An intelligent cardiopulmonary resuscitation training system
Through the intelligent cardiopulmonary resuscitation training system, human-computer interactive equipment and video signal recognition technology, the existing CPR training venues and time restrictions and inconsistency are solved, efficient action evaluation and multiple exercises are achieved, and training is improved Quality and popularization effect.
Patent Information
- Application Number
- CN202211368002.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-03
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2042-11-03
AI Technical Summary
The existing CPR training has problems such as venue and time limitations, inconsistent and standardization of teaching, lack of multiple practice opportunities, and difficulty in passing rate statistics, resulting in low quality and difficulty in popularizing training.
Design an intelligent cardiopulmonary resuscitation training system, including human-computer interaction equipment, host, camera equipment and simulated persons, provide unified independent practice and assessment through video signal recognition and action evaluation, realize action accuracy evaluation and score scoring, and support Practice and statistics multiple times.
It improves the unity and standardization of CPR training, provides flexible learning and assessment opportunities, improves the quality of training, and promotes the popularization and promotion of CPR.
Smart Images

Figure CN115586834B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of online training, and in particular to an intelligent cardiopulmonary resuscitation training system. Background Art
[0002] CPR (Cardio Pulmonary Resuscitation) training is the top priority of medical skills training at present. However, existing CPR training is almost all managed manually, which is not conducive to trainees truly mastering this technology. First, CPR requires centralized training. Sometimes, due to the limitation of venue and time, students may not be able to attend the training, which delays the learning content and makes it impossible to make up for the missed lessons, resulting in failure in the final assessment. At the same time, the time and location of the assessment also face the same problem. Second, because the teaching cannot be unified and standardized, different teaching details may be different, and different assessment standards may also be different, resulting in uneven actual operation of students. Third, because offline training is managed in a unified way, students cannot practice multiple times in one class, resulting in half-knowledge. Even if they pass the assessment, it will eventually lead to serious consequences. Fourth, because the training time is fixed, there are only 1-2 opportunities for assessment, resulting in students failing to pass the exam due to mistakes, and they need to wait a lot of time to continue the assessment. Fifth, although CPR training has greatly increased in recent years, it is very difficult to count the pass rate of students, the number of students who have passed, the age distribution and the regional distribution, which is not conducive to the promotion and popularization of CPR training. Therefore, it is very important to launch an intelligent CPR training system with good uniformity and standardization, high accuracy, low cost and strong practicality. Summary of the invention
[0003] The purpose of the present invention is to provide an intelligent cardiopulmonary resuscitation training system to improve the training quality of cardiopulmonary resuscitation training.
[0004] To achieve the above object, the present invention provides the following solutions:
[0005] An intelligent cardiopulmonary resuscitation training system, the system comprising: a human-computer interaction device, a host, a camera device and a simulated person;
[0006] The human-computer interaction device is connected to the host, and is used to receive a trainee's autonomous practice request or autonomous assessment request, and obtain and display cardiopulmonary resuscitation training content or cardiopulmonary resuscitation assessment content from the host according to the request;
[0007] The trainee performs first aid operations on the simulator according to the cardiopulmonary resuscitation training content or cardiopulmonary resuscitation assessment content; the simulator is used to collect the first aid operation data of the trainee and transmit the first aid operation data to the host; the camera device is used to capture the video image of the trainee performing the first aid operation and transmit the video image to the host;
[0008] The host is used to judge the accuracy of the first aid operation according to the first aid operation data when the trainee is performing independent practice, judge the correctness of the action after performing action recognition according to the video signal, and score the first aid operation according to the first aid operation data when the trainee is performing independent assessment, and score the action after performing action recognition according to the video signal, so as to finally obtain the assessment score.
[0009] Optionally, the host includes: a practice unit, an assessment unit and a processing unit;
[0010] The practice unit is used to store cardiopulmonary resuscitation teaching videos, and transmit the cardiopulmonary resuscitation teaching videos to the human-computer interaction device for display upon receiving an autonomous practice request;
[0011] The processing unit is used to judge the accuracy of the first aid operation according to the first aid operation data, judge the correctness of the action after performing action recognition according to the video signal, and display an error operation prompt on the human-computer interaction device when it is judged that the first aid operation is wrong;
[0012] The assessment unit is used to store the autonomous assessment content, and transmit the autonomous assessment content to the human-computer interaction device for display when receiving the autonomous assessment request;
[0013] The processing unit is further used to score the first aid operation according to the first aid operation data, score the action after performing action recognition according to the video signal, and finally obtain an assessment score.
[0014] Optionally, the processing unit includes:
[0015] A skeleton point extraction subunit is used to extract the key points of the human skeleton of the trainee in the video image by using the OpenPose method, and connect the key points of the human skeleton to form a human skeleton model;
[0016] An action recognition subunit, used for performing action recognition using a spatiotemporal graph convolutional neural network based on the human skeleton model;
[0017] The judging subunit is used to identify the correctness of the action according to the preset action judgment standard.
[0018] Optionally, the spatiotemporal graph convolutional neural network includes 9 layers of ST-GCN modules connected in sequence, the last layer of the first three layers of ST-GCN modules outputs 64-dimensional features, the last layer of the middle three layers of ST-GCN modules outputs 128-dimensional features, and the last layer of the last three layers of ST-GCN modules outputs 256-dimensional features. The time kernel Γ of each layer of ST-GCN modules is 9, and a resnet mechanism is set for each layer of ST-GCN modules.
[0019] Optionally, the process of training the spatiotemporal graph convolutional neural network is:
[0020] Randomly extract clips from historical video images captured by the camera, and use the OpenPose method to extract the student's human skeleton key points from the randomly extracted clips to obtain a human skeleton model sequence;
[0021] Performing affine transformation on the human skeleton model sequence;
[0022] The convolution formula of the spatiotemporal graph convolutional neural network is determined as In the formula, f out (v ti ) is the convolution output, f in (v tj ) is the convolution input, v tj v ti Coordinates of directly adjacent nodes, l ti (v tj ) is v tj The feature set where the node is located, Z ti (v tj ) is the standardized item, Z ti (v tj )=|{v tk |l ti (v tk )=l ti (v tj )}|,v tk is the normalized node coordinate;
[0023] Determine the weight matrix of the spatiotemporal graph convolutional neural network as In the formula, r i 、r j Respectively represent v ti Node, v tj The average distance from the node to the centroid of the skeleton graph;
[0024] According to the human skeleton model sequence after affine transformation, the spatiotemporal graph convolutional neural network is trained based on the convolution formula and weight matrix.
[0025] Optionally, the host further includes: a management unit;
[0026] The management unit is used for data statistics, personnel management, teaching video management, first aid process management, and assessment record management. It can add, delete, modify, and check data, and mark the location information of the students on the map; the statistical data includes student information, age group pass rate, class pass rate, student pass rate, and single action error rate.
[0027] Optionally, the camera device is a binocular camera.
[0028] Optionally, the simulator includes: a single chip microcomputer and a plurality of sensors;
[0029] The single chip microcomputer is connected to a plurality of sensors and a host computer respectively, and the single chip microcomputer is used for standardizing the data measured by the plurality of sensors and transmitting the data to the host computer.
[0030] Optionally, the plurality of sensors include a resistive pressure-sensitive sensor, a pressure sensor, a gas flow sensor and a photoelectric gate sensor.
[0031] Optionally, the single chip microcomputer is an STM32 single chip microcomputer;
[0032] Multiple sensors are connected to the STM32 microcontroller via DuPont cables;
[0033] The STM32 single chip microcomputer is used to standardize the data measured by multiple sensors and transmit them to the host through the Bluetooth or WiFi module.
[0034] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:
[0035] The present invention discloses an intelligent cardiopulmonary resuscitation training system. Trainees can choose to practice or assess themselves on a human-computer interaction device, and after obtaining the corresponding content, they can perform first aid operations on a simulator. The simulator collects the first aid operation data of the trainees, and the camera device captures the video images of the trainees performing the first aid operations. The host can judge and score the accuracy of the first aid operations, and judge and score the correctness of the first aid actions. The present invention is an online training system that is not limited by place and time. Trainees can practice multiple times, and the teaching and assessment standards are unified and standardized, which can effectively improve the training quality of cardiopulmonary resuscitation training. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0037] Figure 1 A schematic diagram of the structure of an intelligent cardiopulmonary resuscitation training system provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0039] The purpose of the present invention is to provide an intelligent cardiopulmonary resuscitation training system to improve the training quality of cardiopulmonary resuscitation training.
[0040] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0041] The embodiment of the present invention provides an intelligent cardiopulmonary resuscitation training system. Figure 1 As shown, it includes: a human-computer interaction device, a host, a camera device and a simulator. The human-computer interaction device is connected to the host, and is used to receive the trainee's autonomous practice request or autonomous assessment request, and obtain the cardiopulmonary resuscitation training content or cardiopulmonary resuscitation assessment content from the host and display it according to the request. The trainee performs first aid operations on the simulator according to the cardiopulmonary resuscitation training content or cardiopulmonary resuscitation assessment content; the simulator is used to collect the trainee's first aid operation data and transmit the first aid operation data to the host; the camera device is used to shoot the video image of the trainee performing the first aid operation, and transmit the video image to the host. The host is used to judge the accuracy of the first aid operation according to the first aid operation data when the trainee performs autonomous practice, judge the correctness of the action after performing action recognition according to the video signal, and score the first aid operation according to the first aid operation data when the trainee performs autonomous assessment, and score the action after performing action recognition according to the video signal, and finally obtain the assessment score.
[0042] Exemplarily, the host includes: a practice unit, an assessment unit and a processing unit. The practice unit is used to store cardiopulmonary resuscitation teaching videos, and transmit the cardiopulmonary resuscitation teaching videos to the human-computer interaction device for display when receiving an autonomous practice request. The processing unit is used to judge the accuracy of the first aid operation based on the first aid operation data, judge the correctness of the action after performing action recognition based on the video signal, and display an error operation prompt on the human-computer interaction device when an error occurs in the first aid operation. The assessment unit is used to store autonomous assessment content, and transmit the autonomous assessment content to the human-computer interaction device for display when receiving an autonomous assessment request. The processing unit is also used to score the first aid operation based on the first aid operation data, score the action after performing action recognition based on the video signal, and finally obtain an assessment score.
[0043] The host is used to process data at high speed. First, after the user operates on the front-end page, the generated data signal will be processed by the host and then transmitted to the front-end to execute the corresponding command; secondly, the host is also used to receive the signal transmitted by the camera unit and the artificial human unit, and transmit it to the display screen (human-computer interaction device) for display after data processing.
[0044] The processing unit extracts the relative position relationship between the key points of the human body in the three-dimensional space in each frame of the binocular camera video, and performs human action recognition based on the spatial position changes of the key points of the human body in the video. The specific steps include human key point extraction and action recognition. The processing unit specifically includes: a skeleton point extraction subunit, which is used to extract the 18 human skeleton key points of the trainee in the video image using the OpenPose method, and connect the human skeleton key points to form a human skeleton model. An action recognition subunit is used to perform action recognition based on the human skeleton model using a spatiotemporal graph convolutional neural network. A judgment subunit is used to identify the correctness of the action according to the preset action judgment standard. The judgment standard for action recognition is formulated by professional medical emergency personnel and the corresponding scoring standard is given, and the host judges the correctness of the action according to the standard.
[0045] What makes OpenPose different from traditional recognition algorithms is that the study uses bottom-up human key point information recognition, that is, finding various parts of the human body and then connecting the parts through part affinity fields. The human key point information output by OpenPose recognition can be regarded as a topological structure, so the spatial-temporal graph neural network (STGCN) is finally used to extract the spatial features of the topological graph. It combines the temporal convolutional neural network (TCN) and the graph neural network (GCN) to process graph structure data with time series relationships. It can perform convolution and extract features in both time and space dimensions to derive a model.
[0046] The specific steps of the action recognition part are as follows: First, since different nodes in the same layer of GCN share the weight matrix, in order to keep the scale of the input data consistent, batch normalization is performed before inputting the data; secondly, ST-GCN consists of 9 layers of ST-GCN modules, the first three layers output 64-dimensional features, the middle 3 layers output 128-dimensional features, and the last 3 layers output 256-dimensional features. The temporal kernel Γ of each layer is 9, and each layer has a resnet mechanism; then, in order to avoid overfitting, dropout = 0.5 is added to each layer, and the pooling stride of the 4th and 7th layers is 2; finally, the graph embedding is converted into an n-dimensional vector and sent to the classic classifier softmax for graph classification.
[0047] During the training process, the sampling function and weight function are redefined and used in the convolution process. The convolution formula is as follows:
[0048]
[0049] Among them, v tj v ti Coordinates of directly adjacent nodes, l ti (v tj ) is the feature set of this node, Z ti (v tj ) is the standardized item, and the formula of the standardized item is:
[0050] z ti (v tj )=|{v tk |l ti (v tk )=l ti (v tj )}|
[0051] At the same time, during the training process, weight values are used to measure the characteristics of different actions to achieve the purpose of action recognition. Specifically, it is defined according to the centripetal and eccentric relationship of the joint points. r represents the average distance from the node to the center of gravity of the skeleton graph. At this time, for the convolution operation of a key point of the human body, its weight matrix includes three weight vectors:
[0052]
[0053] The node's 1 neighborhood is divided into three subsets. The first subset connects the neighbor nodes that are farther away from the entire skeleton than the root node in space, the second subset connects the neighbor nodes that are closer to the center, and the third subset is the root node itself, representing the motion characteristics of centrifugal motion, centripetal motion, and stillness, respectively.
[0054] Finally, in order to prevent overfitting, the training data was optimized: segments were randomly selected from the original sequence, and the OpenPose method was used to extract the student's human skeleton key points in the randomly selected segments to obtain a human skeleton model sequence, and then an affine transformation was performed on the skeleton sequence (simulating camera movement).
[0055] In one example, the host also includes: a management unit. The management unit is used for data statistics, personnel management, teaching video management, first aid process management, and assessment record management. It can add, delete, modify, and check data, and mark the location information of the trainees on the map; the statistical data includes trainee information, age group pass rate, class pass rate, trainee pass rate, and single action error rate. Specifically, for personnel management, each trainee's personal information will be displayed on the personnel management tab; for teaching video management, you can upload, edit, and delete teaching videos, and modify the text prompts of the teaching; for first aid process management, you can edit the module information in cardiopulmonary resuscitation; for assessment record management, you can view and edit the assessment records of each trainee. Personal information includes basic personal information, learning progress, practice records, and assessment records. The information of the trainees will be saved to the management end.
[0056] Correspondingly, the front-end page displayed on the human-computer interaction device contains three modules, namely the student side, the assessment side and the management side. The web pages of these three modules are all built using the B / S architecture. The student side can watch teaching videos and practice independently on the display. During practice, the teaching video, text-based operation prompts and abnormal operation records will be displayed, and the student can make appropriate adjustments according to the prompts; the assessment side includes self-assessment and score distribution functions. The entire assessment process is monitored by binocular cameras and simulated humans, and the assessment results are calculated by the host, all in an unmanned monitoring and unmanned management mode; after the assessment, the score will be announced, and if a certain score is met, the assessment will be considered passed. The assessment results will be saved in personal information, and the operation error information during the assessment process will also be announced. Personal assessment information will also be saved to the management side. The management end is used for data statistics, including statistics, personnel management, teaching video management, first aid process management, and assessment record management. Administrators can add, delete, modify, and query data, and the age group pass rate, class pass rate, student pass rate, and single action error rate will also be displayed visually. At the same time, the student's location information will also be displayed on the map, making it easy to view the geographical distribution of people learning CPR across the country.
[0057] In one example, the camera device is a binocular camera, which is used for students' independent practice and self-assessment. The binocular camera detects the students and their positions, verifies whether the students' fine hand movements are standardized, and transmits the video signal obtained by the binocular camera to the host for processing. The specific steps of the binocular camera include camera calibration and image acquisition. Camera calibration is used to determine the internal and external parameters of the camera to avoid the environment and camera technology affecting the effect; secondly, the image is acquired as the test data.
[0058] The simulator includes: a single chip microcomputer and a plurality of sensors. The single chip microcomputer is connected to the plurality of sensors and a host computer respectively, and the single chip microcomputer is used for standardizing the data measured by the plurality of sensors and transmitting the data to the host computer.
[0059] The human simulator unit is also a key part and plays a decisive role. It can make up for the fact that the binocular camera is only used for "visual" detection. It contains a large number of sensors, each of which is connected to the STM32 microcontroller. After the microcontroller performs preliminary processing on the signal obtained by the sensor, the system host that transmits the data through the Bluetooth or WiFi module is used as part of the basis for the comprehensive judgment of the entire emergency operation. The accuracy of the trainee's simulated operation will be judged, including the detection of hand touch, touch position, artificial respiration, compression frequency and compression degree. The sensors used are resistive pressure sensitive sensors, pressure sensors, gas flow sensors, and photoelectric gate sensors. The generated data (data returned by the sensor after preliminary processing by the microcontroller) is finally uploaded to the host through Bluetooth to score the trainee assessment according to the judgment and scoring standards formulated by professional medical emergency personnel.
[0060] The entire working process of the intelligent cardiopulmonary resuscitation training system is as follows: after the user logs in to the student end, he first registers an account, then enters the account password to log in to the homepage, and clicks on the video cover displayed on the homepage to enter the video interface for learning; click on the practice in the upper right corner of the homepage to open the practice interface. The student only needs to operate on the simulator in front of the binocular camera. If there is an error in the operation, the interface will display the operation prompt, and the student can restart the practice according to the operation prompt. Among them, the practice process of each module will record the analysis of the host data. When the student enters the assessment end, click Start to start the assessment. The assessment is unsupervised and unprompted throughout the process. The entire process is monitored by the binocular camera and the simulator, and its data signal will be transmitted to the host for processing. When a mistake occurs, the host will record the mistake in the score, and will not pop up a prompt like the practice. When the assessment is over, the system will announce the score and the operation error information during the assessment process. The assessment can be passed when a certain score is reached. In addition, the administrator can also log in to the management end to view the student's learning progress, and can also optimize the system based on statistical charts.
[0061] The positive effects of this system:
[0062] (1) Online training allows students to learn and train online at any time, and can remind themselves to standardize their movements based on mistakes made during practice.
[0063] (2) Training makes students’ learning and operation more unified and standardized, and is less likely to result in uneven quality.
[0064] (3) The setting of multiple assessments gives students more opportunities to practice, so that they do not have to wait a long time to verify their learning results due to a mistake in an assessment.
[0065] (4) Online training will allow more people to participate in CPR learning, which will benefit the promotion and popularization of CPR, allowing more patients with cardiac arrest to receive timely and effective assistance and reduce the chance of sudden death.
[0066] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.
[0067] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of the present invention. At the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.
Claims
1. An intelligent cardiopulmonary resuscitation training system, It is characterized in that The system comprises: a human-computer interaction device, a host, a camera device and a simulated person; The human-computer interaction device is connected to the host, and is used to receive a trainee's autonomous practice request or autonomous assessment request, and obtain and display cardiopulmonary resuscitation training content or cardiopulmonary resuscitation assessment content from the host according to the request; The trainee performs first aid operations on the simulator according to the cardiopulmonary resuscitation training content or cardiopulmonary resuscitation assessment content; the simulator is used to collect the first aid operation data of the trainee and transmit the first aid operation data to the host; the camera device is used to capture the video image of the trainee performing the first aid operation and transmit the video image to the host; The host is used to judge the accuracy of the first aid operation according to the first aid operation data when the trainee is practicing independently, judge the correctness of the action after performing action recognition according to the video signal, and score the first aid operation according to the first aid operation data when the trainee is performing self-assessment, and score the action after performing action recognition according to the video signal, and finally obtain the assessment score; The multiple sensors in the simulator include a resistive pressure sensor, a pressure sensor, a gas flow sensor, and a photoelectric gate sensor; The host comprises: a processing unit; the processing unit is used to judge the accuracy of the first aid operation according to the first aid operation data, judge the correctness of the action after performing action recognition according to the video signal, and display an error operation prompt on the human-computer interaction device when judging that an error occurs in the first aid operation; the processing unit is also used to score the first aid operation according to the first aid operation data, score the action after performing action recognition according to the video signal, and finally obtain an assessment score; wherein, the accuracy of the trainee's simulated operation is judged, including detecting whether the hand touch, touch position, artificial respiration and pressing frequency meet the requirements; The processing unit includes: an action recognition subunit, which is used to perform action recognition using a spatiotemporal graph convolutional neural network based on a human skeleton model; The spatiotemporal graph convolutional neural network includes 9 layers of ST-GCN modules connected in sequence, the last layer of the first three layers of ST-GCN modules outputs 64-dimensional features, the last layer of the middle three layers of ST-GCN modules outputs 128-dimensional features, and the last layer of the last three layers of ST-GCN modules outputs 256-dimensional features. The time kernel Γ of each layer of ST-GCN modules is 9, and each layer of ST-GCN modules is set with a resnet mechanism; The process of training the spatiotemporal graph convolutional neural network is: Randomly extract clips from historical video images captured by the camera, and use the OpenPose method to extract the student's human skeleton key points from the randomly extracted clips to obtain a human skeleton model sequence; Performing affine transformation on the human skeleton model sequence; The convolution formula of the spatiotemporal graph convolutional neural network is determined as In the formula, f out (v ti ) is the convolution output, f in (v ij ) is the convolution input, vtj is the node coordinates directly adjacent to vti, lti(vtj) is the feature set where the vtj node is located, Zti(vtj) is the normalization term, and Z ti (v tj )=|{v tk |l ti (v tk )=l ti (v tj )}|, vtk is the standardized node coordinate; Determine the weight matrix of the spatiotemporal graph convolutional neural network as In the formula, r i 、r j Respectively represent v ti Node, v tj The average distance from the node to the centroid of the skeleton graph; According to the human skeleton model sequence after affine transformation, the spatiotemporal graph convolutional neural network is trained based on the convolution formula and weight matrix; The processing unit extracts the relative position relationship between the key points of the human body in the three-dimensional space in each frame of the binocular camera video, and performs human action recognition according to the spatial position changes of the key points of the human body in the video; the specific steps include human key point extraction and action recognition; the processing unit specifically includes: a skeleton point extraction subunit, which is used to extract the 18 human skeleton key points of the trainee in the video image using the OpenPose method, and connect the human skeleton key points to form a human skeleton model; an action recognition subunit, which is used to perform action recognition based on the human skeleton model using a spatiotemporal graph convolutional neural network; a judgment subunit, which is used to identify the correctness of the action according to the preset action judgment standard; the judgment standard for action recognition is formulated by professional medical emergency personnel and the corresponding scoring standard is given, and the host judges the correctness of the action according to the standard; Among them, OpenPose uses bottom-up human key point information recognition, that is, finding out the various parts of the human body, and then connecting the parts through part affinity fields (PartAffinityFields); the human key point information output by OpenPose recognition can be regarded as a topological structure, so the spatial-temporal graph neural network (STGCN) is finally used to extract the spatial features of the topological graph. It combines the temporal convolutional neural network (TCN) and the graph neural network (GCN) to process graph structure data with time series relationships. It can perform convolution and extract features in both time and space dimensions to obtain a model; Management unit: The management unit is used for data statistics, personnel management, teaching video management, first aid process management, and assessment record management. It can add, delete, modify, and check data, and mark the location information of students on the map; the statistical data includes student information, age group pass rate, class pass rate, student pass rate, and single action error rate; specifically, for personnel management, the personal information of each student will be displayed in the personnel management tab; for teaching video management, you can upload, edit, and delete teaching videos, and modify the text prompts of teaching; for first aid process management, you can edit the module information in cardiopulmonary resuscitation; for assessment record management, you can view and edit the assessment records of each student; personal information includes personal basic information, learning progress, practice records, and assessment records; all student information will be saved to the management end; The front-end page displayed on the human-computer interaction device contains three modules, namely the student side, the assessment side and the management side. The web pages of these three modules are all built with B / S architecture. The student side can watch teaching videos on the display and practice independently. During the practice, the teaching video, text version of operation prompts and operation abnormality records will be displayed, and the student can make appropriate adjustments according to the prompts. The assessment side includes self-assessment and score distribution functions. The entire assessment process is monitored by binocular cameras and simulators. The assessment results are calculated by the host, and all adopt unmanned monitoring and unmanned management mode. After the assessment, the score will be announced. If a certain score is met, the assessment is considered to have passed. The assessment results will be saved in personal information, and the operation error information during the assessment process will also be announced; personal assessment information will also be saved to the management side; the management side is used for data statistics, including statistics, personnel management, teaching video management, first aid process management, and assessment record management. The administrator can add, delete, modify, and check the data, and the age group pass rate, class pass rate, student pass rate, and single action error rate will also be visualized. At the same time, the location information of the students will also be displayed on the map, which is convenient for viewing the geographical distribution of CPR learners across the country. The simulator includes: a single chip microcomputer and a plurality of sensors; the single chip microcomputer is connected to the plurality of sensors and the host computer respectively, and the single chip microcomputer is used for standardizing the data measured by the plurality of sensors and transmitting the data to the host computer; The human simulator unit is also a key part and plays a decisive role. It can make up for the fact that the binocular camera is only used for "visual" detection. It contains a large number of sensors, each of which is connected to the STM32 microcontroller. After the microcontroller performs preliminary processing on the signal obtained by the sensor, the system host that transmits the data through the Bluetooth or WiFi module is used as part of the basis for the comprehensive judgment of the entire emergency operation. The accuracy of the trainee's simulated operation will be judged, including the detection of hand touch, touch position, artificial respiration, compression frequency and compression degree. The sensors used are resistive pressure sensitive sensors, pressure sensors, gas flow sensors, and photoelectric gate sensors. The generated data (the data returned by the sensor after preliminary processing by the microcontroller) is finally uploaded to the host via Bluetooth to score the trainee's assessment according to the judgment and scoring standards formulated by professional medical emergency personnel. The entire working process of the intelligent cardiopulmonary resuscitation training system is as follows: after the user logs in to the student end, he first registers an account, then enters the account password to log in to the homepage, and clicks on the video cover displayed on the homepage to enter the video interface for learning; click on practice in the upper right corner of the homepage to open the practice interface, and the student only needs to operate on the simulator in front of the binocular camera. If there is an error in the operation, the interface will display the operation prompt, and the student can restart the practice according to the operation prompt. Among them, the practice process of each module will record the analysis of the host data; when the student enters the assessment end, click start to start the assessment; the assessment is unsupervised and unprompted throughout the process. The entire process is monitored by the binocular camera and the simulator, and its data signal will be transmitted to the host for processing; when an error occurs, the host will record the error in the score, and will not pop up a prompt like the practice; when the assessment is over, the system will announce the score and the operation error information during the assessment process, and the assessment can be passed when a certain score is reached; in addition, the administrator can also log in to the management end to view the student's learning progress, and can also optimize the system based on statistical charts.
2. The intelligent cardiopulmonary resuscitation training system according to claim 1, It is characterized in that The host comprises: a practice unit and an assessment unit; The practice unit is used to store cardiopulmonary resuscitation teaching videos, and transmit the cardiopulmonary resuscitation teaching videos to the human-computer interaction device for display upon receiving an autonomous practice request; The assessment unit is used to store the autonomous assessment content, and transmit the autonomous assessment content to the human-computer interaction device for display when receiving an autonomous assessment request.
3. The intelligent cardiopulmonary resuscitation training system according to claim 2, It is characterized in that The processing unit comprises: A skeleton point extraction subunit is used to extract the key points of the human skeleton of the trainee in the video image by using the OpenPose method, and connect the key points of the human skeleton to form a human skeleton model; The judging subunit is used to identify the correctness of the action according to the preset action judgment standard.
4. The intelligent cardiopulmonary resuscitation training system according to claim 2, It is characterized in that The host also includes: a management unit; The management unit is used for data statistics, personnel management, teaching video management, first aid process management, and assessment record management. It can add, delete, modify, and check data, and mark the location information of the students on the map; the statistical data includes student information, age group pass rate, class pass rate, student pass rate, and single action error rate.
5. The intelligent cardiopulmonary resuscitation training system according to claim 1, It is characterized in that The camera device is a binocular camera.
6. The intelligent cardiopulmonary resuscitation training system according to claim 1, It is characterized in that The simulator comprises: a single chip computer and a plurality of sensors; The single chip microcomputer is connected to a plurality of sensors and a host computer respectively, and the single chip microcomputer is used for standardizing the data measured by the plurality of sensors and transmitting the data to the host computer.
7. The intelligent cardiopulmonary resuscitation training system according to claim 6, It is characterized in that The single chip microcomputer is an STM32 single chip microcomputer; Multiple sensors are connected to the STM32 microcontroller via DuPont cables; The STM32 single chip microcomputer is used to standardize the data measured by multiple sensors and transmit them to the host through the Bluetooth or WiFi module.
Citation Information
Patent Citations
System and method of digitalized emergency skill simulated training
CN103106830A
Visual virtual surge movement protective training device
CN106693338A
Intelligent precise traditional Chinese medicine skill training assessment system
CN107369352A
time sequence behavior detection method based on 3D human body key points
CN109919122A
Electrical equipment insulation test simulation training holographic teacher system based on ring screen interaction
CN110310530A
