000 claims abstract description 14
- 230000000875 corresponding effect Effects 0.000 claims description 76
- 238000004458 analytical method Methods 0.000 claims description 59
- 238000010606 normalization Methods 0.000 claims description 12
- 208000023504 respiratory system disease Diseases 0.000 claims description 5
- 230000002596 correlated effect Effects 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 230000001186 cumulative effect Effects 0.000 claims description 3
- 208000028399 Critical Illness Diseases 0.000 claims 10
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 6
- 229910052760 oxygen Inorganic materials 0.000 description 6
- 239000001301 oxygen Substances 0.000 description 6
- 238000012545 processing Methods 0.000 description 5
- 230000036387 respiratory rate Effects 0.000 description 4
- 230000035945 sensitivity Effects 0.000 description 4
- 230000036772 blood pressure Effects 0.000 description 3
- 239000008280 blood Substances 0.000 description 2
- 210000004369 blood Anatomy 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 230000001960 triggered effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000036391 respiratory frequency Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording for evaluating the cardiovascular system, e.g. pulse, heart rate, blood pressure or blood flow
- A61B5/0205—Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Public Health (AREA)
- Medical Informatics (AREA)
- Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Physiology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Biophysics (AREA)
- Cardiology (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychiatry (AREA)
- Pulmonology (AREA)
- Signal Processing (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention relates to the technical field of multidimensional data anomaly monitoring, in particular to an artificial intelligence-based respiratory severe patient state monitoring and early warning method. According to the method, continuous changes, fluctuation amplitude and fluctuation mutation of the corresponding monitoring data curves of different monitoring types in a suspected abnormal section before the current sampling time are analyzed to obtain abnormal characteristic indexes of each monitoring type; for the monitoring type meeting the preset abnormal condition, the suspected abnormal section is expanded forward in an iterative manner until the abnormal characteristic index meets the preset stop condition to obtain the abnormal section of the monitoring type; the invention adjusts the early warning threshold value for monitoring and early warning according to the distribution range of the abnormal section, abnormal characteristic indexes in the expansion and the association condition among the monitoring types, and the invention enables the early warning information accuracy and timeliness of state monitoring to be higher by adaptively adjusting the early warning threshold value according to the trend abnormality and association condition among the real-time local multi-type data.
Description
Technical Field
The invention relates to the technical field of multidimensional data anomaly monitoring, in particular to an artificial intelligence-based respiratory severe patient state monitoring and early warning method.
Background
In order to help medical students learn the relevant knowledge of monitoring and early warning of patients with severe respiratory diseases through artificial intelligence assistance, such as a teacher carrying out patient state analysis teaching and the like according to different early warning information, early warning information provided by a monitoring system is generally needed, and the monitoring system mainly utilizes monitoring equipment to monitor data of physiological indexes such as respiratory rate, heart rate, blood oxygen saturation, blood pressure and the like in real time and sets an early warning threshold value to acquire early warning information.
However, in the conventional monitoring and early warning method, the selection of the early warning threshold value is usually relatively fixed, which results in relatively weak timeliness of state monitoring, the fixed early warning threshold value can only perform early warning when obvious sign data abnormality occurs, which results in a certain delay between the monitoring condition and the abnormal condition, so that the study of timely judgment of early warning information by medical students is affected, and therefore, the real-time performance of monitoring feedback abnormality is very critical to medical students teaching. When the sign data reaches the abnormal early warning, certain trend abnormal fluctuation exists, so that the early warning threshold value is set, the advanced change condition of the trend abnormality of the data fluctuation reflected by various monitoring data cannot be considered, the early warning threshold value has weaker timeliness, and the early warning information of state monitoring lacks accuracy and timeliness.
Disclosure of Invention
In order to solve the technical problems of weak early warning threshold time effectiveness and lack of accuracy and timeliness of early warning information of state monitoring in the prior art, the invention aims to provide an artificial intelligence-based state monitoring and early warning method for patients suffering from severe respiratory symptoms, and the adopted technical scheme is as follows:
The invention provides an artificial intelligence-based method for monitoring and early warning of the state of a severe respiratory patient, which comprises the following steps:
acquiring a monitoring data curve of each monitoring type at each historical sampling time before the current sampling time;
The method comprises the steps that a preset analysis number of sampling moments before the current sampling moment are used as suspected abnormal intervals; under each monitoring type, according to the continuous change degree, the fluctuation amplitude increase degree and the fluctuation mutation degree of the monitoring data curve in the suspected abnormal interval, obtaining an abnormal characteristic index of each monitoring type in the suspected abnormal interval;
For the suspected abnormal section of each monitoring type, when the abnormal characteristic index meets the preset abnormal condition, the suspected abnormal section is expanded forward in a time sequence, and the abnormal characteristic index after each suspected abnormal section expansion is calculated; obtaining an abnormal interval of each monitoring type until the overall change condition of the abnormal characteristic index meets a preset stop condition;
Obtaining the abnormal adjustment degree of each monitoring type according to the distribution range of the abnormal section of each monitoring type, the distribution condition of the abnormal characteristic index of the corresponding abnormal section in the expansion process and the association condition between the abnormal section of the corresponding monitoring type and the monitoring data curve of other monitoring types;
acquiring an early warning threshold value of each monitoring type according to the abnormal adjustment degree of each monitoring type; and carrying out monitoring and early warning based on the early warning threshold value of each monitoring type.
Further, the method for acquiring the abnormal characteristic index comprises the following steps:
Taking each monitoring type as an analysis type in sequence, and acquiring a slope corresponding to each sampling moment on a monitoring data curve corresponding to the analysis type in a suspected abnormal interval; calculating slope differences between every two adjacent sampling moments in the suspected abnormal interval, and carrying out negative correlation mapping and normalization on accumulated values of all slope differences to obtain a continuous change index of the analysis type in the suspected abnormal interval;
Obtaining peaks of analysis types on the corresponding monitoring data curves in the suspected abnormal intervals, and obtaining corresponding amplitudes of each peak; calculating the difference value between the corresponding amplitude of the next wave crest and the corresponding amplitude of the previous wave crest in every two adjacent wave crests in the suspected abnormal interval to obtain an amplitude difference value; obtaining an amplitude change index of the analysis type in the suspected abnormal interval according to the change degree of all amplitude difference values in the suspected abnormal interval;
For any wave crest of the analysis type in the suspected abnormal interval, calculating the time interval between two minimum values adjacent to the wave crest to obtain the fluctuation span of the wave crest; taking the ratio of the amplitude of the wave crest to the fluctuation span as the aspect ratio of the wave crest; taking the product of the amplitude and the aspect ratio of the wave crest as the wave crest waveform value; normalizing the sum of the waveform values of all wave peaks of the analysis type in the suspected abnormal interval to obtain a fluctuation mutation index of the analysis type in the suspected abnormal interval;
Obtaining an abnormal characteristic index of the analysis type in the suspected abnormal section according to the continuous change index, the amplitude change index and the fluctuation mutation index of the analysis type in the suspected abnormal section; the continuous change index, the amplitude change index and the fluctuation mutation index are positively correlated with the abnormal characteristic index.
Further, the obtaining the amplitude variation index of the analysis type in the suspected abnormal section according to the variation degrees of all the amplitude differences in the suspected abnormal section includes:
mapping all the amplitude difference values of the analysis type in the suspected abnormal interval into a time sequence space, and performing curve fitting to obtain an amplitude difference value curve in the suspected abnormal interval; acquiring the derivative of each amplitude difference in an amplitude difference curve;
And calculating the derivative difference value between the next amplitude difference value and the previous amplitude difference value in every two adjacent amplitude difference values, and carrying out normalization processing on the accumulated value of all the derivative difference values to obtain the amplitude change index of the analysis type in the suspected abnormal interval.
Further, for each monitoring type of suspected abnormal section, when the abnormal characteristic index meets a preset abnormal condition, the suspected abnormal section is iteratively expanded forward according to a time sequence order, and the abnormal characteristic index after each suspected abnormal section expansion is calculated, including:
The preset abnormal condition is that the abnormal characteristic index of the monitoring type in the suspected abnormal interval is larger than a preset abnormal threshold value;
For any monitoring type, when the abnormal characteristic index of the monitoring type in the suspected abnormal section meets a preset abnormal condition, the suspected abnormal section of the monitoring type is expanded forwards by one sampling time to obtain the expanded suspected abnormal section of the monitoring type, and the abnormal characteristic index of the expanded suspected abnormal section is calculated.
Further, the obtaining the abnormal section of each monitoring type until the overall change condition of the abnormal characteristic index meets a preset stop condition includes:
for any monitoring type, in the process of expanding the suspected abnormal section of the monitoring type, obtaining the stopping possibility of each expansion according to the change condition of the abnormal characteristic index of the suspected abnormal section corresponding to each expansion and the abnormal characteristic index of the suspected abnormal section corresponding to the previous expansion;
The preset stopping conditions are as follows: the expanded abnormal characteristic index does not meet the preset abnormal condition, or the stopping possibility of the corresponding expansion is larger than a preset stopping threshold value when the expanded abnormal characteristic index meets the preset abnormal condition;
And stopping the expansion process of the monitoring type when the expanded suspected abnormal section of the monitoring type meets a preset stop condition, and taking the suspected abnormal section after the last expansion as the abnormal section of the monitoring type.
Further, the method for acquiring the stopping possibility includes:
For any one expansion, when the abnormal characteristic index corresponding to the expanded is greater than or equal to the abnormal characteristic index corresponding to the expanded, marking the expanded variation value of the expansion as a preset variation value;
When the abnormal characteristic index corresponding to the expanded is smaller than the abnormal characteristic index corresponding to the expanded, taking the difference between the abnormal characteristic index corresponding to the expanded and the abnormal characteristic index corresponding to the expanded as an expanded change value of the expansion;
The extension change values of each previous extension including the extension are added to obtain the stop possibility of the extension.
Further, the method for acquiring the abnormal adjustment degree comprises the following steps:
taking each monitoring type as a type to be detected in sequence, and recording the abnormal adjustment degree of the type to be detected as zero when the number of sampling moments in an abnormal interval of the type to be detected is equal to the preset analysis number;
When the number of sampling moments in the abnormal interval of the type to be detected is larger than the preset analysis number, taking the total number of the sampling moments in the abnormal interval of the type to be detected as the interval span of the type to be detected; calculating the difference value between the maximum value and the minimum value of the data on the monitored data curve in the abnormal interval to be used as the interval amplitude of the type to be detected; taking the product of the interval span of the type to be measured and the interval amplitude as an interval distribution index of the type to be measured;
Taking the maximum value of all abnormal characteristic indexes calculated in the expansion process of the abnormal interval corresponding to the type to be detected as an abnormal degree index of the type to be detected;
In an abnormal interval of the type to be detected, calculating a Pearson coefficient between the type to be detected and each monitoring data curve of other monitoring types, and solving an absolute value to obtain the association degree of the type to be detected and each other monitoring type; taking the accumulated sum of the association degrees of the type to be detected and all monitoring types as an association index of the type to be detected;
and carrying out cumulative and normalization processing on the distribution index, the abnormality degree index and the association index of the type to be detected to obtain the abnormality adjustment degree of the type to be detected.
Further, the obtaining the early warning threshold value of each monitoring type according to the abnormal adjustment degree of each monitoring type includes:
For any monitoring type, calculating the product of the preset maximum adjustment amount and the abnormal adjustment degree of the monitoring type to obtain the threshold adjustment amount of the monitoring type;
taking the difference value between the preset high threshold value of the monitoring type and the threshold value adjustment amount as the high threshold value of the early warning threshold value of the monitoring type; and taking the sum value of the preset low threshold value and the threshold value adjustment amount of the monitoring type as the low threshold value of the early warning threshold value of the monitoring type.
Further, the monitoring and early warning based on the early warning threshold value of each monitoring type comprises:
If the monitoring type at the current sampling moment corresponds to the data of the monitoring data curve in the abnormal interval, when the data is larger than the high threshold value of the early warning threshold value corresponding to the monitoring type or smaller than the low threshold value of the early warning threshold value corresponding to the monitoring type, the early warning information corresponding to the monitoring type is marked as abnormal, otherwise, the early warning information corresponding to the monitoring type is marked as normal.
Further, the method for acquiring the amplitude comprises the following steps:
Among the two minima adjacent to each peak, the minimum of the two minima is taken as the fluctuation minimum of each peak; and calculating the numerical difference between each wave crest and the minimum value of the fluctuation to obtain the amplitude of each wave crest.
The invention has the following beneficial effects:
According to the invention, through analyzing three aspects of continuous change, fluctuation amplitude and fluctuation mutation of corresponding monitoring data curves of different monitoring types in a suspected abnormal section before the current sampling time, abnormal characteristic indexes of each monitoring type in the suspected abnormal section are obtained, the trend abnormal fluctuation condition which can exist before the real state abnormality is considered, the preliminary abnormal condition of each monitoring type is obtained to carry out subsequent threshold adjustment, meanwhile, for the condition that the preliminary abnormality exists in the more complete analysis, the suspected abnormal section is expanded forward in an iterative manner, the abnormal section of each monitoring type is obtained based on the abnormal characteristic indexes in the expansion process to meet the preset stop adjustment condition, and a certain abnormal fluctuation condition existing before the current sampling time is completely represented. Further, the condition abnormality to be pre-warned is characterized by the common change of the multidimensional data, so that when the adjustment degree is obtained by analyzing the abnormal condition in the abnormal section, the abnormal adjustment degree of each monitoring type is obtained by analyzing the distribution range of the abnormal section and the abnormal characteristic index in the expansion, and the abnormal adjustment degree of each monitoring type is obtained, so that the condition abnormality to be pre-warned can be accurately adjusted subsequently. Finally, the early warning threshold is adjusted according to the abnormal adjustment degree to carry out monitoring early warning. According to the invention, through the tendency abnormal conditions and the association conditions among the real-time local multi-type monitoring data, the early warning threshold value is adaptively adjusted, so that the state abnormality needing early warning can be rapidly and accurately monitored, and the early warning information accuracy and timeliness of the state monitoring are higher.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for monitoring and early warning of a state of a patient suffering from severe respiratory symptoms based on artificial intelligence according to an embodiment of the present invention.
Detailed Description
In order to further explain the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of specific implementation, structure, characteristics and effects of the artificial intelligence-based method for monitoring and early warning of the state of a patient suffering from severe respiratory disease according to the invention, which is provided by the invention, with reference to the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The invention provides a specific scheme of an artificial intelligence-based respiration severe patient state monitoring and early warning method, which is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a respiratory critical patient state monitoring and early warning method based on artificial intelligence according to an embodiment of the invention is shown, and the method comprises the following steps:
S1: and acquiring a monitoring data curve of each monitoring type at each historical sampling time before the current sampling time.
When the monitoring system monitors, the real-time monitoring data are collected to judge the early warning threshold value, early warning information of different monitoring data is obtained, and as different monitoring data have various changes, the early warning threshold value of each monitoring data is usually set based on abnormal conditions of the monitoring data, and as the abnormality of the monitoring data is usually not sudden abnormality, namely, before abnormal data fluctuation occurs, the monitoring data can exist in a certain abnormal fluctuation trend in a normal data range, but the monitoring early warning is not triggered at this time, the early warning information can be triggered until the abnormal increase reaches the early warning threshold value, the real-time performance of the early warning information has short delay, the different monitoring data have early warning errors, the early warning information is insufficient in time accuracy when the teaching judgment state is performed, and the teaching analysis of the monitoring early warning information is inaccurate.
Therefore, in order to improve the timeliness of monitoring and early warning, the judgment of the abnormal trend degree is carried out according to the change condition of the collected data during data collection, the threshold value is adaptively adjusted through possible abnormal conditions, the delay of abnormal early warning is reduced, and the real-time accuracy of early warning information is improved. Because abnormal state changes needing early warning are commonly presented in various monitoring data, for example, under abnormal anoxic state conditions needing early warning, the respiratory rate and the heart rate are accelerated, and simultaneously, the oxygen saturation is reduced and other linkage change conditions, the association judgment is needed by means of various monitoring data types, so that the accuracy of adjusting the early warning condition threshold is improved.
In the embodiment of the invention, monitoring equipment based on artificial intelligence, such as medical monitoring equipment connected with a respiratory sensor, an electrocardiograph, a pulse oxygen saturation meter, a sphygmomanometer and the like, acquires monitoring data of different monitoring types at each acquisition time in real time, wherein the monitoring types are respiratory rate, heart rate, blood oxygen saturation, blood pressure and the like. The collected monitoring data of different monitoring types are mapped into time sequence spaces corresponding to the monitoring types respectively, curve fitting is carried out to obtain a monitoring data curve of each monitoring type, wherein the time sequence space is a time sequence coordinate system, the horizontal axis is represented as time, the vertical axis is represented as the collected monitoring data, the collection frequency is set to be once per second, and a specific numerical value implementer can adjust according to specific real-time scenes without limitation. It should be noted that, curve fitting of data is a technical means well known to those skilled in the art, and will not be described herein.
So far, the real-time collection of various monitoring data is carried out, so that the abnormal condition of each monitoring data can be further analyzed.
S2: the method comprises the steps that a preset analysis number of sampling moments before the current sampling moment are used as suspected abnormal intervals; under each monitoring type, according to the continuous change degree, the fluctuation amplitude increase degree and the fluctuation mutation degree of the monitoring data curve in the suspected abnormal interval, obtaining the abnormal characteristic index of each monitoring type in the suspected abnormal interval.
Since the situation of trend anomaly is a continuous change process, when real-time data analysis is performed, analysis of change trend needs to be performed on the sampled data near the current sampling time, that is, the preset analysis number of sampling times before the current sampling time is taken as a suspected anomaly interval.
By means of the change conditions of the monitoring data of different monitoring types in the suspected abnormal interval, whether the local fluctuation condition of each monitoring type at the current sampling time is abnormal or not can be analyzed, and the abnormal fluctuation condition is analyzed from three aspects, namely a continuous change characteristic, an amplitude change characteristic and a fluctuation mutation characteristic.
For the continuous change feature, normal monitoring data is usually in a relatively stable state, and when the monitoring data shows continuous change, such as continuous rising of respiratory frequency, continuous falling of pulse oxygen saturation and the like, it is indicated that abnormal change features may exist in the condition that the monitoring data reflects in the section, so that the greater the overall continuous change degree in the monitoring data curve, the more remarkable the abnormal features of the monitoring type data in the suspected abnormal section.
For the amplitude variation characteristic, under the condition that the vital sign is more stable, the general fluctuation of the monitored data is smaller fluctuation with continuous similar amplitude, the state of the whole monitored data is more stable, and when the fluctuation amplitude of the monitored data is gradually increased, the condition that the vital sign data is unbalanced and disturbed, the fluctuation intensity of the vital sign data is gradually increased and tends to generate state abnormality is further described, the condition that the monitored data reflects in the section part can have abnormal characteristics, so that when the fluctuation amplitude of the monitored data curve is gradually increased, the more obvious the regular variation is, the more obvious the abnormal characteristics of the monitored data in the suspected abnormal section are.
For the fluctuation mutation feature, under the condition that the vital sign is stable, the fluctuation change of the monitoring data is gentle, the waveform of the monitoring data curve is smoother, and the abnormal condition caused by the abnormal state can cause the data of some monitoring types to have mutation fluctuation, the monitoring data curve is shown as the steep feature of the waveform, so that when the fluctuation mutation feature of the monitoring data curve is more remarkable, the abnormal feature of the monitoring types of data in the suspected abnormal section is more remarkable.
The three aspects of the abnormal characteristics can be comprehensively characterized, and the abnormal condition of each monitoring type in the suspected abnormal interval is analyzed, so that under each monitoring type, the abnormal characteristic index of each monitoring type in the suspected abnormal interval is obtained according to the continuous change degree, the fluctuation amplitude increase degree and the fluctuation mutation degree of the monitoring data curve in the suspected abnormal interval.
Preferably, each monitoring type is used as an analysis type in sequence, each monitoring type is analyzed, the slope corresponding to each sampling moment on the monitoring data curve corresponding to the analysis type in the suspected abnormal interval is obtained, and the degree of continuous change is reflected through the change between the slopes. And calculating the slope difference between every two adjacent sampling moments in the suspected abnormal interval, carrying out negative correlation mapping and normalization processing on accumulated values of all the slope differences to obtain a continuous change index of the analysis type in the suspected abnormal interval, wherein the smaller the slope difference is, the more stable the change is, the smaller the fluctuation inflection point is, the more likely the higher the continuous change is, and the greater the abnormal condition is. It should be noted that, the obtaining of the slope on the curve is a technical means well known to those skilled in the art, and will not be described herein.
Further, both the change of the fluctuation amplitude and the fluctuation mutation degree need to be analyzed for the wave, so that the peak on the corresponding monitoring data curve of the analysis type in the suspected abnormal interval is obtained.
In order to reflect the amplitude of the wave, the amplitude corresponding to each wave crest is further obtained, the amplitude reflects the difference degree of extremum in each wave crest, and in the embodiment of the invention, the minimum value in two minimum values adjacent to each wave crest is taken as the wave crest minimum value of each wave crest, and the numerical value difference between each wave crest and the wave crest minimum value is calculated to obtain the amplitude of each wave crest. In other embodiments of the present invention, the average value of the two minima may be taken as the minimum value of the fluctuation, and the difference between the peak and the minimum value of the fluctuation may be taken as the amplitude of each peak, which is not limited herein.
Because the amplitude is continuously increased under the abnormal condition, the difference value between the amplitude corresponding to the next wave crest and the amplitude corresponding to the previous wave crest in every two adjacent wave crests in the suspected abnormal interval is calculated, and the amplitude difference value is obtained, namely the amplitude difference value between each wave crest and the previous wave crest is calculated, and the increase condition of the amplitude along with the time sequence is reflected through the amplitude difference value.
Further, according to the change degree of all the amplitude differences in the suspected abnormal interval, an amplitude change index of the analysis type in the suspected abnormal interval is obtained. The derivative of each amplitude difference in the amplitude difference curve is obtained, the derivative reflects the change degree of each amplitude difference on the curve, when the derivative is positive and is continuously increased, the amplitude change is more consistent with the abnormal condition, therefore, the derivative difference between the next amplitude difference and the previous amplitude difference in every two adjacent amplitude differences is calculated, the accumulated value of all the derivative differences is normalized, the amplitude change index of the analysis type in the suspected abnormal interval is obtained, and the abnormal condition on the fluctuation amplitude change is reflected.
For any wave crest of the analysis type in the suspected abnormal interval, calculating a time interval between two minimum values adjacent to the wave crest to obtain a fluctuation span of the wave crest, wherein the fluctuation span is reflected as the fluctuation duration of each fluctuation. The ratio of the amplitude to the fluctuation span of the peak is taken as the aspect ratio of the peak, the degree of abrupt change of each peak is reflected by the aspect ratio, and the larger the aspect ratio is, the more remarkable the abrupt change is. Taking the product of the amplitude and the aspect ratio of the wave crest as the waveform value of the wave crest, integrating the abrupt change condition at each wave crest and the fluctuation amplitude, and reflecting the corresponding steepness degree of the wave crest through the waveform value. Normalizing the sum of the waveform values of all the peaks of the analysis type in the suspected abnormal interval to obtain a fluctuation mutation index of the analysis type in the suspected abnormal interval, and reflecting the abnormal condition of the fluctuation degree.
Finally, according to the continuous change index, the amplitude change index and the fluctuation mutation index of the analysis type in the suspected abnormal section, obtaining the abnormal characteristic index of the analysis type in the suspected abnormal section, synthesizing different abnormal conditions, and integrally analyzing the abnormal conditions of the suspected abnormal section. The continuous change index, the amplitude change index and the fluctuation mutation index are all positively correlated with the abnormal characteristic index, and in the embodiment of the invention, the expression of the abnormal characteristic index is as follows:
；
In the method, in the process of the invention, Expressed as/>Abnormal characteristic index of species monitoring type in suspected abnormal section,/>Expressed as the total number of sampling instants in the suspected anomaly interval,/>Expressed as the/>, in the suspected abnormal sectionSlope of each sampling instant,/>Expressed as the/>, in the suspected abnormal sectionSlope of each sampling instant,/>Expressed as the total number of amplitude differences in suspected anomaly intervals,/>Expressed as the/>, in the suspected abnormal sectionDerivative of the difference of the individual amplitudes,/>Expressed as the/>, in the suspected abnormal sectionDerivative of the difference of the individual amplitudes,/>Expressed as the total number of peaks in the suspected anomaly interval,/>Expressed as the/>, in the suspected abnormal sectionWave span of individual peaks,/>Expressed as the/>, in the suspected abnormal sectionThe amplitude of the individual peaks. /(I)Expressed as an absolute value extraction function,/>Expressed as an exponential function with a base of natural constant,/>It should be noted that, normalization is a technical means well known to those skilled in the art, and the normalization function may be selected by linear normalization or standard normalization, and the specific normalization method is not limited herein.
Wherein,Represented as the difference in slope between adjacent sampling instants in the suspected abnormal interval,Expressed as/>The continuous change index of the species monitoring type in the suspected abnormal section,Expressed as the derivative difference between every two adjacent amplitude differences in the suspected anomaly interval,Expressed as/>The amplitude change index of the species monitoring type in the suspected abnormal section,Expressed as the/>, in the suspected abnormal sectionAspect ratio of the peaks,/>Indicated as the first in suspected abnormal sectionWaveform value of each wave crest,/>Expressed as/>And monitoring the fluctuation mutation index of the type in the suspected abnormal region. When the continuous change index, the amplitude change index or the fluctuation mutation index are larger, the abnormal characteristic index is larger when the suspected abnormal section has abnormal fluctuation. In other embodiments of the present invention, other basic mathematical operations may be used to reflect that the continuous change index, the amplitude change index, and the fluctuation mutation index are all positively correlated with the abnormal characteristic index, such as addition or exponentiation, without limitation.
Thus, analysis of trend abnormal conditions of monitoring data of each monitoring type in real time and locally is completed.
S3: for the suspected abnormal section of each monitoring type, when the abnormal characteristic index meets the preset abnormal condition, the suspected abnormal section is expanded forward in a time sequence, and the abnormal characteristic index after each suspected abnormal section expansion is calculated; and obtaining the abnormal interval of each monitoring type until the overall change condition of the abnormal characteristic index meets the preset stop condition.
Further, the monitoring type data with abnormal conditions continues to search for complete abnormal intervals which are possibly abnormal conditions, so that the early warning threshold value can be conveniently adjusted according to the continuous characteristics of the complete abnormal conditions, therefore, when the abnormal characteristic index of each monitoring type of suspected abnormal interval meets the preset abnormal conditions, the suspected abnormal interval is expanded forward in a time sequence and the abnormal characteristic index after each suspected abnormal interval expansion is calculated.
Preferably, the preset abnormal condition is that an abnormal characteristic index of the monitoring type in the suspected abnormal section is greater than a preset abnormal threshold, and in the embodiment of the present invention, the preset abnormal threshold is 0.3, and a specific numerical value implementer can adjust according to specific implementation conditions, that is, when the abnormal characteristic index of the monitoring type in the suspected abnormal section is greater than 0.3, it is indicated that the abnormal characteristic index meets the preset abnormal condition, and the abnormal section of the monitoring type needs to be expanded.
For any monitoring type, when the abnormal characteristic index of the monitoring type in the suspected abnormal section meets the preset abnormal condition, the abnormal condition of the suspected abnormal section is indicated, and more complete abnormal condition judgment needs to be searched forward. The suspected abnormal section of the monitoring type is expanded forwards by one sampling time, the suspected abnormal section of the monitoring type after expansion is obtained, and an abnormal characteristic index of the suspected abnormal section after expansion is calculated and used for subsequent analysis of the change condition of abnormal conditions.
When the preset stop condition is not met, the suspected abnormal section of the monitoring type can be expanded forwards in an iterative mode, an abnormal characteristic index is calculated after each expansion, in the continuous expansion process, the abnormal condition is continuously reduced, two stop conditions exist in the process, one abnormal characteristic index is reduced to be normal, namely not to be in full of the preset abnormal condition, the other abnormal condition is reduced to be larger in change degree, the abnormal condition reaches a demarcation point which can be divided, and abnormal judgment is not needed any more, so that the abnormal section of each monitoring type is obtained until the integral change condition of the abnormal characteristic index meets the preset stop condition.
Preferably, for any monitoring type, in the process of expanding the suspected abnormal section of the monitoring type, according to the change condition of the abnormal characteristic index of the suspected abnormal section corresponding to each expansion and the abnormal characteristic index of the suspected abnormal section corresponding to the previous expansion, the stopping possibility of each expansion is obtained, and the stopping possibility is the possibility that the obtained expansion is a stopping point according to the analysis of the abnormal change condition before and after the expansion at the moment. In the embodiment of the invention, for any expansion, the possibility of stopping each time is calculated, when the abnormal characteristic index corresponding to the expanded is greater than or equal to the abnormal characteristic index corresponding to the expanded, the trend that the abnormal condition is still not reduced after the forward expansion is indicated, and the probability of stopping the expansion is extremely low, so that the expansion change value of the expansion is recorded as a preset change value. When the abnormal characteristic index corresponding to the after expansion is smaller than the abnormal characteristic index corresponding to the before expansion, the abnormal condition is described to be smaller, the change condition of the abnormal condition is required to be recorded, the difference between the abnormal characteristic index corresponding to the before expansion and the abnormal characteristic index corresponding to the after expansion is used as the expansion change value of the expansion, the abnormal change condition before and after expansion is recorded, the expansion change value of each expansion before the expansion is added, the stopping possibility of the expansion is obtained, and the stopping possibility of the expansion is obtained through the integral change condition before the expansion, and in the embodiment of the invention, the expression of the stopping possibility is as follows:
；
In the method, in the process of the invention, Expressed as/>Species monitoring type is at/>Extended change value of secondary extension,/>Expressed as/>Species monitoring type is at/>Abnormal characteristic index of suspected abnormal section after secondary expansion,/>Expressed as/>Species monitoring type is at/>Abnormal characteristic index of suspected abnormal section after secondary expansion,/>Expressed as/>Extended times of species monitoring type,/>Expressed as/>Species monitoring type is at/>The stopping possibility of the secondary expansion. Wherein, the/>The species monitoring type is in the firstThe abnormal characteristic index of the suspected abnormal section after secondary expansion is the/>The abnormal characteristic index before secondary expansion has an abnormal characteristic value obtained by a suspected abnormal section with an initial size before 1 st expansion.
The preset stop conditions for stopping the expansion process are therefore: the expanded abnormal characteristic index does not meet the preset abnormal condition, or the possibility of stopping the corresponding expansion when the expanded abnormal characteristic index meets the preset abnormal condition is larger than a preset stopping threshold, namely, when the abnormal characteristic index is in a normal range, or the degree of variation of the abnormal characteristic index is reduced enough when the abnormal characteristic index does not reach the normal range yet, the expansion process is stopped, and the section corresponding to the expansion at the moment is the section meeting the abnormal condition. It should be noted that, when the initial suspected abnormal section of the monitoring type is normal, that is, when the extended preset stop condition is directly satisfied, the initial suspected abnormal section of the monitoring type is an abnormal section.
And stopping the expansion process of the monitoring type when the expanded suspected abnormal section of the monitoring type meets a preset stop condition, and taking the suspected abnormal section after the last expansion as the abnormal section of the monitoring type. And analyzing the abnormal conditions of each monitoring type through the abnormal interval conditions corresponding to the monitoring types, so as to obtain real-time adjustment conditions.
S4: and obtaining the abnormal adjustment degree of each monitoring type according to the distribution range of the abnormal section of each monitoring type, the distribution condition of the abnormal characteristic index of the corresponding abnormal section in the expansion process and the association condition between the abnormal section of the corresponding monitoring type and the monitoring data curve of other monitoring types.
The method comprises the steps of integrating the performance degrees of abnormal intervals of different monitoring types, judging approaching abnormal conditions reflected by different monitoring types, and if the abnormal fluctuation conditions of the abnormal intervals are more remarkable and the range is larger, indicating that the fluctuation conditions of corresponding vital sign data tend to be abnormal, the duration time of the abnormal conditions is longer, and the early warning threshold value which needs to be adjusted for abnormality is larger as the duration time of the abnormal conditions is longer and is closer to the abnormal conditions of real states. Meanwhile, because the abnormal state needing early warning is characterized by common change of data of multiple monitoring types, for example, under the abnormal condition of the anoxic state needing early warning, the respiratory rate is increased, the heart rate is increased, the pulse oxygen saturation is reduced, the temperature is increased or the blood pressure is changed, and the like, when the association degree of the common change of the multiple monitoring types is tighter, the more reliable the abnormal state needing early warning at the moment can be illustrated, and the finally obtained abnormal adjustment degree is also more reliable. Therefore, according to the distribution range of the abnormal section of each monitoring type and the distribution condition of the abnormal characteristic index of the corresponding abnormal section in the expansion process and the association condition between the abnormal section of the corresponding monitoring type and the monitoring data curves of other monitoring types, the abnormal adjustment degree of each monitoring type is obtained.
Preferably, each monitoring type is sequentially used as a type to be detected, when the number of sampling moments in an abnormal section of the type to be detected is equal to the preset analysis number, the abnormal adjustment degree of the type to be detected is recorded as zero, when the number of sampling moments is equal to the preset analysis number, the abnormal section is an initial suspected abnormal section which is not expanded, and further the abnormal degree in the section is not satisfied with the abnormal expansion condition at the initial time, at this time, the abnormal section is in normal data fluctuation, and the sensitivity adjustment of a threshold value is not needed, so the abnormal adjustment degree is recorded as zero.
When the number of sampling moments in the abnormal section of the type to be detected is larger than the preset analysis number, the expansion of the abnormal condition in the abnormal section is indicated, and the threshold value needs to be adjusted, so that the abnormal characterization condition is further analyzed to acquire the abnormal adjustment degree. It should be noted that, since the abnormal section is obtained by expanding a suspected abnormal section composed of a preset analysis number of sampling times, the number of sampling times in the abnormal section may be only equal to or greater than the preset analysis number, and no less than the preset analysis number may be possible.
Firstly, analyzing the distribution range of an abnormal section, taking the total number of middle sampling moments of the abnormal section of the type to be tested as the section span of the type to be tested, reflecting the duration of the abnormality from the duration range of the abnormal section in time, wherein the larger the section span is, the more likely the state abnormality is approached. And calculating the difference value between the data maximum value and the data minimum value on the monitoring data curve in the abnormal section, and taking the difference value as the section amplitude of the type to be detected, wherein the maximum fluctuation range is reflected from the numerical value floating condition of the abnormal section, and the larger the section amplitude is, the closer the abnormal fluctuation is to the state abnormal condition. Taking the product of the interval span of the type to be measured and the interval amplitude as an interval distribution index of the type to be measured, and analyzing the distribution range of the abnormal interval to obtain the degree to which the early warning threshold needs to be adjusted.
Further analyzing the degree of the abnormal condition, taking the maximum value of all abnormal characteristic indexes calculated in the expansion process of the abnormal section corresponding to the type to be detected as the abnormal degree index of the type to be detected, reflecting the degree of the most abnormal condition of the abnormal section, and when the degree of the most abnormal condition is larger, indicating that the possibility of the abnormal degree of the state occurring under the monitoring type is larger, the threshold value of the abnormal degree index needs to be more sensitive, namely the degree of the early warning threshold value needs to be adjusted is larger.
Further analyzing the association change between the monitoring types, calculating the Pearson coefficient between the monitoring data curves of the type to be detected and each other monitoring type in the abnormal interval of the type to be detected, namely in the time sequence range of the abnormal interval of the type to be detected, and solving the absolute value to obtain the association degree between the type to be detected and each other monitoring type, wherein the association between the data is reflected through the Pearson coefficient. And taking the accumulated sum of the association degrees of the type to be detected and all the monitoring types as an association index of the type to be detected, and integrating the association degrees among all the monitoring types, wherein the larger the overall association degree is, the more likely the approaching abnormal fluctuation before the state abnormal condition is. It should be noted that, the calculation of the pearson coefficient is a technical means well known to those skilled in the art, and will not be described herein.
Finally, carrying out cumulative and normalization processing on the distribution index, the abnormality degree index and the association index of the type to be tested to obtain the abnormality adjustment degree of the type to be tested, and integrating the abnormality probability of the type to be tested to reflect the degree of the threshold value of the type to be tested to be adjusted, wherein in the embodiment of the invention, the expression of the abnormality adjustment degree is as follows:
；
In the method, in the process of the invention, Expressed as/>Abnormal adjustment degree of species monitoring type,/>Expressed as/>Interval span of species monitoring type,/>Expressed as/>Interval amplitude of species monitoring type,/>Expressed as/>Abnormality degree index of species monitoring type,/>Expressed as except for the first ]Total number of other monitoring types than the species monitoring type,/>Expressed as/>Species monitoring type and other/>Correlation of species monitoring type,/>Expressed as an absolute value extraction function,/>Represented as a normalization function.
Wherein,Expressed as/>Interval distribution index of species monitoring type,/>Expressed as/>And (5) a monitoring type of associated index. When the interval distribution index is larger, the abnormality degree index is larger, the association index is larger, which indicates that the abnormality in the abnormal interval corresponding to the monitoring type has long duration and obvious abnormality, and the association with common change of other monitoring types is strong, which indicates that the approaching abnormality is more likely to be a state abnormality, the range of an early warning threshold is required to be reduced in time to a larger extent, the early warning sensitivity is improved, and the abnormality adjustment degree is larger.
So far, the degree of self-adaptive adjustment threshold value required by each monitoring type, namely the abnormal adjustment degree, is obtained through the analysis of the abnormal condition of each monitoring type in the abnormal section and the linkage condition of each monitoring type with other monitoring types.
S5: acquiring an early warning threshold value of each monitoring type according to the abnormal adjustment degree of each monitoring type; and carrying out monitoring and early warning based on the early warning threshold value of each monitoring type.
According to the abnormal adjustment degree of each monitoring type, the self-adaptive range adjustment is carried out on the threshold value of each monitoring type, and the more obvious monitoring types are abnormal, the greater the adjustment degree is, so that the early warning threshold value of each monitoring type is obtained according to the abnormal adjustment degree of each monitoring type. Preferably, for any monitoring type, the product of the preset maximum adjustment amount and the abnormal adjustment degree of the monitoring type is calculated to obtain the threshold adjustment amount of the monitoring type, wherein the threshold adjustment amount is the value to be adjusted according to the abnormal condition. The difference value between the preset high threshold value and the threshold value adjustment amount of the monitoring type is used as the high threshold value of the early warning threshold value of the monitoring type, the sum value of the preset low threshold value and the threshold value adjustment amount of the monitoring type is used as the low threshold value of the early warning threshold value of the monitoring type, the high threshold value of the early warning threshold value is reduced, the low threshold value is improved, the early warning sensitivity is improved, and the early warning threshold value which is more sensitive to the trend of abnormal fluctuation with abnormal state is given.
Finally, the real-time monitoring is carried out through the early warning threshold value after the adjustment of each monitoring type, so that the early warning can be timely carried out according to trend fluctuation of the abnormal situation before the abnormal situation is reached during early warning, the early warning timeliness is improved, namely the early warning is carried out based on the early warning threshold value of each monitoring type. In one embodiment of the present invention, if the monitoring type at the current sampling time corresponds to the data of the monitoring data curve in the abnormal section, if the data is greater than the high threshold value of the early warning threshold value corresponding to the monitoring type or less than the low threshold value of the early warning threshold value corresponding to the monitoring type, the early warning information corresponding to the monitoring type is marked as abnormal, otherwise the early warning information corresponding to the monitoring type is marked as normal. Through the real-time early warning threshold adjustment of more sensitivity, the timeliness of the early warning information is improved when the accuracy is guaranteed, through the more accurate early warning information, follow-up teachers can carry out state judgment teaching and the like on medical students according to different early warning information conditions, and also can input all monitoring type early warning information into a trained neural network model, output state results to assist teaching and displaying and the like, and further description is omitted here.
In summary, the invention analyzes three aspects of continuous change, fluctuation amplitude and fluctuation mutation of corresponding monitoring data curves of different monitoring types in a suspected abnormal section before the current sampling time to obtain abnormal characteristic indexes of each monitoring type in the suspected abnormal section, considers the tendency abnormal fluctuation condition which can exist before the real state abnormality to obtain the preliminary abnormal condition of each monitoring type, adjusts the subsequent threshold value, simultaneously, for more complete analysis of the condition of the preliminary abnormality, iteratively expands the suspected abnormal section forwards for the monitoring type meeting the preset abnormal condition, and obtains the abnormal section of each monitoring type based on the abnormal characteristic indexes in the expansion process meeting the preset stop adjustment condition, thereby completely representing the certain abnormal fluctuation condition existing before the current sampling time. Further, the condition abnormality to be pre-warned is characterized by the common change of the multidimensional data, so that when the adjustment degree is obtained by analyzing the abnormal condition in the abnormal section, the abnormal adjustment degree of each monitoring type is obtained by analyzing the distribution range of the abnormal section and the abnormal characteristic index in the expansion, and the abnormal adjustment degree of each monitoring type is obtained, so that the condition abnormality to be pre-warned can be accurately adjusted subsequently. Finally, the early warning threshold is adjusted according to the abnormal adjustment degree to carry out monitoring early warning. According to the invention, through the tendency abnormal conditions and the association conditions among the real-time local multi-type monitoring data, the early warning threshold value is adaptively adjusted, so that the state abnormality needing early warning can be rapidly and accurately monitored, and the early warning information accuracy and timeliness of the state monitoring are higher.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410354888.2A CN117954116B (en) | 2024-03-27 | 2024-03-27 | Breathing severe patient state monitoring and early warning method based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410354888.2A CN117954116B (en) | 2024-03-27 | 2024-03-27 | Breathing severe patient state monitoring and early warning method based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117954116A CN117954116A (en) | 2024-04-30 |
| CN117954116B true CN117954116B (en) | 2024-05-24 |
Family
ID=90805223
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410354888.2A Active CN117954116B (en) | 2024-03-27 | 2024-03-27 | Breathing severe patient state monitoring and early warning method based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117954116B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118571490B (en) * | 2024-07-31 | 2024-09-27 | 绿色医疗科技(大连)有限公司 | Pregnant woman pregnancy health state monitoring method based on big data analysis |
| CN118571491A (en) * | 2024-08-01 | 2024-08-30 | 江苏盖睿健康科技有限公司 | Health monitoring and early warning system based on wearable equipment |
| CN119207802B (en) * | 2024-11-27 | 2025-02-07 | 绿色医疗科技(大连)有限公司 | A method and system for monitoring patient blood pressure for clinical nursing |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110368019A (en) * | 2019-07-03 | 2019-10-25 | 北京必安必恒科技发展有限公司 | A kind of cardiechema signals feature extraction, detection model building and detection device |
| CN116328122A (en) * | 2023-03-22 | 2023-06-27 | 深圳市科曼医疗设备有限公司 | Trigger threshold adjusting method and device, equipment and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190015614A1 (en) * | 2017-07-13 | 2019-01-17 | Royal Commission Yanbu Colleges & Institutes | Systems, devices, and methodologies to provide protective and personalized ventilation |
-
2024
- 2024-03-27 CN CN202410354888.2A patent/CN117954116B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110368019A (en) * | 2019-07-03 | 2019-10-25 | 北京必安必恒科技发展有限公司 | A kind of cardiechema signals feature extraction, detection model building and detection device |
| CN116328122A (en) * | 2023-03-22 | 2023-06-27 | 深圳市科曼医疗设备有限公司 | Trigger threshold adjusting method and device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117954116A (en) | 2024-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117954116B (en) | Breathing severe patient state monitoring and early warning method based on artificial intelligence | |
| Rong et al. | A multi-type features fusion neural network for blood pressure prediction based on photoplethysmography | |
| US9042973B2 (en) | Apparatus and method for measuring physiological signal quality | |
| WO2019161609A1 (en) | Method for analyzing multi-parameter monitoring data and multi-parameter monitor | |
| CN109276241B (en) | Pressure identification method and equipment | |
| CN106037720B (en) | A medical application system of mixed continuous information analysis technology | |
| US20100057490A1 (en) | Methods, systems, and computer program products for evaluating a patient in a pediatric intensive care unit | |
| WO2019153578A1 (en) | Electrocardiosignal-based non-invasive blood glucose detection method and system | |
| CN113317794B (en) | Vital sign analysis method and system | |
| US20210345896A1 (en) | Heart Rate Variability Monitoring and Analysis | |
| CN102688027A (en) | Ambulatory blood pressure monitor | |
| Choudhary et al. | Analyzing seismocardiographic approach for heart rate variability measurement | |
| Liu et al. | Refined generalized multiscale entropy analysis for physiological signals | |
| CN118053531B (en) | Intelligent management method and system for clinical data of medical examination | |
| JP5544365B2 (en) | Improvements in multi-parameter monitoring or improvements related to multi-parameter monitoring | |
| KR20220123376A (en) | Methods and systems for determining cardiovascular parameters | |
| Roy et al. | BePCon: A photoplethysmography-based quality-aware continuous beat-to-beat blood pressure measurement technique using deep learning | |
| Zaman et al. | Estimating reliability of signal quality of physiological data from data statistics itself for real-time wearables | |
| CN118430815A (en) | Remote monitoring method and system for patient data for medical care | |
| Ma et al. | PPG-based continuous BP waveform estimation using polarized attention-guided conditional adversarial learning model | |
| WO2019153579A1 (en) | Electrocardiographic signal-based airbag-free blood pressure measurement method and system | |
| Liu et al. | The impact of noise on the reliability of heart-rate variability and complexity analysis in trauma patients | |
| Xie et al. | Prediction of chronic obstructive pulmonary disease exacerbation using physiological time series patterns | |
| CN117045216B (en) | A non-invasive medical analysis system for blood indicators based on fuzzy control | |
| CN119074038B (en) | Cardiopulmonary sound data monitoring method and system based on big data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |