CN115910320A - An acute respiratory distress syndrome early warning system for critically ill ICU patients - Google Patents

Abstract

The invention belongs to the technical field of acute respiratory distress syndrome early warning, and discloses an acute respiratory distress syndrome early warning system for ICU severe patients, which comprises: the device comprises a physiological index detection module, a respiratory rate detection module, a central control module, an analysis module, a prediction module, a health assessment module, an early warning module and a display module. According to the invention, the respiratory frequency of the detected ICU critical patient is determined by the respiratory frequency detection module according to the change curve of the area of the detection area along with the change of time, the detection process does not need manual participation, and the respiratory frequency can be more conveniently detected; meanwhile, machine learning is applied to the death rate prediction of the ARDS patient through the prediction module, the death rate of the ARDS patient can be accurately and objectively predicted through a model trained through the machine learning, and more effective and feasible prediction information is provided for clinicians to serve as reference.

Description

An acute respiratory distress syndrome early warning system for critically ill patients in ICU

Technical Field

The invention belongs to the technical field of acute respiratory distress syndrome early warning, and in particular relates to an acute respiratory distress syndrome early warning system for critically ill patients in an ICU.

Background Art

Acute respiratory distress syndrome (ARDS) is a clinical syndrome caused by intrapulmonary and/or extrapulmonary causes, with refractory hypoxemia as a prominent feature. It has attracted much attention due to its high mortality rate. There are many causes of ARDS, and the pathogenesis of ARDS caused by different causes is also different. Clinical manifestations are mostly acute onset, respiratory distress, and hypoxemia that is difficult to correct with conventional oxygen therapy. The "Berlin definition" is often used internationally to diagnose and stratify the severity of ARDS, and it needs to be differentially diagnosed with a variety of diseases. Clinical examinations involve: diagnosis and differential diagnosis, treatment monitoring and guidance, criticality and prognosis evaluation, etc. The treatment of ARDS includes two major categories: mechanical ventilation therapy and non-mechanical ventilation therapy, and its effective treatment methods are still being explored. However, the existing ARDS early warning system for critically ill ICU patients requires manual observation of the chest or abdominal rise and fall of the critically ill ICU patients when detecting the respiratory rate, which is very inconvenient. At the same time, it cannot accurately predict the mortality rate of ARDS patients.

Through the above analysis, the problems and defects of the prior art are as follows:

(1) When detecting the respiratory rate, the existing early warning system for acute respiratory distress syndrome used in ICU patients requires manual observation of the chest or abdominal rise and fall of the ICU patient, which is very inconvenient.

(2) The mortality rate of ARDS patients cannot be accurately predicted.

Summary of the invention

In view of the problems existing in the prior art, the present invention provides an acute respiratory distress syndrome early warning system for critically ill patients in an ICU.

The present invention is implemented in this way: an acute respiratory distress syndrome early warning system for critically ill patients in an ICU comprises:

Physiological index detection module, respiratory rate detection module, central control module, analysis module, prediction module, health assessment module, early warning module, display module;

A physiological index detection module, connected to the central control module, is used to detect the patient's physiological index through medical equipment;

The physiological index detection module includes a data normalization processing unit, which is used to receive the pre-processing operation of detecting the physiological index of the patient from the medical device and perform normalization processing on the data;

A respiratory rate detection module, connected to the central control module, is used to detect the patient's respiratory rate;

The central control module is connected with the physiological index detection module, the respiratory rate detection module, the analysis module, the prediction module, the health assessment module, the early warning module, and the display module to control the normal operation of each module;

The prediction module is connected to the central control module and is used to predict the mortality rate of acute respiratory distress syndrome in critically ill patients in the ICU;

The analysis module is connected to the central control module, and based on the results of the respiratory rate detection module and the prediction module, it comprehensively analyzes the acute respiratory distress syndrome of critically ill patients in the ICU by analyzing the respiratory rate of the patient and the predicted value of the acute respiratory distress syndrome mortality rate;

A health assessment module, connected to the central control module, is used to assess the patient's health;

The early warning module is connected to the central control module and is used to warn of acute respiratory distress syndrome for critically ill patients in the ICU;

The early warning module includes an early warning calculation unit, which compares the patient's respiratory rate value detected by the respiratory rate detection module, the predicted value of acute respiratory distress syndrome mortality, and the results of the analysis module with the given health threshold, and if the health threshold is exceeded, an alarm message is issued; otherwise, the health information is fed back to the central control module and displayed in the display module;

The display module is connected to the central control module and is used to display the patient's physiological index, respiratory rate, analysis results, prediction results, and health assessment results through the display.

Further, the respiratory rate detection module detection method is as follows:

(1) acquiring a plurality of ICU critically ill patient images including the ICU critically ill patient to be detected; performing enhancement processing on the ICU critically ill patient images; identifying the outline of the ICU critically ill patient to be detected from each ICU critically ill patient image; and determining a detection area from each of the outlines;

Determining a detection area from each of the contours comprises:

For each ICU critically ill patient image, identifying the head region of the detected ICU critically ill patient from the current ICU critically ill patient image;

Determine the lowest point of the head area, and the head distance between the highest point of the head area and the lowest point in the vertical direction;

Determine a first straight line parallel to the horizontal direction passing through the lowest point; determine a second straight line below the first straight line and having a distance from the first straight line that is a multiple of a first preset value of the head distance;

Determine a third straight line below the second straight line, the distance from the second straight line being a multiple of a second preset value of the head distance;

determining a closed area between the second straight line and the third straight line formed by the second straight line, the third straight line and the contour;

Using the closed area as the detection area corresponding to the current ICU critically ill patient image;

(2) Determine the area of each detection area; determine a change curve of the area over time; and determine the respiratory rate of the critically ill ICU patient being detected based on the change curve.

Further, the step of identifying the head region of the detected ICU critically ill patient from the current ICU critically ill patient image includes:

Determine the R, G, and B of each pixel in the current ICU critically ill patient image;

According to the conversion formula and the R, G, B of each pixel, the parameter Cb and the parameter Cr corresponding to each pixel are determined, wherein the conversion formula is:

According to Formula 1, the intermediate parameters n ₁ and n ₂ of each pixel are determined, where Formula 1 is:

According to Formula 2, the judgment parameter P of each pixel is determined, where Formula 2 is:

Determine whether the judgment parameter of each pixel point is less than or equal to 1, and if so, determine that the pixel point belongs to the head area corresponding to the current ICU critically ill patient image; otherwise, determine that the pixel point does not belong to the head area corresponding to the current ICU critically ill patient image;

The head region corresponding to the current ICU critically ill patient image is determined according to the pixel points belonging to the head region.

Furthermore, the detection method further comprises: establishing a rectangular coordinate system;

Loading each of the ICU critically ill patients' images into the rectangular coordinate system;

The determining of the lowest point of the head region and the head distance between the highest point of the head region and the lowest point in the vertical direction comprises:

Taking the point with the smallest vertical axis coordinate in the head region as the lowest point;

determining a maximum longitudinal coordinate in the head region;

The difference between the maximum longitudinal coordinate and the longitudinal coordinate of the lowest point is taken as the head distance;

The determining of a first straight line parallel to the horizontal direction passing through the lowest point comprises:

A straight line passing through the lowest point and parallel to the horizontal axis of the rectangular coordinate system is taken as the first straight line.

Further, the prediction module prediction method is as follows:

1) Building a medical database to obtain medical data of critically ill patients in the ICU; storing the obtained medical data in the medical database; and preprocessing the obtained medical data, wherein the preprocessing includes sample screening and feature extraction;

2) Classifying the preprocessed data according to the number of days the patients survived to obtain positive and negative groups of three mortality prediction models, respectively. The three mortality prediction models include an in-hospital mortality prediction model, a 30-day mortality prediction model, and a one-year mortality prediction model;

3) Inter-group analysis was performed on the positive and negative groups of the three mortality prediction models to screen out the features with significant inter-group differences; a random forest algorithm was used to establish an acute respiratory distress syndrome mortality prediction model based on the features with significant inter-group differences; the acute respiratory distress syndrome mortality prediction model was used to predict the objects to be predicted, and the prediction results of acute respiratory distress syndrome mortality were obtained.

Further, the obtaining of medical data of critically ill patients in the ICU includes downloading the medical data of critically ill patients in the ICU from a MIMIC-III database.

Further, the pretreatment method comprises:

The obtained medical data were sampled and screened according to the inclusion criteria and exclusion criteria to obtain the screened samples. The inclusion criteria included patients aged 18 years or older who were admitted to the intensive care unit and diagnosed with acute respiratory distress syndrome according to the Berlin standard. The exclusion criteria included any one of the following: data with incomplete data records in the MIMIC-III database, patients aged less than 18 years old, patients receiving palliative therapy, and patients with an ICU record time of less than 48 hours;

For the screened samples, the variable features of each sample for modeling are extracted.

Furthermore, the preprocessing method further comprises:

Multiple imputation of missing data in acquired medical data.

Furthermore, the positive group and negative group of the three mortality prediction models are specifically:

The positive group of the in-hospital mortality prediction model is the data of patients who died in hospital, and the negative group of the in-hospital mortality prediction model is the data of patients who survived during hospitalization; the positive group of the 30-day mortality prediction model is the data of patients who died within 30 days after hospitalization, and the negative group of the 30-day mortality prediction model is the data of patients who survived within 30 days after hospitalization; the positive group of the one-year mortality prediction model is the data of patients who died within one year after hospitalization, and the negative group of the one-year mortality prediction model is the data of patients who survived within one year after hospitalization.

Furthermore, the model building method includes:

The K-fold cross-validation method was used to divide the features with significant differences between groups into the first training set and the test set;

The K-fold cross validation method is used to divide the first training set into the second training set and the validation set;

The grid optimization method was used to find the best model parameters based on the second training set and the validation set, and then several prediction models for acute respiratory distress syndrome mortality were constructed using the random forest algorithm based on the best model parameters.

Each acute respiratory distress syndrome mortality prediction model was used to test the test set and obtain the prediction results of each fold;

The prediction results of each fold were averaged to obtain the prediction performance results corresponding to each acute respiratory distress syndrome mortality prediction model;

According to the obtained prediction performance results, the model with the best prediction performance results is selected from various acute respiratory distress syndrome mortality prediction models as the final acute respiratory distress syndrome mortality prediction model.

In combination with the above technical solutions and the technical problems solved, please analyze the advantages and positive effects of the technical solutions to be protected by the present invention from the following aspects:

First, in view of the technical problems existing in the above-mentioned prior art and the difficulty of solving the problems, the technical solutions to be protected by the present invention and the results and data during the research and development process are closely combined to analyze in detail and deeply how the technical solutions of the present invention solve the technical problems, and some creative technical effects brought about after solving the problems. The specific description is as follows:

The present invention identifies the outline of the detected ICU critically ill patient from each ICU critically ill patient image containing the detected ICU critically ill patient through a respiratory rate detection module, determines the detection area in each outline, and determines the respiratory rate of the detected ICU critically ill patient according to a change curve of the area of the detection area over time. The detection process does not require human intervention, and the respiratory rate can be detected more conveniently. At the same time, after obtaining the medical data of the ICU critically ill patient through the prediction module, an acute respiratory distress syndrome mortality prediction model is trained through a machine learning method, and finally the acute respiratory distress syndrome mortality prediction model is used to predict the mortality rate of ARDS patients. Machine learning is applied to the prediction of the mortality rate of ARDS patients, and the model trained by machine learning can accurately and objectively predict the mortality rate of ARDS patients, providing more effective and feasible prediction information for clinical physicians as a reference.

Second, considering the technical solution as a whole or from the perspective of the product, the technical effects and advantages of the technical solution to be protected by the present invention are described in detail as follows:

The present invention identifies the outline of the detected ICU critically ill patient from each ICU critically ill patient image containing the detected ICU critically ill patient through a respiratory rate detection module, determines the detection area in each outline, and determines the respiratory rate of the detected ICU critically ill patient according to a change curve of the area of the detection area over time. The detection process does not require human intervention, and the respiratory rate can be detected more conveniently. At the same time, after obtaining the medical data of the ICU critically ill patient through the prediction module, an acute respiratory distress syndrome mortality prediction model is trained through a machine learning method, and finally the acute respiratory distress syndrome mortality prediction model is used to predict the mortality rate of ARDS patients. Machine learning is applied to the prediction of the mortality rate of ARDS patients, and the model trained by machine learning can accurately and objectively predict the mortality rate of ARDS patients, providing more effective and feasible prediction information for clinical physicians as a reference.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a structural block diagram of an acute respiratory distress syndrome early warning system for critically ill ICU patients provided by an embodiment of the present invention.

FIG. 2 is a flow chart of a detection method of a respiratory rate detection module provided in an embodiment of the present invention.

FIG3 is a flow chart of a prediction method of a prediction module provided by an embodiment of the present invention.

In Figure 1: 1. Physiological index detection module; 11. Data normalization processing unit; 2. Respiratory rate detection module; 3. Central control module; 4. Analysis module; 5. Prediction module; 6. Health assessment module; 7. Early warning module; 71. Early warning calculation unit; 8. Display module.

DETAILED DESCRIPTION

In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

1. Explanatory Examples In order to enable those skilled in the art to fully understand how to implement the present invention, this section provides an illustrative example that expands and describes the technical solution of the claims.

As shown in Figure 1, the acute respiratory distress syndrome early warning system for critically ill patients in the ICU provided by an embodiment of the present invention includes: a physiological index detection module 1, a respiratory rate detection module 2, a central control module 3, an analysis module 4, a prediction module 5, a health assessment module 6, an early warning module 7, and a display module 8.

The physiological index detection module 1 is connected to the central control module 3 and is used to detect the patient's physiological index through medical equipment;

The physiological index detection module 1 includes a data normalization processing unit 11, which is used to receive pre-processing operations of detecting physiological indexes of patients from medical equipment and perform normalization processing on the data;

A respiratory rate detection module 2, connected to the central control module 3, is used to detect the patient's respiratory rate;

The central control module 3 is connected with the physiological index detection module 1, the respiratory rate detection module 2, the analysis module 4, the prediction module 5, the health assessment module 6, the early warning module 7, and the display module 8, and is used to control the normal operation of each module;

The prediction module 5 is connected to the central control module 3 and is used to predict the mortality rate of acute respiratory distress syndrome in critically ill patients in the ICU;

The analysis module 4 is connected to the central control module 3, and based on the results of the respiratory rate detection module and the prediction module, performs a comprehensive analysis of the acute respiratory distress syndrome of the critically ill patients in the ICU by analyzing the respiratory rate of the patient and the predicted value of the acute respiratory distress syndrome mortality rate;

The health assessment module 6 is connected to the central control module 3 and is used to assess the patient's health;

The early warning module 7 is connected to the central control module 3 and is used to warn of acute respiratory distress syndrome for critically ill patients in the ICU;

The early warning module 7 includes an early warning calculation unit 71, which compares the patient's respiratory rate value detected by the respiratory rate detection module 2, the predicted value of acute respiratory distress syndrome mortality, and the result of the analysis module 4 with the given health threshold value. If the health threshold value is exceeded, an alarm message is issued; otherwise, the health information is fed back to the central control module 3 and displayed in the display module 8;

The display module 8 is connected to the central control module 3 and is used to display the patient's physiological index, respiratory rate, analysis results, prediction results, and health assessment results through a display.

The present invention identifies the outline of the detected ICU critically ill patient from each ICU critically ill patient image containing the detected ICU critically ill patient through a respiratory rate detection module, determines the detection area in each outline, and determines the respiratory rate of the detected ICU critically ill patient according to a change curve of the area of the detection area over time. The detection process does not require human intervention, and the respiratory rate can be detected more conveniently. At the same time, after obtaining the medical data of the ICU critically ill patient through the prediction module, an acute respiratory distress syndrome mortality prediction model is trained through a machine learning method, and finally the acute respiratory distress syndrome mortality prediction model is used to predict the mortality rate of ARDS patients. Machine learning is applied to the prediction of the mortality rate of ARDS patients, and the model trained by machine learning can accurately and objectively predict the mortality rate of ARDS patients, providing more effective and feasible prediction information for clinical physicians as a reference.

As shown in FIG. 2 , the respiratory rate detection module detection method provided by the present invention is as follows:

S101, acquiring a plurality of ICU critically ill patient images including a detected ICU critically ill patient; performing enhancement processing on the ICU critically ill patient images; identifying the outline of the detected ICU critically ill patient from each ICU critically ill patient image; and determining a detection area from each of the outlines;

Determining a detection area from each of the contours comprises:

For each ICU critically ill patient image, identifying the head region of the detected ICU critically ill patient from the current ICU critically ill patient image;

Determine the lowest point of the head area, and the head distance between the highest point of the head area and the lowest point in the vertical direction;

Determine a first straight line parallel to the horizontal direction passing through the lowest point; determine a second straight line below the first straight line and having a distance from the first straight line that is a multiple of a first preset value of the head distance;

Determine a third straight line below the second straight line, the distance from the second straight line being a multiple of a second preset value of the head distance;

determining a closed area between the second straight line and the third straight line formed by the second straight line, the third straight line and the contour;

Using the closed area as the detection area corresponding to the current ICU critically ill patient image;

S102, determining the area of each detection area; determining a change curve of the area over time; and determining the respiratory rate of the detected ICU critically ill patient based on the change curve.

The present invention uses a respiratory rate detection module to identify the outline of a detected ICU critically ill patient from each ICU critically ill patient image that includes the detected ICU critically ill patient, determine a detection area in each outline, and determine the respiratory rate of the detected ICU critically ill patient according to a curve of the change of the area of the detection area over time. The detection process does not require human intervention, and the respiratory rate can be detected more conveniently.

The method provided by the present invention for identifying the head region of the detected ICU critically ill patient from the current ICU critically ill patient image includes:

Determine the R, G, and B of each pixel in the current ICU critically ill patient image;

According to the conversion formula and the R, G, B of each pixel, the parameter Cb and the parameter Cr corresponding to each pixel are determined, wherein the conversion formula is:

According to Formula 1, the intermediate parameters n ₁ and n ₂ of each pixel are determined, where Formula 1 is:

According to Formula 2, the judgment parameter P of each pixel is determined, where Formula 2 is:

Determine whether the judgment parameter of each pixel point is less than or equal to 1, and if so, determine that the pixel point belongs to the head area corresponding to the current ICU critically ill patient image; otherwise, determine that the pixel point does not belong to the head area corresponding to the current ICU critically ill patient image;

The head region corresponding to the current ICU critically ill patient image is determined according to the pixel points belonging to the head region.

The detection method provided by the present invention further comprises: establishing a rectangular coordinate system;

Loading each of the ICU critically ill patients' images into the rectangular coordinate system;

The determining of the lowest point of the head region and the head distance between the highest point of the head region and the lowest point in the vertical direction comprises:

Taking the point with the smallest vertical axis coordinate in the head region as the lowest point;

determining a maximum longitudinal coordinate in the head region;

The difference between the maximum longitudinal coordinate and the longitudinal coordinate of the lowest point is taken as the head distance;

The determining of a first straight line parallel to the horizontal direction passing through the lowest point comprises:

A straight line passing through the lowest point and parallel to the horizontal axis of the rectangular coordinate system is taken as the first straight line.

As shown in FIG3 , the prediction method of the prediction module 5 provided by the present invention is as follows:

S201, constructing a medical database, acquiring medical data of critically ill patients in the ICU; storing the acquired medical data in the medical database; and preprocessing the acquired medical data, wherein the preprocessing includes sample screening and feature extraction;

S202, classifying the preprocessed data according to the number of survival days of the patients to obtain positive groups and negative groups of three mortality prediction models, respectively, wherein the three mortality prediction models include an in-hospital mortality prediction model, a 30-day mortality prediction model, and a one-year mortality prediction model;

S203, performing inter-group analysis on the positive group and negative group of the three mortality prediction models respectively, and screening out the features with significant inter-group differences; using the random forest algorithm to establish an acute respiratory distress syndrome mortality prediction model based on the features with significant inter-group differences; using the acute respiratory distress syndrome mortality prediction model to predict the object to be predicted, and obtaining the prediction result of acute respiratory distress syndrome mortality.

After acquiring the medical data of critically ill patients in the ICU through a prediction module, the present invention trains an acute respiratory distress syndrome mortality prediction model through a machine learning method, and finally adopts the acute respiratory distress syndrome mortality prediction model to predict the mortality of ARDS patients. By applying machine learning to the prediction of the mortality of ARDS patients, the mortality of ARDS patients can be accurately and objectively predicted through the model trained by machine learning, providing clinical physicians with more effective and feasible prediction information as a reference.

The method for obtaining medical data of critically ill patients in ICU provided by the present invention includes downloading the medical data of critically ill patients in ICU from a MIMIC-III database.

The pretreatment method provided by the present invention comprises:

The obtained medical data were sampled and screened according to the inclusion criteria and exclusion criteria to obtain the screened samples. The inclusion criteria included patients aged 18 years or older who were admitted to the intensive care unit and diagnosed with acute respiratory distress syndrome according to the Berlin standard. The exclusion criteria included any one of the following: data with incomplete data records in the MIMIC-III database, patients aged less than 18 years old, patients receiving palliative therapy, and patients with an ICU record time of less than 48 hours;

For the screened samples, the variable features of each sample for modeling are extracted.

The pretreatment method provided by the present invention also includes:

Multiple imputation of missing data in acquired medical data.

The positive group and negative group of the three mortality prediction models provided by the present invention are specifically:

The positive group of the in-hospital mortality prediction model is the data of patients who died in hospital, and the negative group of the in-hospital mortality prediction model is the data of patients who survived during hospitalization; the positive group of the 30-day mortality prediction model is the data of patients who died within 30 days after hospitalization, and the negative group of the 30-day mortality prediction model is the data of patients who survived within 30 days after hospitalization; the positive group of the one-year mortality prediction model is the data of patients who died within one year after hospitalization, and the negative group of the one-year mortality prediction model is the data of patients who survived within one year after hospitalization.

The model building method provided by the present invention comprises:

The K-fold cross-validation method was used to divide the features with significant differences between groups into the first training set and the test set;

The K-fold cross validation method is used to divide the first training set into the second training set and the validation set;

The grid optimization method was used to find the best model parameters based on the second training set and the validation set, and then several prediction models for acute respiratory distress syndrome mortality were constructed using the random forest algorithm based on the best model parameters.

Each acute respiratory distress syndrome mortality prediction model was used to test the test set and obtain the prediction results of each fold;

The prediction results of each fold were averaged to obtain the prediction performance results corresponding to each acute respiratory distress syndrome mortality prediction model;

According to the obtained prediction performance results, the model with the best prediction performance results is selected from various acute respiratory distress syndrome mortality prediction models as the final acute respiratory distress syndrome mortality prediction model.

2. Application Examples: In order to prove the creativity and technical value of the technical solution of the present invention, this section provides application examples of the claimed technical solution on specific products or related technologies.

When the present invention is working, firstly, the physiological index of the patient is detected by medical equipment through the physiological index detection module 1; the respiratory rate of the patient is detected by the respiratory rate detection module 2; secondly, the central control module 3 analyzes the acute respiratory distress syndrome of the critically ill patients in the ICU through the analysis module 4; the mortality rate of the acute respiratory distress syndrome of the critically ill patients in the ICU is predicted through the prediction module 5; the health of the patient is evaluated through the health assessment module 6; then, the acute respiratory distress syndrome of the critically ill patients in the ICU is warned through the early warning module 7; finally, the patient's physiological index, respiratory rate, analysis results, prediction results, and health assessment results are displayed on a display through the display module 8.

It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and/or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits 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., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

3. Evidence of the effects of the embodiments. The embodiments of the present invention have achieved some positive effects during the development or use process, and indeed have great advantages over the prior art. The following content is described in conjunction with the data, charts, etc. of the test process.

The present invention identifies the outline of the detected ICU critically ill patient from each ICU critically ill patient image containing the detected ICU critically ill patient through a respiratory rate detection module, determines the detection area in each outline, and determines the respiratory rate of the detected ICU critically ill patient according to a change curve of the area of the detection area over time. The detection process does not require human intervention, and the respiratory rate can be detected more conveniently. At the same time, after obtaining the medical data of the ICU critically ill patient through the prediction module, an acute respiratory distress syndrome mortality prediction model is trained through a machine learning method, and finally the acute respiratory distress syndrome mortality prediction model is used to predict the mortality rate of ARDS patients. Machine learning is applied to the prediction of the mortality rate of ARDS patients, and the model trained by machine learning can accurately and objectively predict the mortality rate of ARDS patients, providing more effective and feasible prediction information for clinical physicians as a reference.

The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims (10)

1. An acute respiratory distress syndrome early warning system for critically ill patients in an ICU, characterized in that the acute respiratory distress syndrome early warning system for critically ill patients in the ICU comprises: Physiological index detection module, respiratory rate detection module, central control module, analysis module, prediction module, health assessment module, early warning module, display module; A physiological index detection module, connected to the central control module, is used to detect the patient's physiological index through medical equipment; The physiological index detection module includes a data normalization processing unit, which is used to receive the pre-processing operation of detecting the physiological index of the patient from the medical device and perform normalization processing on the data; A respiratory rate detection module, connected to the central control module, is used to detect the patient's respiratory rate; The central control module is connected with the physiological index detection module, the respiratory rate detection module, the analysis module, the prediction module, the health assessment module, the early warning module, and the display module to control the normal operation of each module; The prediction module is connected to the central control module and is used to predict the mortality rate of acute respiratory distress syndrome in critically ill patients in the ICU; The analysis module is connected to the central control module, and based on the results of the respiratory rate detection module and the prediction module, conducts a comprehensive analysis of the acute respiratory distress syndrome of critically ill patients in the ICU by analyzing the respiratory rate of the patient and the predicted value of the acute respiratory distress syndrome mortality rate; A health assessment module, connected to the central control module, is used to assess the patient's health; The early warning module is connected to the central control module and is used to warn of acute respiratory distress syndrome for critically ill patients in the ICU; The early warning module includes an early warning calculation unit, which compares the patient's respiratory rate value detected by the respiratory rate detection module, the predicted value of acute respiratory distress syndrome mortality, and the results of the analysis module with the given health threshold, and if the health threshold is exceeded, an alarm message is issued; otherwise, the health information is fed back to the central control module and displayed in the display module; The display module is connected to the central control module and is used to display the patient's physiological index, respiratory rate, analysis results, prediction results, and health assessment results through the display. 2. The acute respiratory distress syndrome early warning system for critically ill patients in the ICU according to claim 1, characterized in that the respiratory rate detection module detection method is as follows: (1) acquiring a plurality of ICU critically ill patient images including the ICU critically ill patient to be detected; performing enhancement processing on the ICU critically ill patient images; identifying the outline of the ICU critically ill patient to be detected from each ICU critically ill patient image; and determining a detection area from each of the outlines; Determining a detection area from each of the contours comprises: For each ICU critically ill patient image, identifying the head region of the detected ICU critically ill patient from the current ICU critically ill patient image; Determine the lowest point of the head area, and the head distance between the highest point of the head area and the lowest point in the vertical direction; Determine a first straight line parallel to the horizontal direction passing through the lowest point; determine a second straight line below the first straight line and having a distance from the first straight line that is a multiple of a first preset value of the head distance; Determine a third straight line below the second straight line, the distance from the second straight line being a multiple of a second preset value of the head distance; determining a closed area between the second straight line and the third straight line formed by the second straight line, the third straight line and the contour; Using the closed area as the detection area corresponding to the current ICU critically ill patient image; (2) Determine the area of each detection area; determine a change curve of the area over time; and determine the respiratory rate of the critically ill ICU patient being detected based on the change curve. 3. The acute respiratory distress syndrome early warning system for ICU critically ill patients according to claim 2, characterized in that the step of identifying the head area of the detected ICU critically ill patient from the current ICU critically ill patient image comprises: Determine the R, G, and B of each pixel in the current ICU critically ill patient image; According to the conversion formula and the R, G, B of each pixel, the parameter Cb and the parameter Cr corresponding to each pixel are determined, wherein the conversion formula is:

According to Formula 1, the intermediate parameters n ₁ and n ₂ of each pixel are determined, where Formula 1 is:

According to Formula 2, the judgment parameter P of each pixel is determined, where Formula 2 is:

Determine whether the judgment parameter of each pixel point is less than or equal to 1, and if so, determine that the pixel point belongs to the head area corresponding to the current ICU critically ill patient image; otherwise, determine that the pixel point does not belong to the head area corresponding to the current ICU critically ill patient image; The head region corresponding to the current ICU critically ill patient image is determined according to the pixel points belonging to the head region. 4. The acute respiratory distress syndrome early warning system for critically ill patients in the ICU according to claim 2, characterized in that the detection method further comprises: establishing a rectangular coordinate system; Loading each of the ICU critically ill patients' images into the rectangular coordinate system; The determining of the lowest point of the head region and the head distance between the highest point of the head region and the lowest point in the vertical direction comprises: Taking the point with the smallest vertical axis coordinate in the head region as the lowest point; determining a maximum longitudinal coordinate in the head region; The difference between the maximum longitudinal coordinate and the longitudinal coordinate of the lowest point is taken as the head distance; The determining of a first straight line parallel to the horizontal direction passing through the lowest point comprises: A straight line passing through the lowest point and parallel to the horizontal axis of the rectangular coordinate system is taken as the first straight line. 5. The acute respiratory distress syndrome early warning system for critically ill patients in the ICU according to claim 1, characterized in that the prediction method of the prediction module is as follows: 1) Building a medical database to obtain medical data of critically ill patients in the ICU; storing the obtained medical data in the medical database; and preprocessing the obtained medical data, wherein the preprocessing includes sample screening and feature extraction; 2) Classifying the preprocessed data according to the number of days the patients survived to obtain positive and negative groups of three mortality prediction models, respectively. The three mortality prediction models include an in-hospital mortality prediction model, a 30-day mortality prediction model, and a one-year mortality prediction model; 3) Inter-group analysis was performed on the positive and negative groups of the three mortality prediction models to screen out the features with significant inter-group differences; a random forest algorithm was used to establish an acute respiratory distress syndrome mortality prediction model based on the features with significant inter-group differences; the acute respiratory distress syndrome mortality prediction model was used to predict the objects to be predicted, and the prediction results of acute respiratory distress syndrome mortality were obtained. 6. The acute respiratory distress syndrome early warning system for critically ill ICU patients as claimed in claim 5, characterized in that obtaining the medical data of critically ill ICU patients comprises downloading the medical data of critically ill ICU patients from the MIMIC-III database. 7. The acute respiratory distress syndrome early warning system for critically ill ICU patients according to claim 5, characterized in that the pretreatment method comprises: The obtained medical data were sampled and screened according to the inclusion criteria and exclusion criteria to obtain the screened samples. The inclusion criteria included patients aged 18 years or older who were admitted to the intensive care unit and diagnosed with acute respiratory distress syndrome according to the Berlin standard. The exclusion criteria included any one of incomplete data records in the MIMIC-III database, patients aged less than 18 years, patients receiving palliative therapy, and patients with an ICU record time of less than 48 hours. For the screened samples, the variable features of each sample for modeling are extracted. 8. The acute respiratory distress syndrome early warning system for critically ill ICU patients according to claim 7, characterized in that the pretreatment method further comprises: Multiple imputation of missing data in acquired medical data. 9. The acute respiratory distress syndrome early warning system for critically ill ICU patients according to claim 5, characterized in that the positive group and the negative group of the three mortality prediction models are specifically: The positive group of the in-hospital mortality prediction model is the data of patients who died in hospital, and the negative group of the in-hospital mortality prediction model is the data of patients who survived during hospitalization; the positive group of the 30-day mortality prediction model is the data of patients who died within 30 days after hospitalization, and the negative group of the 30-day mortality prediction model is the data of patients who survived within 30 days after hospitalization; the positive group of the one-year mortality prediction model is the data of patients who died within one year after hospitalization, and the negative group of the one-year mortality prediction model is the data of patients who survived within one year after hospitalization. 10. The acute respiratory distress syndrome early warning system for critically ill ICU patients according to claim 5, characterized in that the model establishment method comprises: The K-fold cross-validation method was used to divide the features with significant differences between groups into the first training set and the test set; The K-fold cross validation method is used to divide the first training set into the second training set and the validation set; The grid optimization method was used to find the best model parameters based on the second training set and the validation set, and then several prediction models for acute respiratory distress syndrome mortality were constructed using the random forest algorithm based on the best model parameters. Each acute respiratory distress syndrome mortality prediction model was used to test the test set and obtain the prediction results of each fold; The prediction results of each fold were averaged to obtain the prediction performance results corresponding to each acute respiratory distress syndrome mortality prediction model; According to the obtained prediction performance results, the model with the best prediction performance results is selected from various acute respiratory distress syndrome mortality prediction models as the final acute respiratory distress syndrome mortality prediction model.

Images