CN115910310A - Method for determining standard threshold of cardio-pulmonary resuscitation pressing posture and processor - Google Patents

Abstract

The invention relates to a method for determining a standard threshold value of a cardio-pulmonary resuscitation (CPR) compression posture and a processor, wherein the method at least comprises the following steps: the method comprises the steps of simultaneously collecting CPR actions by adopting a first optical component and a second optical component which have different collection angles, extracting bone point data of a human body based on first action data collected by the first optical component and second action data collected by the second optical component, calculating at least double-arm posture angle data and gravity center matching angle data related to the CPR actions, and selecting single-side skewed state distribution rules of the double-arm posture angle data and the gravity center matching angle data with a plurality of CPR action specifications to determine a reasonable range of the double-arm posture angle data and a reasonable range of the gravity center matching angle data. According to the CPR compression posture standard threshold, the CPR action has a data standard for evaluation, an objective standardized evaluation standard is constructed, and the deviation of subjective judgment of a mentor is avoided.

Description

Method for determining standard threshold of cardio-pulmonary resuscitation pressing posture and processor

Technical Field

The invention relates to the technical field of artificial intelligence emergency training, in particular to a method and a processor for determining a standard threshold value of a cardio-pulmonary resuscitation pressing posture.

Background

Once sudden cardiac arrest occurs, if the sudden cardiac arrest cannot be rescued in time, irreversible damage to the brain and other important organs and tissues of a human body can be caused after 4-6 min, and even the life is threatened. There is a need to immediately perform high quality Cardiopulmonary resuscitation (CPR) on site. High quality cardiopulmonary resuscitation is the basis for basic and advanced life support. The reason why the quality is enhanced is that if the standard is not met, tissues and organs cannot obtain enough perfusion, so that the recovery success rate is obviously reduced, particularly, the nervous system is very sensitive to ischemia and hypoxia, and many patients recover circulation through recovery but cause irreversible brain injury, thereby seriously affecting the life quality after recovery. The first manual of CPR was issued in 1966 by a special CPR committee of the national research council of the national academy of medicine of the american national academy of sciences. Cardiac arrest remains a leading cause of life and health risks to humans more than half a century after the release of the first guidelines. Guidelines have always been directed to optimizing CPR operation using current relevant evidence to formulate definitive actionable criteria to save lives and be highly effective. However, applying these indicators to CPR rescue is not an easy task and requires professional training of CPR many times and even repeatedly. Most of the current CPR training depends on the personal ability and subjective judgment of a training instructor, and a unified and standardized monitoring means is lacked, so that the training level is not uniform.

Moreover, most of the currently known human skeleton extraction algorithms are constructed based on natural standing positions or other special moving positions, however, the CPR operation is kneeling, and no skeleton extraction algorithm specially used for identifying CPR actions exists at present. The accuracy of the existing bone extraction algorithm for extracting CPR actions is poor, and the bone extraction algorithm is the basis of CPR in subsequent contact with artificial intelligence technology and is very important. Therefore, in order to overcome the defects of the prior art, the cardiopulmonary resuscitation compression posture standard threshold is determined based on a skeleton extraction algorithm.

The prior art also proposes a wide variety of training systems and detection systems in order to achieve high quality training for cardiopulmonary resuscitation. For example, chinese patent publication No. CN105224383A discloses a cardiopulmonary resuscitation simulation system, in which a user can select an operation mode and an execution parameter of cardiopulmonary resuscitation by setting a display unit, a processing unit can simulate cardiopulmonary resuscitation according to the set operation mode and execution parameter and simulate a simulation result, a curve generating unit can generate a pressing force waveform, a cardiac output waveform, and a cerebral/coronary blood flow waveform and display them in a curve display area, a cycle average value calculating unit can calculate a cycle average value of each parameter and display them in the parameter display area, and an evaluating unit can obtain a comparison result of whether the cardiopulmonary resuscitation performed is effective or not by comparing the simulation result with a standard threshold value.

For another example, chinese patent publication No. CN110990649A discloses a cardiopulmonary resuscitation interactive training system based on gesture recognition technology, which includes a somatosensory recognition probe and interactive software, where the interactive software is installed in a computer device, the somatosensory recognition probe is used for connecting with the computer device and being collected somatosensory information by the interactive software, and the interactive software is executed by a computer to complete the following steps: acquiring real-time posture data of human body feeling through a body feeling identification probe; comparing the real-time posture data with the cardiopulmonary resuscitation standard posture data in the database and outputting a comparison result; and displaying the comparison result on a display device. The posture acquisition device can acquire the posture of a person during cardio-pulmonary resuscitation training through the somatosensory recognition probe, can compare the posture with standard cardio-pulmonary resuscitation actions, and can output a comparison result, so that whether the actions are standard or not can be checked in real time, action correction can be performed, and training is completed.

The standard of the monitoring basis of the prior art comes from direct training of the neural network on the collected CPR actions, but the collected CPR actions have larger difference with the skeleton actions of an operator due to the shielding of the clothes on the surface of the human body and the hardness of the difference characteristics of the human body, and the detection standard of the CPR action detection model obtained based on the neural network training is not accurate enough on the basis. Moreover, the CPR motion detection model obtained by directly training CPR motion based on the neural network requires that the acquisition angle for acquiring CPR motion images and the acquisition angle during modeling can accurately detect two-dimensional CPR motion, otherwise, accurate detection cannot be performed. This has just formed the restriction to the locating place of collection equipment for collection equipment can not put wantonly based on the condition in actual place.

Therefore, based on the defects that the detection standard progress of the CPR detection model in the prior art is low and the acquisition angle of the acquisition device is limited, the invention hopes to provide a new standard threshold determination method and also provide a standard threshold which can enable the detection model to realize high-precision detection by arbitrarily setting the acquisition angle of the acquisition device.

Furthermore, on the one hand, due to the differences in understanding to the person skilled in the art; on the other hand, since the applicant has studied a great deal of literature and patents when making the present invention, but the disclosure is not limited thereto and the details and contents thereof are not listed in detail, it is by no means the present invention has these prior art features, but the present invention has all the features of the prior art, and the applicant reserves the right to increase the related prior art in the background.

Disclosure of Invention

In view of the deficiencies of the prior art, the present invention provides a method for determining a standard threshold value of a cardiopulmonary resuscitation compression posture, the method at least comprising: the method comprises the steps of simultaneously collecting CPR actions by adopting a first optical component and a second optical component which have different collection angles, extracting bone point data of a human body based on first action data collected by the first optical component and second action data collected by the second optical component, calculating at least double-arm posture angle data and gravity center matching angle data related to the CPR actions, and selecting single-side skewed state distribution rules of the double-arm posture angle data and the gravity center matching angle data with a plurality of CPR action specifications to determine a reasonable range of the double-arm posture angle data and a reasonable range of the gravity center matching angle data.

Aiming at the defect that a CPR motion detection model in the prior art does not have a clear standard threshold, the CPR motion detection model establishes the standard threshold with the CPR compression posture, so that the CPR motion detection model can accurately provide more accurate judgment and comparison for the monitored motion based on the standard threshold, and can be applied to the training of CPR motion and the standard examination of CPR motion so as to improve the accuracy of the training and the examination of CPR motion.

Specifically, the method and the device calculate the double-arm angle and the gravity center matching angle of each sample data by counting and calculating the sample data. The posture of the CPR action has high correlation with the angle of both arms and the matching angle of the center of gravity. When the angle of the two arms and the matching angle of the center of gravity are in a reasonable range, the force of the two arms of the CPR operator is balanced, and when the matching angle of the center of gravity is in a reasonable range, the forward tilting amplitude of the CPR operator is reasonable, so that the waist and back fatigue caused by unreasonable back force can be avoided. When the angle of the double arms and the central matching angle are in reasonable ranges, a triangle formed by connecting lines between straight lines of shoulders and wrists of the double arms and the double shoulders is closer to an isosceles triangle, so that the force difference of the double arms is not too large, and the lateral deviation of the gravity center of an operator is smaller.

If the range of the double arms of the operator and the range of the matching angle of the gravity center are unreasonable, the situation that the double arms of the operator exert force unevenly is shown, meanwhile, the lateral deviation of the gravity center is large, and the limb fatigue of the operator is easy to occur due to the fact that the double arms exert force unevenly. Unreasonable matching angle of the center of gravity indicates insufficient or excessive forward tilting amplitude of the operator, so that the back of the operator is hard, and the waist of the operator is easy to fatigue. When fatigue is likely to occur to the CPR operator, the operator may experience an insufficient subsequent degree of compression, thereby reducing the success rate of cardiopulmonary resuscitation.

Therefore, the two-arm angle range and the gravity center matching angle range are simultaneously in a reasonable range, namely a threshold range, the posture of the CPR action operator can be obviously regulated to be a standard posture, the time that the operator feels fatigue is delayed, and the success rate of the cardiopulmonary resuscitation is improved.

Preferably, a reasonable range of the two-arm posture angle data is determined by selecting a 5% percentile based on a single-side skewed distribution rule of the two-arm posture angle data, and a reasonable range of the gravity center matching angle data is determined by selecting a 95% percentile based on a single-side skewed distribution rule of the gravity center matching angle data.

In the conventional CPR movement detection, the system only focuses on the movement amplitude or the bending angle of the two arms and neglects the influence of body weight and mind on the force, so that even if the two-arm angle is qualified and the compression force is qualified, the distribution of the two-arm force of the trainee is balanced, and therefore, the trainee of the CPR movement feels tired and the compression force is insufficient in the later period. Compared with the traditional detection means which only restrains the angles of the two arms and lacks the gravity center restraint, the invention uses the gravity center matching angle as an objective parameter to detect the standard degree of the posture, so that the CPR action operator has balanced forces of the two arms during operation, the forward leaning range of the back is reasonable, the core standard posture requirement of balanced forces of the two arms and reasonable forward leaning range can be more easily realized and understood without the guidance of a human instructor, and the training and learning of the standard posture of the CPR action are easier.

Preferably, the collection angle deviation of the first optical assembly and the second optical assembly with different collection angles ranges from 30 degrees to 90 degrees.

The newly published article of literature, "application of multi-modal system in CPR", also mentions monitoring of CPR posture, which collects multi-channel signals of a Kinect camera and a wearable myoelectric cuff at the same time, and designs an intelligent algorithm for monitoring arm posture and gravity center change during compression, but the research has obvious limitations. The study is a black box algorithm obtained by machine learning, and the equipment must be kept as completely consistent as possible, otherwise, the experimental result cannot be widely applied. For example, moving the Kinect camera to a different location, or adding or removing a certain sensor from the current setup, the developed algorithm will no longer work. Unlike this study, the present invention first extracts the CPR operator's skeletal points using an intelligent algorithm and then compares them to the standard ranges obtained in this study. The parameters detected based on the invention are the arm angle and the gravity center angle, and an AI plus statistical method is adopted, so that the angles and the distances of the camera are not required to be completely the same in each experiment and future application, and the result is not obviously influenced as long as the angles and the distances are changed within a certain range. Secondly, in the multi-modal research, a single camera is adopted to collect the pressing posture of a subject, the specific distance and angle of the camera are not described, and when the research is carried out, the fact that the single camera has a blind area is found, and the pressing posture data can be collected more accurately in multiple angles by collecting at least 2 angles simultaneously. In addition, the trainee need not to wear any equipment in this research, does not also receive the influence of other equipment, and its convenience, generalization and compatibility are better, and the feasibility of popularization and application later is higher.

Preferably, the collection angle deviation range of the first optical assembly and the second optical assembly of the non-same collection angle is 45 degrees, wherein the first optical assembly collects the first motion data of the CPR motion in a first angular direction, the first angular direction is towards the front of the body of the CPR motion operator, and the second optical assembly collects the second motion data of the CPR motion in a second angular direction.

When the collection angle deviation range is 45 degrees, the collected first motion data and second motion data can more accurately calculate the double-arm posture angle data and the gravity center matching angle data, the collection angle deviation range is 45 degrees, which is optimal, and the most accurate deviation angle is calculated.

Preferably, the two-arm posture angle data and the gravity center matching angle data are analyzed based on a line segment formed by connecting human skeletal points; the right arm posture angle refers to an angle formed by connecting lines of skeleton points of a right shoulder, a right elbow joint and a right wrist, the left arm posture angle refers to an angle formed by connecting lines of skeleton points of a left shoulder, a left elbow joint and a left wrist, and the gravity center matching angle refers to an included angle between the gravity center moving direction of a CPR action operator and a plane normal vector.

Preferably, the manner of selecting the two-arm posture angle data and the gravity center matching angle data of a plurality of CPR action specifications at least comprises: selecting two-arm posture angle data and gravity center matching angle data with qualified confidence; and marking indexes of CPR actions by at least two professionals, and selecting the two-arm posture angle data and the gravity center matching angle data of the CPR actions with qualified indexes of the CPR actions for distribution statistics.

Preferably, the method further comprises: the labeling content of the professional for indicating the CPR action at least comprises the following contents: arm extension and its index, and the matching angle of gravity center and its index.

Preferably, the method further comprises: performing data preprocessing on the first motion data acquired by the first optical assembly and the second motion data acquired by the second optical assembly before extracting the bone point data, wherein the data preprocessing method comprises data missing value and abnormal value analysis, data cleaning, feature selection and/or data transformation.

Preferably, the two-arm posture angle data comprises left arm posture angle data and right arm posture angle data, the reasonable range of the left arm posture angle data is 169.24-180 degrees, the reasonable range of the right arm posture angle data is 168.49-180 degrees, and the reasonable range of the gravity center matching angle data is 0-18.46 degrees.

The invention also provides a processor capable of running a method for determining a cardiopulmonary resuscitation compression posture standard threshold, the processor being configured to: receiving first motion data collected by the first optical assembly and second motion data collected by the second optical assembly at a non-same collection angle, extracting skeleton point data of a human body based on the first motion data and the second motion data, calculating at least double-arm posture angle data and gravity center matching angle data related to CPR motions, and selecting a plurality of single-side skewed state distribution rules of the double-arm posture angle data and the gravity center matching angle data with CPR motion specifications to determine a reasonable range of the double-arm posture angle data and a reasonable range of the gravity center matching angle data.

The invention constructs the standard cardiopulmonary resuscitation compression posture threshold special for recognizing CPR actions for the first time, compared with other skeleton extraction algorithms, the invention extracts the CPR actions more accurately and lays an important foundation for future research, particularly with artificial intelligence algorithms.

Drawings

FIG. 1 is a schematic illustration of a chest compression bone extraction of a preferred embodiment of the present invention;

FIG. 2 is an exemplary diagram of human key points for AlphaPose recognition in a preferred embodiment of the present invention;

FIG. 3 is a confidence statistics table of extracted human bone points for a preferred embodiment of the present invention;

FIG. 4 is a list of major compression errors and incidence rates for a preferred embodiment of the present invention;

FIG. 5 is a right arm and left arm pose angle histogram for a preferred embodiment of the present invention;

FIG. 6 is a center of gravity matching angle histogram of a preferred embodiment of the present invention;

FIG. 7 is a compression gesture specification standard range for a preferred embodiment of the present invention;

figure 8 is a schematic diagram of the center of gravity matching angle for a CPR compression gesture of the present invention;

FIG. 9 is a schematic illustration of the dual arm angle of a CPR compression gesture of the present invention;

FIG. 10 is a processor for implementing the standard threshold determination of the present invention.

List of reference numerals

0: a nose portion; 1: cervical vertebrae; 2: a right shoulder; 3: the right elbow joint; 4: a right wrist; 5: a left shoulder; 6: the left elbow; 7: a left wrist; 8: a right hip; 9: a right knee; 10: a right ankle; 11: a left hip; 12: the left knee; 13: a left ankle; 14: a right eye; 15: a left eye; 16: a right ear; 17: a left ear; 100: a collection end; 110: a ZED camera device; 120: a human simulator; 200: a data extraction module; 300: a preprocessing module; 400: a gesture detection module; 500: a receiving terminal; 60: a bone line segment; 61: a bone end point; 70: an actual distribution curve; 71: a normal distribution curve; γ: matching the angle of the center of gravity; α: a right arm angle; beta: left arm angle.

Detailed Description

The following detailed description is made with reference to the accompanying drawings.

In the present invention, the arm angle refers to the angle of bending of the upper arm and forearm of the arm, as shown in fig. 9.

The gravity center matching angle is as follows: the center of gravity of the operator moving in the CPR action moves the angle between the normal vector and the plane. As shown in FIG. 8, the motion of the midpoint A of the line connecting the right shoulder 2 and the left shoulder 5 to the midpoint B of the line connecting the right wrist 4 and the left wrist 7 generates a vector

(Vector)

And an included angle between the normal vector of the plane and the normal vector of the plane is a gravity center matching angle.

Currently, the known human skeleton extraction algorithm is mostly constructed based on a natural standing position or other special moving positions, however, the CPR operation is kneeling, and at present, there is no skeleton extraction algorithm specially used for identifying CPR actions. The accuracy of the existing bone extraction algorithm for extracting CPR actions is poor, and the bone extraction algorithm is the basis of CPR subsequently connected with artificial intelligence technology and is very important. Accordingly, the present invention provides a compression posture standard threshold value capable of recognizing CPR operation and a method for determining the same. Based on the cardiopulmonary resuscitation compression posture standard threshold determination method, a CPR posture detection model can be constructed and applied to CPR training and clinical CPR rescue.

When the method is applied, only an optical component capable of shooting is used for collecting CPR motion images and sending the images to a processor provided with a CPR posture detection model, so that the data of the skeleton points of CPR motion can be converted into numerical values in real time and judged to obtain a standard analysis result. The invention can also use the CPR posture detection model to obtain a suggestion on how to adjust the CPR action, so that unskilled personnel can also implement emergency rescue of the cardiopulmonary resuscitation in an emergency situation.

The invention relates to a method for determining a standard threshold value of a Cardio Pulmonary Resuscitation (CPR) compression posture, which at least comprises the following steps:

s1: the first optical component and the second optical component with different collection angles are adopted to simultaneously collect CPR actions,

s2: extracting skeletal point data of a human body based on first motion data acquired by the first optical assembly and second motion data acquired by the second optical assembly;

s3: preprocessing the extracted bone point data;

s4: calculating at least two-arm posture angle data and gravity center matching angle data related to CPR actions, and selecting a plurality of standard single-side skewed distribution rules of the two-arm posture angle data and the gravity center matching angle data of the CPR actions to determine a reasonable range of the two-arm posture angle data and a reasonable range of the gravity center matching angle data.

The invention relates to a method for determining a standard threshold value of a cardio-pulmonary resuscitation compression posture, which further comprises the following steps:

and marking indexes of CPR actions by at least two professionals, and selecting the two-arm posture angle data and the gravity center matching angle data of the CPR actions with qualified indexes of the CPR actions for distribution statistics. As shown in fig. 10, the construction apparatus for determining the compression posture standard threshold for CPR action includes at least a collecting terminal 100, a processor and a receiving terminal 500. The processor establishes a connection with the acquisition terminal 100 and the receiving terminal in a wired or wireless manner to transmit information. Each of the acquisition terminal 100, the processor, and the receiving terminal 500 is provided with an independent power line to supply power.

The collection end 100 includes a first optical assembly and a second optical assembly. The first optical assembly is used for collecting first motion information of the CPR gesture at a first coordinate. The first coordinate serves as a reference coordinate system. The second optical assembly is used for collecting second motion information of the CPR gesture at a second coordinate. The first coordinate system is different from the second coordinate system. The first optical assembly acquires first reference information of a CPR movement operator in a first coordinate system.

Preferably, the first optical assembly and the second optical assembly simultaneously acquire dynamic images of the CPR pose from different angles. Preferably, the collection angle deviation between the first optical assembly and the second optical assembly is 45 degrees. Preferably, the collection angle deviation between the first optical assembly and the second optical assembly is not limited to 45 degrees, but may be 30 degrees, 60 degrees, and the like. Preferably, the collection angle deviation between the first optical assembly and the second optical assembly ranges from 30 to 90 degrees. If the collection angle between the first optical assembly and the second optical assembly is offset by 90 degrees, the collection angle for collecting the side of the CPR operator cannot easily collect the center of gravity offset vector of the CPR operator. The collection angle deviation between the first optical assembly and the second optical assembly is preferably less than 90 degrees.

Preferably, the first optical assembly acquires the first motion data at a zero degree acquisition angle directly in front of the CPR motion. The second optical assembly acquires second motion data at a 45 degree acquisition angle lateral to the CPR motion. The two-arm posture angle data is determined from the bone point data collected by the first optical assembly. The center of gravity matching angle data is determined from the bone point data collected by the second optical assembly.

Alternatively, the first optical assembly collects at a side 45 degree angle and the second optical assembly collects at a front right angle. The center of gravity matching angle data is determined from the bone point data collected by the first optical assembly. The two-arm posture angle data is determined from the bone point data acquired by the second optical assembly.

The first optical assembly includes at least a camera assembly and a computing assembly. The second optical assembly includes at least a camera assembly and a computing assembly. The image pickup assembly is, for example, a ZED image pickup device 110. The computing assembly comprises a data extraction module 200 for extracting the skeletal point data of the human body based on the AlphaPose algorithm through the CPR motion image collected by the camera assembly. The bone point data includes features of the bone line segments 60 formed based on the bone points and their bone end points 61. The data extraction module 200 sends the bone point data to the processor. The data extraction module 200 is a calculator capable of running an alphaPose algorithm, and an application specific integrated circuit chip is arranged inside the calculator. The data extraction module 200 is connected to the camera module through a data line to receive and process the image data. The data extraction module 200 is connected to the processor through a data bus to extract the bone point data from the image data and send the bone point data to the processor.

Preferably, the data extraction module 200 of the present invention may not be disposed in the optical assembly, but disposed in the processor to become a part of the processor.

As shown in fig. 1, when an operator of CPR action is performing CPR action operations on the simulator 120, the data extraction module 200 can extract the operator's time-dependent bone endpoints 61 and bone line segments 60.

As shown in fig. 2, the bone points extracted by the data extraction module 200 include at least 18 main parts. The skeletal points mainly include a nose 0, a cervical vertebra 1, a right shoulder 2, a right elbow joint 3, a right wrist 4, a left shoulder 5, a left elbow 6, a left wrist 7, a right hip 8, a right knee 9, a right ankle 10, a left hip 11, a left knee 12, a left ankle 13, a right eye 14, a left eye 15, a right ear 16 and a left ear 17.

The data extraction module 200 also performs confidence statistics on the bone point data. Fig. 9 shows the average confidence statistics of the individual human bone points collected by the first optical assembly and the second optical assembly.

The data extraction modules 200 of the two optical assemblies each send the respective extracted bone data to the processor.

The processor may be one of a server, a remote server, an application specific integrated chip. The processor is used for executing the analysis step and the statistic step of the preprocessed bone point data. Preferably, the processor may be a combination of at least two application specific integrated chips or CPU processors, or the processor may be a separate application specific integrated chip or CPU capable of running the data pre-processing module program and the gesture detection module program. The application specific integrated chip or CPU can be applied in the form of a server or a cloud server.

The processor includes at least a pre-processing module 300 and a gesture detection module 400. Both the pre-processing module 300 and the gesture detection module 400 may be dedicated integrated chips or separate hardware modules of a CPU processor. When the preprocessing module 300 and the gesture detection module 400 are integrated on the same asic or CPU processor, the preprocessing module 300 and the gesture detection module 400 are running programs using the processor as a hardware carrier.

The processor is also provided with a first data transmission port and a second data transmission port. In the case where the preprocessing module 300 and the gesture detection module 400 are hardware modules of an application specific integrated chip or CPU, respectively, the preprocessing module 300 and the first data transmission port are connected by a data transmission line. The preprocessing module 300 and the posture detecting module 400 are connected through a data transmission line. The gesture detection module 400 and the second data transmission port are connected by a data transmission line. The first data transmission port and the second data transmission port may be a wired data transmission port component or a wireless data transmission port component, specifically which data transmission mode is wired transmission or wireless transmission. The wired data transmission port component is, for example, various types of USB transmission line ports. The wireless data transmission port component is, for example, a bluetooth data transmission communication component, a WIFI data transmission communication component, a ZigBee data transmission communication component, and the like.

Preferably, the processor may also be provided with only one data transmission port, which is connected in parallel with the preprocessing module 300 and the gesture detection module 400, for respectively transmitting and receiving data to and from the preprocessing module 300 and the gesture detection module 400.

Preferably, the preprocessing module 300 also performs data preprocessing on the received bone point data.

The data preprocessing step at least comprises the following steps:

s31: analyzing missing values and abnormal values of the data;

s32: data cleaning;

s33: selecting characteristics;

s34: and (5) data transformation.

The preprocessing module 300 transmits the preprocessed data to the gesture detection module 400 through a transmission line. Preferably, the gesture detection module 400 is capable of running a gesture detection model. The pose detection model is capable of calculating both-arm pose angle data and center-of-gravity matching angle data.

The posture detection module 400 calculates at least two-arm posture angle data of a two-arm posture of the CPR action operator based on the bone point data acquired by the first optical assembly. In particular, the bone point data collected by the first optical assembly comprises at least a right shoulder 2, a right elbow joint 3, a right wrist 4, a left shoulder 5, a left elbow 6 and a left wrist 7. The confidence coefficients of the right shoulder 2, the right elbow joint 3, the right wrist 4, the left shoulder 5, the left elbow 6 and the left wrist 7 are 0.94, 0.89, 0.93, 0.95, 0.90 and 0.87 respectively. As shown in fig. 9, the arm posture angle means an angle between the hand, elbow and shoulder, that is, the right arm posture angle is an angle α formed between the right shoulder 2, right elbow joint 3 and right wrist 4, and the left arm posture angle is an angle β formed between the left shoulder 5, left elbow 6 and left wrist 7.

The gesture detection module 400 calculates at least the CPR movement operator's center of gravity matching angle data based on the bone point data collected by the second optical assembly. The skeletal point data here includes at least a right shoulder 2, a left shoulder 5, a right wrist 4 and a left wrist 7. As shown in fig. 9, the confidence of the bone point data of the right shoulder 2, the left shoulder 5, the right wrist 4, and the left wrist 7 is 0.91, 0.81, 0.89, 0.88, respectively. As shown in fig. 8, the center of gravity matching angle is an angle in which the center of gravity of the CPR administrator moves in a direction perpendicular to the patient. As shown in FIG. 10, the motion of the midpoint A of the line connecting the right shoulder 2 and the left shoulder 5 to the midpoint B of the line connecting the right wrist 4 and the left wrist 7 generates a vector

Vector->

The included angle between the normal vector and the surface is a gravity center matching angle gamma.

The pose detection module 400 calculates an angle formed by the bone line segments or an angle between the vector and the normal vector based on a preset angle calculation formula.

The calculation formula of the included angle is as follows:

m ₁ representing the slope of the first line, m ₂ Showing the slope of the second line.

If the first straight line passes through the point P ₁ ＝[x ₁ ，y ₁ ]And P ₂ ＝[x ₂ ，y ₂ ]Definition of

The slope m is calculated as:

epsilon is 10 ^-9 。

In the prior art, the latest research on the application of a multi-modal system in CPR also mentions the monitoring of the CPR posture, collects the multichannel signals of a Kinect camera and a wearable myoelectric sleeve simultaneously, and designs an intelligent algorithm for monitoring the change of the arm posture and the gravity center during compression, but the research has obvious limitations. The study is a black box algorithm obtained by machine learning, and the equipment must be kept as completely consistent as possible, otherwise, the experimental result cannot be widely applied. For example, moving the Kinect camera to a different location, or adding or removing a certain sensor from the current setup, the developed algorithm will no longer work. Unlike this study, the present invention first extracts the CPR operator's skeletal points using an intelligent algorithm and then compares them to the standard ranges obtained in this study. The parameters detected based on the invention are the arm angle and the gravity center angle, and an AI plus statistical method is adopted, so that the angles and the distances of the camera are not required to be completely the same in each experiment and future application, and the result is not obviously influenced as long as the angles and the distances are changed within a certain range. Secondly, in the multi-modal research, a single camera is adopted to collect the pressing posture of a subject, the specific distance and angle of the camera are not described, and when the research is carried out, the fact that the single camera has a blind area is found, and the pressing posture data can be collected more accurately in multiple angles by collecting at least 2 angles simultaneously. In addition, the trainee need not to wear any equipment in this research, does not also receive the influence of other equipment, and its convenience, generalization and compatibility are better, and the feasibility of popularization and application later is higher. Moreover, the AI model in the prior art directly determines the angle of each posture of both arms by the motion image of the CPR posture, and is affected by the physical and mental image characteristics of the CPR operator, for example, the CPR posture is determined to be biased and cannot be eliminated by the fat and thin characteristics and the clothing shielding characteristics. The AI model in the prior art is not accurate enough in the assessment of the CPR posture, and has a large error.

In the prior art, because the center of gravity of a CPR posture cannot be determined due to the influence of human body characteristics and clothing shielding, the center of gravity shift of the movement of the CPR posture cannot be checked and judged, whether the center of gravity shift of a CPR operator is correct or not cannot be judged, and whether the force of the CPR operator is correct or not cannot be further judged.

The gesture detection module of the invention obtains angle data not by a deep learning algorithm but by a specific calculation formula. Regardless of the arrangement of the acquisition end, accurate angle data of the CPR operator can be calculated and obtained as long as the angle of the acquisition end is reasonable. Therefore, the position of the acquisition end is set randomly, the setting condition is simple, and the position of the acquisition end does not influence the recognition progress of the gesture detection module. Namely, in an actual application scene, the influence of the change of the acquisition angle of the acquisition end on the angle result is small, and the setting range of the acquisition end is wide.

Furthermore, the posture detection model of the present invention extracts and calculates both-arm posture angle data and gravity center matching data by using the human skeleton points. The skeleton data formed by the skeleton points of the human body cannot be influenced by the weight of the human body and the shielding of clothes, so that the calculation and judgment of the CPR posture angle can be more accurate.

The invention utilizes the skeleton point data to calculate, gets rid of the hardness of the body characteristic difference of people and the difference of clothes and clothes, can not directly determine the center of gravity of the action, but can determine the center of gravity offset direction through the vector formed by the middle point of the connecting line of the shoulders and the middle point of the connecting line of the wrists. Whether the exertion of the CPR operator is balanced and proper can be determined according to the angle between the vector and the normal vector of the plane. If the center of gravity matches the angle reasonably, the CPR operator's effort is balanced and appropriate.

Preferably, the processor is connected, preferably in a wireless or wireless manner, with at least one terminal. Specifically, the posture detection module 400 is connected to at least one terminal through a line or a wireless signal to transmit video data of motion images of the CPR operator, and calculated two-arm posture angle data and center-of-gravity matching angle data to the terminal. The terminal is used for displaying the motion image, the double-arm posture angle data and the gravity center matching angle data of the CPR operator to at least one expert. The terminal at least comprises a display component, an interaction component and an information storage component. I.e. the terminal is an electronic device allowing interaction. The terminal is an electronic device such as a tablet computer iPad, a notebook computer, a desktop computer, a smart phone, a smart watch, and smart glasses.

Preferably, the both-arm posture angle data and the center-of-gravity matching angle data are displayed on the display screen of the terminal so as not to obstruct the motion of the CPR operator.

The terminal is used by a professional familiar with CPR operating standards. Preferably, a terminal is provided to a professional. Preferably, the professional is preferably composed of three persons based on the principles of scientific statistics. The motion images of the CPR operator are individually labeled by three professionals. The professional is labeled based on the specified index content. The index content at least comprises two items: arm extension and its index, and center of gravity matching angle and its index.

The index of arm straightening refers to judging whether the arm posture is correct in the process of cardiopulmonary resuscitation. The index of the center of gravity matching angle indicates whether the center of gravity of the CPR administrator moves in a direction perpendicular to the patient. During the operation of a CPR operator, compression errors of CPR postures mainly include: wrist exertion, non-tilting fingers, center of gravity shifting (including basic center of gravity skew, center of gravity moving forward and backward, center of gravity moving left and right), elbow bending, etc. The wrist force, the fingers are not tilted, the center of gravity shifts (including the basic center of gravity is inclined, the center of gravity moves forwards and backwards, and the center of gravity moves leftwards and rightwards), and the elbow bending belongs to the errors with the highest incidence.

Therefore, the invention also carries out professional labeling on the CPR action based on the labeling index by a professional looking up the CPR action image, and excludes irregular data. The labeling indicators at least include whether the arm is straight and the center of gravity is correct.

The terminal sends the motion image labeled by the professional to the gesture detection module 400 through the second data transmission port of the processor. The posture detection module 400 receives the motion image containing the labeling information, and takes the two-arm posture angle data and the center-of-gravity matching angle data of the CPR posture conforming to the arm extension and the index thereof, the center-of-gravity matching angle and the index thereof as the specification data. Preferably, when the motion image acquired by the acquisition end is a professional group consisting of professionals, the two-arm posture angle data and the center-of-gravity matching angle data of the CPR posture conforming to each index are professional group specification data. Professional team normative data as a data set for the standard set of cardiopulmonary resuscitation. The posture detection module 400 receives the motion image containing the labeling information, and takes the two-arm posture angle data and the center-of-gravity matching angle data of the CPR posture which does not conform to the arm extension and the index thereof, the center-of-gravity matching angle and the index thereof as the non-standard data.

The gesture detection module 400 performs statistics on the screened professional group specification data.

Specifically, as shown in fig. 4, 28800 groups of human skeletal point coordinate data of a professional group specification data set are acquired. 7200 groups of human skeleton point coordinate data of a non-professional group data set are collected. In fig. 5 and 6, the thick curve represents the actual distribution curve 70. The thin curve represents a normal distribution curve 71. It can be seen that both arm posture angle data and center of gravity matching angle data are in accordance with the skewed distribution. And taking a 5% quantile as a normal value range for the angle data of the left arm and the right arm, and taking a 95% quantile as a normal value range for the gravity center matching angle data.

Specifically, the measurement data is described by mean ± standard deviation, and the interclass mean comparison is performed by independent sample t test. Because the arm angle is the unilateral skewed distribution data, 5% -10% percentile is used for calculating the reasonable range threshold, and similarly, the gravity center matching angle range is the unilateral skewed distribution data, and 90% -95% percentile is used for calculating the reasonable range threshold. All statistical analyses will be counted at a two-sided 0.05 significance level.

As shown in fig. 7, in the case of taking the 5% percentile, the reasonable range of the posture angle of the left arm is 169.24 to 180 degrees, and the reasonable range of the posture angle of the right arm is 168.49 to 180 degrees. In the case of a 95% percentile, a reasonable range of the center of gravity matching angle is 0-18.46 degrees. This is also the standard cardiopulmonary resuscitation compression posture threshold achieved by the present invention.

Preferably, when the cardiopulmonary resuscitation compression posture standard threshold of the present invention is applied to CPR motion monitoring or quality control, the CPR motion detection model or quality control model it constitutes can be continuously optimized.

As shown in fig. 10, the posture detection module 400 constructs a posture detection model based on a machine learning model using CPR operation data that meets the CPR compression posture standard threshold and has a qualified manual labeling index and standard threshold data as normative data, and may also be used as a posture quality control model.

Aiming at the posture detection model or the posture quality control model, the following steps can be realized:

s5: and (6) optimizing the model.

The step of model optimization comprises at least:

s51: establishing a 3D moving point model through standard actions of professionals; converting the 3D moving point model into a second 2DCPR action with the same collection angle as that of the collected first 2DCPR action; comparing a first detection result of the first 2DCPR action with a second detection result of the second 2DCPR action; if the first detection result is consistent with the second detection result, the posture detection model or the posture quality control model does not need to be optimized. And if the first detection result is inconsistent with the second detection result, optimizing the posture detection model or the posture quality control model.

It should be noted that the above-mentioned embodiments are exemplary, and that those skilled in the art, having benefit of the present disclosure, may devise various arrangements that are within the scope of the present disclosure and that fall within the scope of the invention. It should be understood by those skilled in the art that the present specification and figures are illustrative only and are not limiting upon the claims. The scope of the invention is defined by the claims and their equivalents. The present description contains several inventive concepts, such as "preferably", "according to a preferred embodiment" or "optionally", each indicating that the respective paragraph discloses a separate concept, the applicant reserves the right to submit divisional applications according to each inventive concept.

Claims (10)

1. A method for determining a cpr compression posture standard threshold, the method comprising at least:

the first optical component and the second optical component with different collection angles are adopted to simultaneously collect CPR actions,

extracting skeletal point data of a human body based on first motion data acquired by the first optical assembly and second motion data acquired by the second optical assembly and calculating at least two-arm posture angle data and center-of-gravity matching angle data related to CPR motion,

selecting a plurality of standard CPR action single-side skewed state distribution rules of the double-arm posture angle data and the gravity center matching angle data to determine the reasonable range of the double-arm posture angle data and the reasonable range of the gravity center matching angle data.

2. The method for determining a CPR compression posture standard threshold value according to claim 1,

selecting 5% percentile based on the unilateral skewness distribution rule of the double-arm posture angle data to determine the reasonable range of the double-arm posture angle data,

and selecting 95% percentile to determine the reasonable range of the gravity center matching angle data based on the unilateral skewness distribution rule of the gravity center matching angle data.

3. The method for determining the CPR compression posture standard threshold value according to claim 1 or 2, wherein the first optical component and the second optical component of the different collection angles have a collection angle deviation ranging from 30 to 90 degrees.

4. The method for determining CPR compression posture standard threshold according to any one of claims 1-3, wherein the non-identical acquisition angle of the first optical component and the second optical component has an acquisition angle deviation of 45 degrees, wherein,

the first optical assembly acquires first motion data of CPR motions in a first angular direction, the first angular direction being toward a front of a torso of a CPR motion operator,

the second optical assembly acquires second motion data of the CPR motions at a second angular orientation.

5. The method for determining CPR compression posture standard threshold according to any one of claims 1 to 4, wherein the two-arm posture angle data and the center-of-gravity matching angle data are analyzed based on a line segment formed by connecting human skeletal points; wherein the content of the first and second substances,

the posture angle of the right arm is an angle formed by connecting the skeleton points of the right shoulder, the right elbow joint and the right wrist,

the posture angle of the left arm is the angle formed by the connecting lines of the skeleton points of the left shoulder, the left elbow joint and the left wrist,

the gravity center matching angle refers to an included angle between the gravity center moving direction of the CPR action operator and a plane normal vector.

6. The method for determining CPR compression posture standard threshold value according to any one of claims 1-5, wherein the manner of selecting the angle data of both arms posture and the angle data of center of gravity matching for a plurality of CPR action specifications at least comprises:

selecting two-arm posture angle data and gravity center matching angle data with qualified confidence;

and marking indexes of the CPR actions by at least two professionals, and selecting the angle data of the posture of the two arms of the CPR actions and the angle data matched with the gravity center of the CPR actions qualified according to the indexes of the CPR actions for distribution statistics.

7. The method for determining a CPR compression posture standard threshold according to any one of claims 1 to 6, further comprising:

the labeling content of the professional for indicating the CPR action at least comprises the following contents: arm extension and its index, and the center of gravity matching angle and its index.

8. The method for determining the CPR compression posture standard threshold value according to any one of claims 1 to 7, wherein the method further comprises:

pre-processing data of the first motion data acquired by the first optical assembly and the second motion data acquired by the second optical assembly prior to extracting the bone point data,

the data preprocessing method comprises the steps of data missing value and abnormal value analysis, data cleaning, feature selection and/or data transformation.

9. The method for determining CPR compression posture standard threshold according to any one of claims 1 to 7, wherein the two-arm posture angle data comprises left-arm posture angle data and right-arm posture angle data,

the reasonable range of the left arm posture angle data is 169.24-180 degrees, the reasonable range of the right arm posture angle data is 168.49-180 degrees,

the center of gravity matches a reasonable range of angular data of 0-18.46.

10. A processor capable of running a method of determining a cardiopulmonary resuscitation compression posture standard threshold, the processor configured to:

receiving first motion data acquired by the first optical assembly and second motion data acquired by the second optical assembly at a non-identical acquisition angle,

extracting skeletal point data of a human body based on the first motion data and the second motion data and calculating at least two-arm posture angle data and center-of-gravity matching angle data related to CPR motion,

selecting a plurality of standard CPR action single-side skewed state distribution rules of the double-arm posture angle data and the gravity center matching angle data to determine the reasonable range of the double-arm posture angle data and the reasonable range of the gravity center matching angle data.

Images