Emergency intelligent alarm method and system for emergency treatment

Abstract

The invention provides an emergency intelligent alarm method and system for emergency treatment, and the method comprises the steps: obtaining current video data and current audio data collected by an intelligent device of an emergency medical worker, and recognizing violence trend information corresponding to the medical worker, and the semantic information and keyword segment information of the current audio data, analyzing the violence probability of each violence event type corresponding to the medical personnel; generating current initial alarm information based on the target violence probability greater than the violence probability threshold and the target violence event type, or generating the current initial alarm information through a keyword segment alarm strategy based on the keyword segment information; and collecting current position information of the medical personnel, and generating emergency alarm information of the medical personnel based on the current position information and the current initial alarm information. By adopting the scheme, the emergency alarm efficiency of medical staff in the face of violent events can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of catalysts and hydrogen production equipment, and in particular to an emergency intelligent alarm method and system for emergency treatment. Background Art

[0002] As a department where violent incidents frequently occur, the emergency department often causes great physical and mental interference to medical staff due to the sudden nature of the incidents. Therefore, how to ensure the personal safety of medical staff and maintain medical order is the current research focus.

[0003] The traditional way to protect against emergency violent incidents is to add security personnel and set up a one-touch alarm button at a location that is easily accessible to medical staff in each department. However, the sudden outbreak areas of violent incidents and the sudden outbreaks of violent incidents are often uncertain. Manual protection and alarm protection at fixed locations are still not comprehensive enough to protect the safety of medical staff, resulting in low efficiency of emergency alarms for medical staff in the face of violent incidents. Summary of the invention

[0004] The main purpose of the present invention is to provide an emergency intelligent alarm method and system for emergency treatment, aiming to solve the problem of low efficiency of emergency alarm in the prior art, because the sudden areas of violent incidents and the sudden incidents of violent incidents are often uncertain, and manual protection and alarm protection at fixed positions are still not comprehensive enough for the safety protection of medical staff.

[0005] To achieve the above object, the present invention provides an emergency intelligent alarm method for emergency treatment, the method comprising:

[0006] Acquire current video data collected by the smart device of the emergency medical staff and current audio data collected by the smart device, and identify the violence trend information corresponding to the medical staff through a behavior trend recognition strategy based on the current video data;

[0007] Based on the current audio data, the semantic information of the current audio data and the key segment information of the current audio data are identified by a semantic recognition strategy, and when the key segment information of the current audio data is 0, based on the violence trend information and the semantic information, the types of violence events corresponding to the medical staff and the violence probability of each type of violence event are analyzed by a violence trend analysis strategy;

[0008] When there is a violence probability greater than a violence probability threshold, based on the target violence probability greater than the violence probability threshold and the target violence event type corresponding to the target violence probability, generate current initial alarm information, and when the keyword segment information of the current audio data is not 0, based on the keyword segment information, generate current initial alarm information through a keyword segment alarm strategy;

[0009] The current location information of the medical staff is collected, and based on the current location information and the current initial alarm information, emergency alarm information of the medical staff is generated.

[0010] Optionally, the identifying, based on the current video data, the violence trend information corresponding to the medical staff through a behavior trend identification strategy includes:

[0011] Splitting the current video data into image data of each frame, and identifying the facial range of each person and the limb range of each person in each image data by an image edge recognition algorithm;

[0012] The body movement information of the body image of each person corresponding to the body range of each person is identified through the action recognition network, and the facial expression information of the facial image of each person corresponding to the facial range of each person is identified through the expression recognition network;

[0013] According to the time sequence of each of the image data, the body movement information of each person's image data is sorted and processed to obtain the body movement change information of each person, and according to the time sequence of each of the image data, the facial expression information of each person's image data is sorted and processed to obtain the facial expression change information of each person;

[0014] Based on the body movement change information of each person and the facial expression change information of each person, predict the body movement change trend of each person and the facial expression change trend of each person, and based on the body movement change trend of each person and the facial expression change trend of each person, identify the sub-violence trend information of each person through the violence trend identification strategy;

[0015] The sub-violence trend information of all personnel is used as the violence trend information corresponding to the medical staff.

[0016] Optionally, the identifying semantic information of the current audio data and key word segment information of the current audio data by using a semantic recognition strategy based on the current audio data includes:

[0017] The current audio data is converted into a current audio text through a text recognition network, and a keyword segment search is performed in the current audio text to identify the keyword segment information in the current audio text;

[0018] The text semantic content of the current audio text is identified through a semantic recognition network, and the text semantic content is used as the semantic information of the current audio data.

[0019] Optionally, the analyzing, based on the violence trend information and the semantic information, the types of violence events corresponding to the medical staff and the violence probability of each type of violence event through a violence trend analysis strategy includes:

[0020] Based on the semantic information, identifying each violent semantic content in the semantic information through a semantic screening network, and identifying the recurrence frequency of each violent word corresponding to each violent semantic content in the current audio text;

[0021] Identify the type of violent event corresponding to each violent semantic content, and identify the initial violent probability of each violent event type based on the recurrence frequency of each violent word corresponding to each violent semantic content;

[0022] Based on the sub-violence trend information of each person, the violence information corresponding to each person and the violence risk corresponding to each violence information are identified, and the initial violence probability of each type of violence event is adjusted to obtain the violence probability of each type of violence event.

[0023] Optionally, the generating current initial alarm information based on the target violence probability greater than the violence probability threshold and the target violence event type corresponding to the target violence probability includes:

[0024] Based on the target violent event type, identifying the violent information corresponding to the target violent event type, and based on the target violence probability of the target violent event type and the violence risk level of the violent information corresponding to each person, screening each target violent information and the target person corresponding to each target violent information;

[0025] A preset alarm information template is obtained, and each target violence information, a target person corresponding to each target violence information, and a violence risk level of each target violence information are filled into the alarm information template to obtain the current initial alarm information.

[0026] Optionally, the generating of current initial alarm information based on the keyword segment information and through a keyword segment alarm strategy includes:

[0027] Based on the keyword segment information, identifying each keyword contained in the keyword segment information, and querying the alarm information corresponding to each keyword in the database;

[0028] The alarm information corresponding to all keywords is used as the current initial alarm information.

[0029] In addition, to achieve the above-mentioned purpose, the present invention also provides an emergency intelligent alarm system for emergency treatment, the emergency intelligent alarm system comprising:

[0030] An acquisition module, used to acquire current video data collected by the smart device of the emergency medical staff and current audio data collected by the smart device, and identify the violence trend information corresponding to the medical staff through a behavior trend recognition strategy based on the current video data;

[0031] an identification module, configured to identify semantic information of the current audio data and key segment information of the current audio data through a semantic identification strategy based on the current audio data, and to analyze the types of violence events corresponding to the medical staff and the violence probability of each type of violence event through a violence trend analysis strategy based on the violence trend information and the semantic information when the key segment information of the current audio data is 0;

[0032] A generating module, configured to generate current initial alarm information based on a target violence probability greater than the violence probability threshold and a target violence event type corresponding to the target violence probability when there is a violence probability greater than the violence probability threshold, and to generate current initial alarm information based on the keyword segment information and through a keyword segment alarm strategy when the keyword segment information of the current audio data is not 0;

[0033] The alarm module is used to collect the current location information of the medical staff and generate emergency alarm information of the medical staff based on the current location information and the current initial alarm information.

[0034] Optionally, the acquisition module is specifically used to:

[0035] Splitting the current video data into image data of each frame, and identifying the facial range of each person and the limb range of each person in each image data by an image edge recognition algorithm;

[0036] The body movement information of the body image of each person corresponding to the body range of each person is identified through the action recognition network, and the facial expression information of the facial image of each person corresponding to the facial range of each person is identified through the expression recognition network;

[0037] According to the time sequence of each of the image data, the body movement information of each person's image data is sorted and processed to obtain the body movement change information of each person, and according to the time sequence of each of the image data, the facial expression information of each person's image data is sorted and processed to obtain the facial expression change information of each person;

[0038] Based on the body movement change information of each person and the facial expression change information of each person, predict the body movement change trend of each person and the facial expression change trend of each person, and based on the body movement change trend of each person and the facial expression change trend of each person, identify the sub-violence trend information of each person through the violence trend identification strategy;

[0039] The sub-violence trend information of all personnel is used as the violence trend information corresponding to the medical staff.

[0040] Optionally, the identification module is specifically used to:

[0041] The current audio data is converted into a current audio text through a text recognition network, and a keyword segment search is performed in the current audio text to identify the keyword segment information in the current audio text;

[0042] The text semantic content of the current audio text is identified through a semantic recognition network, and the text semantic content is used as the semantic information of the current audio data.

[0043] Optionally, the identification module is specifically used to:

[0044] Based on the semantic information, identifying each violent semantic content in the semantic information through a semantic screening network, and identifying the recurrence frequency of each violent word corresponding to each violent semantic content in the current audio text;

[0045] Identify the type of violent event corresponding to each violent semantic content, and identify the initial violent probability of each violent event type based on the recurrence frequency of each violent word corresponding to each violent semantic content;

[0046] Based on the sub-violence trend information of each person, the violence information corresponding to each person and the violence risk corresponding to each violence information are identified, and the initial violence probability of each type of violence event is adjusted to obtain the violence probability of each type of violence event.

[0047] Optionally, the generating module is specifically used for:

[0048] Based on the target violent event type, identifying the violent information corresponding to the target violent event type, and based on the target violence probability of the target violent event type and the violence risk level of the violent information corresponding to each person, screening each target violent information and the target person corresponding to each target violent information;

[0049] A preset alarm information template is obtained, and each target violence information, a target person corresponding to each target violence information, and a violence risk level of each target violence information are filled into the alarm information template to obtain the current initial alarm information.

[0050] Optionally, the generating module is specifically used for:

[0051] Based on the keyword segment information, identifying each keyword contained in the keyword segment information, and querying the alarm information corresponding to each keyword in the database;

[0052] The alarm information corresponding to all keywords is used as the current initial alarm information.

[0053] In a third aspect, the present application provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of any one of the methods in the first aspect are implemented.

[0054] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of any one of the methods in the first aspect are implemented.

[0055] In a fifth aspect, the present application provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of any one of the methods in the first aspect are implemented.

[0056] The present invention provides an emergency intelligent alarm method and system for emergency department, the method comprising: obtaining current video data collected by the intelligent device of the emergency medical staff and current audio data collected by the intelligent device, and based on the current video data, identifying the violence trend information corresponding to the medical staff through a behavior trend recognition strategy; based on the current audio data, identifying the semantic information of the current audio data and the key segment information of the current audio data through a semantic recognition strategy, and when the key segment information of the current audio data is 0, based on the violence trend information and the semantic information, identifying the violence trend information through a violence trend classification strategy; Analysis strategy, analyze the types of violent events corresponding to the medical staff and the violence probability of each type of violent event; when there is a violence probability greater than the violence probability threshold, based on the target violence probability greater than the violence probability threshold and the target violent event type corresponding to the target violence probability, generate the current initial alarm information, and when the keyword segment information of the current audio data is not 0, based on the keyword segment information, generate the current initial alarm information through the keyword segment alarm strategy; collect the current location information of the medical staff, and generate the emergency alarm information of the medical staff based on the current location information and the current initial alarm information. This solution uses the current video data and current audio data collected in real time by smart devices (such as camera equipment, smart watches, smart bracelets, portable image acquisition equipment) set on the body surface of the medical staff to comprehensively analyze the violence trend information corresponding to the medical staff, the semantic information corresponding to the current audio data, and the keyword segment information. Thus, it is judged whether there is a risk of violence through dual channels. First, the violence probability of medical staff facing various types of violent incidents is comprehensively judged through violence trend information and semantic information. In this way, when medical staff cannot report the alarm in time, when it is judged that the violence probability is greater than the violence probability threshold, the initial alarm information is automatically generated, or when the medical staff speaks the keyword segment information, the alarm information corresponding to the keyword in the keyword segment is directly obtained through the database, and then the location information of the positioning intelligent device is used to generate an emergency alarm message sent to each target police officer, thereby ensuring that medical staff can actively voice alarm or intelligently analyze the alarm in a passive situation, greatly reducing the risk of violence caused by the inconvenient location of the one-button alarm device or the untimely alarm causing harm to the physical and mental safety of medical staff. Finally, this scheme predicts the risk of violence of medical staff efficiently and timely through positioning and dual-channel alarm information recognition, thereby effectively improving the efficiency of emergency alarms for medical staff in the face of violent incidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0058] Figure 1 is a flow chart of an emergency intelligent alarm method for emergency treatment provided by an embodiment of the present invention;

[0059] Figure 2 is a structural diagram of an emergency intelligent alarm system for emergency treatment provided by an embodiment of the present invention;

[0060] Figure 3 An internal structure diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The emergency intelligent alarm method provided by the embodiment of the present invention is applied to the emergency intelligent alarm system of the hospital emergency department. Unless otherwise defined, all technical and scientific terms used in this article have the same meaning as those commonly understood by technicians in the technical field of this application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned drawings and any variations thereof are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0062] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0063] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0064] The emergency intelligent alarm method for emergency provided by the embodiment of the present application can be applied to the application environment of dietary control. Among them, the method can be applied to a terminal, and can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. Among them, the terminal can be, but is not limited to, various personal computers, laptops, etc. Among them, the terminal uses the current video data and current audio data collected in real time by an intelligent device (such as a camera device, a smart watch, a smart bracelet, a portable image acquisition device) set on the body surface of the medical staff, so as to comprehensively analyze the violence trend information corresponding to the medical staff, the semantic information corresponding to the current audio data, and the key word segment information. Thus, it is judged whether there is a risk of violence through dual channels. First, the violence probability of medical staff facing various types of violent incidents is comprehensively judged through violence trend information and semantic information. In this way, when medical staff cannot report the alarm in time, when it is judged that the violence probability is greater than the violence probability threshold, the initial alarm information is automatically generated, or when the medical staff speaks the keyword segment information, the alarm information corresponding to the keyword in the keyword segment is directly obtained through the database, and then the location information of the positioning intelligent device is used to generate an emergency alarm message sent to each target police officer, thereby ensuring that medical staff can actively voice alarm or intelligently analyze the alarm in a passive situation, greatly reducing the risk of violence caused by the inconvenient location of the one-button alarm device or the untimely alarm causing harm to the physical and mental safety of medical staff. Finally, this scheme predicts the risk of violence of medical staff efficiently and timely through positioning and dual-channel alarm information recognition, thereby effectively improving the efficiency of emergency alarms for medical staff in the face of violent incidents.

[0065] In one embodiment, Figure 1 As shown, an emergency intelligent alarm method for emergency treatment is provided, and the method is applied to a terminal as an example for explanation, including the following steps:

[0066] Step S101, obtaining the current video data collected by the smart device of the emergency medical staff and the current audio data collected by the smart device, and based on the current video data, identifying the violence trend information corresponding to the medical staff through the behavior trend recognition strategy.

[0067] In this embodiment, the terminal collects video data of the direction in which the medical staff is facing and the ambient audio data of the environment in which the medical staff is located in real time through the intelligent device set on the surface of the emergency medical staff. Among them, the intelligent device can be but not limited to smart watches, smart bracelets, portable cameras and other devices. Then, based on the current video data, the terminal identifies the violence trend information corresponding to the medical staff through the behavior trend recognition strategy. Among them, the trend recognition strategy is an identification strategy for identifying the sub-violence trend information of each person. The specific recognition process will be described in detail later. Among them, the violence trend information includes the sub-violence trend information of each person appearing in the video data. The sub-violence trend information is used to characterize the violence information generated by each person and the violence risk of the violence information. Among them, the violence risk of the same violence information is different. For example, the violence risk of a tall man hammering a doctor is much greater than the violence risk of a petite woman hammering a doctor. The violence risk of multiple people hammering a doctor is greater than the violence risk of a single person hammering a doctor. The specific recognition process will be described in detail later.

[0068] Step S102, based on the current audio data, the semantic information of the current audio data and the key segment information of the area range are identified through a semantic recognition strategy, and when the key segment information of the current audio data is 0, based on the violence trend information and the semantic information, the types of violence events corresponding to the medical staff and the violence probability of each type of violence event are analyzed through a violence trend analysis strategy.

[0069] In this embodiment, the terminal identifies the semantic information of the current audio data and the key segment information of the regional range through the semantic recognition strategy based on the current audio data, and when the key segment information of the current audio data is 0, based on the violence trend information and the semantic information, the terminal analyzes the types of violence events corresponding to the medical staff and the violence probability of each type of violence event through the violence trend analysis strategy. Among them, the semantic recognition strategy can not only identify the semantic information of the text content corresponding to the current audio data, but also identify the keyword information in the text content corresponding to the current audio data. Among them, the keyword is the alarm risk word preset by the staff for different alarm information, wherein different alarm risk words correspond to different alarm information. For example, the alarm risk word is A1dd, and the corresponding alarm information is a multi-person fight alarm information, the alarm risk word is A2cc, and the corresponding alarm information is a sexual assault alarm information, and the alarm risk word is A3ff. The alarm information is an instrument wounding alarm information, etc. Among them, the type of violence event is the type to which each violence information belongs. The type includes, but is not limited to, a gun violence event type, an instrument violence event type, a fight violence event type, a sexual assault violence event type, a suicide violence event type, etc. Each type of violence event corresponds to one or more violence information. Each piece of violence information corresponds to a specific violent behavior. For example, the types of weapon violence events include knife violence event information, stick violence event information, and other weapon violence event information. The specific analysis process will be described in detail later.

[0070] Step S103, when there is a violence probability greater than the violence probability threshold, based on the target violence probability greater than the violence probability threshold and the target violence event type corresponding to the target violence probability, generates the current initial alarm information, and when the keyword segment information of the current audio data is not 0, based on the keyword segment information, generates the current initial alarm information through the keyword segment alarm strategy.

[0071] In this embodiment, when there is a violence probability greater than the violence probability threshold, the terminal generates the current initial alarm information based on the target violence probability greater than the violence probability threshold and the target violence event type corresponding to the target violence probability, and generates the current initial alarm information based on the keyword segment information and the keyword segment alarm strategy when the keyword segment information of the current audio data is not 0. Among them, when the keyword segment information is 0, the current initial alarm information is each target violence information, the target person corresponding to each target violence information, and the violence risk of each target violence information. When the keyword segment information is not 0, the current initial alarm information is the alarm information corresponding to the keyword contained in the keyword segment information. The specific generation process will be described in detail later.

[0072] Step S104, collecting the current location information of the medical staff, and generating emergency alarm information of the medical staff based on the current location information and the current initial alarm information.

[0073] In this embodiment, the terminal obtains the current location information of the medical staff in real time through the positioning device set in the smart device carried by the medical staff, and generates the emergency alarm information of the medical staff based on the current location information and the current initial alarm information. Finally, the terminal sends the emergency alarm information to the alarm terminal of the target police officer. The target police officer includes but is not limited to the alarm terminal of the police department closest to the hospital, the alarm terminal of the hospital security department, etc.

[0074] Based on the above scheme, the current video data and current audio data collected in real time by the smart device (such as a camera, a smart watch, a smart bracelet, a portable image acquisition device) set on the body of the medical staff are used to comprehensively analyze the violence trend information corresponding to the medical staff, the semantic information corresponding to the current audio data, and the key word segment information. Therefore, it is judged whether there is a risk of violence through dual channels. First, the violence probability of the medical staff facing each type of violent event is comprehensively judged through the violence trend information and the semantic information. In the case that the medical staff cannot report the alarm in time, when it is judged that the violence probability is greater than the violence probability threshold, the initial alarm information is automatically generated, or when the medical staff speaks the key word segment information, the alarm information corresponding to the key word in the key word segment is directly obtained from the database, and then the location information of the smart device is located to generate an emergency alarm message sent to each target police officer, thereby ensuring that the medical staff can actively voice alarm or intelligently analyze the alarm in a passive situation, which greatly reduces the risk of violence incidents caused by the inconvenient location of the one-button alarm device or the untimely alarm causing harm to the physical and mental safety of the medical staff. Finally, this solution uses positioning and dual-channel alarm information recognition to efficiently and timely predict the risk of violence among medical staff, thereby effectively improving the efficiency of emergency alarms for medical staff in the face of violent incidents.

[0075] Optionally, based on the current video data, a behavior trend recognition strategy is used to identify the violence trend information corresponding to the medical staff, including: splitting the current video data into image data of each frame, and for each image data, using an image edge recognition algorithm to identify the facial range of each person in each image data, and the limb range of each person; using a motion recognition network to identify the body motion information of the limb image corresponding to the limb range of each person, and using an expression recognition network to identify the facial expression information of the facial image corresponding to the facial range of each person; sorting the body motion information of each image data of each person according to the time sequence of the image data, and performing Processing is performed to obtain the body movement change information of each person, and the facial expression information of each image data of each person is sorted and processed according to the time sequence of each image data to obtain the facial expression change information of each person; based on the body movement change information of each person and the facial expression change information of each person, the body movement change trend of each person and the facial expression change trend of each person are predicted, and based on the body movement change trend of each person and the facial expression change trend of each person, the sub-violence trend information of each person is identified through the violence trend identification strategy; the sub-violence trend information of all persons is used as the violence trend information corresponding to the medical staff.

[0076] In this embodiment, the terminal splits the current video data into image data of each frame, and for each image data, identifies the facial range of each person and the limb range of each person in each image data through an image edge recognition algorithm. The image edge recognition algorithm includes a face edge recognition algorithm and a body edge recognition algorithm. The face edge recognition algorithm is used to identify the facial range of each person, and the body edge recognition algorithm is used to identify the limb range of each person.

[0077] Then, the terminal uses the action recognition network obtained by training the convolutional neural network based on the self-attention mechanism to identify the body movement information of the body image of each person corresponding to the body range of each person, and uses the expression recognition network obtained by training the convolutional neural network based on the self-attention mechanism to identify the facial expression information of the facial image of each person corresponding to the facial range of each person. The action recognition network and the expression recognition network are respectively body image data marked by a large number of body movements and face image data marked by a large number of expression information. After the convolutional neural network is classified and trained, an action recognition network for identifying body movement information and an expression recognition network for identifying expression information are obtained.

[0078] Then, the terminal sorts the body movement information of each image data of each person according to the time sequence of each image data to obtain the body movement change information of each person, and sorts the facial expression information of each image data of each person according to the time sequence of each image data to obtain the facial expression change information of each person. Then, based on the body movement change information of each person and the facial expression change information of each person, the terminal predicts the body movement change trend of each person and the facial expression change trend of each person, and based on the body movement change trend of each person and the facial expression change trend of each person, identifies the sub-violence trend information of each person through the violence trend identification strategy. Specifically, the terminal identifies the body movement change trend of the person based on the body movement change information, and identifies the facial expression change trend of the person based on the facial expression change information through the reinforcement learning neural network. Among them, the body movement change trend is the change trend of the action type corresponding to the body movement of the person, and the facial expression change trend is the change trend of the expression type corresponding to the facial expression of the person. Among them, the action types include but are not limited to swinging type, kicking type, jumping type, grabbing type, holding arm swinging type, holding stabbing type, etc. The violence trend identification strategy includes a violence trend information corresponding to each action type range and expression type range. The terminal identifies the violence trend information of the person by matching the action type range to which the change trend of each action type belongs, and the expression type range to which the change trend of the expression type belongs. Each violence trend information corresponds to a piece of violence information. Then, the terminal identifies the person's body shape characteristics, gender characteristics, and age characteristics based on the image feature recognition network (CNN, Convolutional Neural Networks, Convolutional Neural Networks) of the self-attention mechanism and the personnel feature recognition network. After that, the terminal determines the violence risk of each violence information based on the person's body shape characteristics, gender characteristics, and age characteristics. Each person's physical feature corresponds to a violence risk range, each person's gender feature corresponds to a violence risk range, and each person's age feature corresponds to a violence risk range. The terminal then screens the intersection of the violence risk ranges as the target violence risk range for each violence information, and screens the highest violence risk in the target violence risk range as the violence risk for the violence information.

[0079] Finally, the terminal uses the sub-violence trend information of all personnel as the violence trend information corresponding to the medical staff.

[0080] Based on the above solution, after identifying the trend of expression changes and movement changes, the violence trend information of the person is identified, which improves the comprehensiveness and accuracy of identifying violence trend information.

[0081] Optionally, based on the current audio data, a semantic recognition strategy is used to identify the semantic information of the current audio data and key word segment information in a regional range, including: converting the current audio data into the current audio text through a text recognition network, and performing key word segment retrieval in the current audio text to identify the key word segment information in the current audio text; identifying the text semantic content of the current audio text through a semantic recognition network, and using the text semantic content as the semantic information of the current audio data.

[0082] In this embodiment, the terminal converts the current audio data into the current audio text through the text recognition network, and performs a keyword segment search in the current audio text to identify the keyword segment information in the current audio text. The keyword segment search is to query the keywords contained in the current audio text and the number of occurrences of the keywords in the current audio text to obtain the keyword segment information.

[0083] Afterwards, the terminal identifies the text semantic content of the current audio text through the semantic recognition network corresponding to the large language model, and uses the text semantic content as the semantic information of the current audio data.

[0084] Based on the above scheme, the accuracy and comprehensiveness of identifying the text semantic content and key word segment information of the current audio text are improved by searching keywords and performing semantic recognition.

[0085] Optionally, based on the violence trend information and semantic information, a violence trend analysis strategy is used to analyze the types of violent events corresponding to medical staff and the violence probability of each type of violent event, including: based on the semantic information, through a semantic screening network, identifying each violent semantic content in the semantic information, and identifying the recurrence frequency of each violent word corresponding to each violent semantic content in the current audio text; identifying the type of violent event corresponding to each violent semantic content, and based on the recurrence frequency of each violent word corresponding to each violent semantic content, identifying the initial violence probability of each violent event type; based on the sub-violence trend information of each person, identifying the violent information corresponding to each person and the violence risk corresponding to each violent information, adjusting the initial violence probability of each violent event type, and obtaining the violence probability of each violent event type.

[0086] In this embodiment, the terminal identifies each violent semantic content in the semantic information through a semantic screening network based on the semantic information, and identifies the recurrence frequency of each violent word corresponding to each violent semantic content in the current audio text. Among them, the semantic screening network is a semantic classification network based on a large language model. The semantic classification network can split the semantic information into units to obtain the semantic information of each unit audio text, and then identify the semantic type corresponding to the semantic information of each unit audio text respectively, and use the semantic information with a violent semantic type as the violent semantic content. Among them, the unit splitting is a splitting method of splitting the audio text according to commas or periods.

[0087] Then, the terminal uses a preset range of violent semantic content corresponding to each violent time type and a similarity recognition algorithm to identify the similarity between each violent semantic content and each semantic content in the range of violent semantic content of each violent event type, and selects the violent event type corresponding to the semantic content with the highest similarity as the violent event type corresponding to the violent semantic content.

[0088] Then, the terminal identifies the initial violence probability of each type of violent event based on the recurrence frequency of each violent word corresponding to each violent semantic content. The initial violence probability is the initial violence probability of each type of violent event obtained by calculating the violence probability corresponding to each recurrence frequency based on the ratio between each recurrence frequency and the violence probability preset in the terminal.

[0089] Finally, based on the sub-violence trend information of each person, the terminal identifies the violence information corresponding to each person and the violence risk corresponding to each violence information, adjusts the initial violence probability of each violent event type, and obtains the violence probability of each violent event type. The adjustment process is that each violent event type includes one or more violence information, and the terminal averages the violence risk of the violence information corresponding to each Paris event type to obtain the average violence risk. Then, the terminal uses the average violence risk as a weight value to perform weighted calculation on the initial violence probability to obtain the violence probability of each violent event.

[0090] Based on the above scheme, the initial violence probability of each violent event type is adjusted through the violence risk level of the violent information contained in each violent event type to obtain the violence probability of each violent event type, thereby improving the accuracy and comprehensiveness of the obtained violence probability of each violent event type.

[0091] Optionally, based on a target violence probability greater than a violence probability threshold and a target violence event type corresponding to the target violence probability, generating current initial alarm information, including: based on the target violence event type, identifying violence information corresponding to the target violence event type, and based on the target violence probability of the target violence event type and the violence risk of the violence information corresponding to each person, screening each target violence information and a target person corresponding to each target violence information; obtaining a preset alarm information template, and filling each target violence information, the target person corresponding to each target violence information, and the violence risk of each target violence information into the alarm information template to obtain the current initial alarm information.

[0092] In this embodiment, the terminal identifies the violence information corresponding to the target violence event type based on the target violence event type, and screens each target violence information and the target person corresponding to each target violence information based on the target violence probability of the target violence event type and the violence risk of the violence information corresponding to each person.

[0093] Then, the terminal obtains a preset alarm information template, and fills each target violence information, the target person corresponding to each target violence information, and the violence risk level of each target violence information into the alarm information template to obtain the current initial alarm information.

[0094] Based on the above scheme, by screening the target violence information, target personnel, and violence risk of each target violence information corresponding to the target violence event type, the simplicity, intuitiveness, and comprehensiveness of the alarm information are improved.

[0095] Optionally, based on the keyword segment information, the current initial alarm information is generated through the keyword segment alarm strategy, including: based on the keyword segment information, identifying each keyword contained in the keyword segment information, and querying the alarm information corresponding to each keyword in the database; using the alarm information corresponding to all keywords as the current initial alarm information.

[0096] In this embodiment, the terminal identifies each keyword contained in the keyword segment information based on the keyword segment information, and searches the database for the alarm information corresponding to each keyword. Finally, the alarm information corresponding to all keywords of the terminal is used as the current initial alarm information.

[0097] Based on the above scheme, the current initial alarm information is generated by querying the alarm information corresponding to each keyword, thereby improving the generation accuracy and generation efficiency of the current initial alarm information.

[0098] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0099] Based on the same inventive concept, the embodiment of the present application also provides an emergency intelligent alarm system for implementing the emergency intelligent alarm method involved above. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the emergency intelligent alarm system for emergency provided below can refer to the limitations of the emergency intelligent alarm method for emergency above, and will not be repeated here.

[0100] Further references Figure 2 , as a response to the above Figure 1 The implementation of the method shown in the figure, the present application provides an embodiment of an emergency intelligent alarm system 200 for emergency treatment, the emergency intelligent alarm system for emergency treatment includes an acquisition module 210, an identification module 220, a generation module 230 and an alarm module 240, wherein:

[0101] The acquisition module 210 is used to acquire the current video data collected by the smart device of the emergency medical staff and the current audio data collected by the smart device, and identify the violence trend information corresponding to the medical staff through the behavior trend recognition strategy based on the current video data;

[0102] The identification module 220 is used to identify the semantic information of the current audio data and the key segment information of the current audio data through a semantic identification strategy based on the current audio data, and analyze the types of violence events corresponding to the medical staff and the violence probability of each type of violence event through a violence trend analysis strategy based on the violence trend information and the semantic information when the key segment information of the current audio data is 0;

[0103] A generating module 230 is used to generate current initial alarm information based on a target violence probability greater than the violence probability threshold and a target violence event type corresponding to the target violence probability when there is a violence probability greater than the violence probability threshold, and to generate current initial alarm information based on the keyword segment information and through a keyword segment alarm strategy when the keyword segment information of the current audio data is not 0;

[0104] The alarm module 240 is used to collect the current location information of the medical staff, and generate emergency alarm information of the medical staff based on the current location information and the current initial alarm information.

[0105] Optionally, the acquisition module 210 is specifically configured to:

[0106] Splitting the current video data into image data of each frame, and identifying the facial range of each person and the limb range of each person in each image data by an image edge recognition algorithm;

[0107] The body movement information of the body image of each person corresponding to the body range of each person is identified through the action recognition network, and the facial expression information of the facial image of each person corresponding to the facial range of each person is identified through the expression recognition network;

[0108] According to the time sequence of each of the image data, the body movement information of each person's image data is sorted and processed to obtain the body movement change information of each person, and according to the time sequence of each of the image data, the facial expression information of each person's image data is sorted and processed to obtain the facial expression change information of each person;

[0109] Based on the body movement change information of each person and the facial expression change information of each person, predict the body movement change trend of each person and the facial expression change trend of each person, and based on the body movement change trend of each person and the facial expression change trend of each person, identify the sub-violence trend information of each person through the violence trend identification strategy;

[0110] The sub-violence trend information of all personnel is used as the violence trend information corresponding to the medical staff.

[0111] Optionally, the identification module 220 is specifically configured to:

[0112] The current audio data is converted into a current audio text through a text recognition network, and a keyword segment search is performed in the current audio text to identify the keyword segment information in the current audio text;

[0113] The text semantic content of the current audio text is identified through a semantic recognition network, and the text semantic content is used as the semantic information of the current audio data.

[0114] Optionally, the identification module 220 is specifically configured to:

[0115] Based on the semantic information, identifying each violent semantic content in the semantic information through a semantic screening network, and identifying the recurrence frequency of each violent word corresponding to each violent semantic content in the current audio text;

[0116] Identify the type of violent event corresponding to each violent semantic content, and identify the initial violent probability of each violent event type based on the recurrence frequency of each violent word corresponding to each violent semantic content;

[0117] Based on the sub-violence trend information of each person, the violence information corresponding to each person and the violence risk corresponding to each violence information are identified, and the initial violence probability of each type of violence event is adjusted to obtain the violence probability of each type of violence event.

[0118] Optionally, the generating module 230 is specifically configured to:

[0119] Based on the target violent event type, identifying the violent information corresponding to the target violent event type, and based on the target violence probability of the target violent event type and the violence risk level of the violent information corresponding to each person, screening each target violent information and the target person corresponding to each target violent information;

[0120] A preset alarm information template is obtained, and each target violence information, a target person corresponding to each target violence information, and a violence risk level of each target violence information are filled into the alarm information template to obtain the current initial alarm information.

[0121] Optionally, the generating module 230 is specifically configured to:

[0122] Based on the keyword segment information, identifying each keyword contained in the keyword segment information, and querying the alarm information corresponding to each keyword in the database;

[0123] The alarm information corresponding to all keywords is used as the current initial alarm information.

[0124] Each module in the above-mentioned emergency intelligent alarm system for emergency treatment can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0125] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input system connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, an emergency intelligent alarm method for emergency treatment is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input system of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0126] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0127] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of any one of the methods in the first aspect are implemented.

[0128] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the methods in the first aspect are implemented.

[0129] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the steps of any one of the methods in the first aspect.

[0130] It should be noted that the patient information (including but not limited to patient device information, patient personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the patient or fully authorized by all parties.

[0131] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0132] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0133] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. An emergency intelligent alarm method for emergency treatment, characterized in that: The method comprises: Acquire current video data collected by the smart device of the emergency medical staff and current audio data collected by the smart device, and identify the violence trend information corresponding to the medical staff through a behavior trend recognition strategy based on the current video data; Based on the current audio data, the semantic information of the current audio data and the key segment information of the current audio data are identified by a semantic recognition strategy, and when the key segment information of the current audio data is 0, based on the violence trend information and the semantic information, the types of violence events corresponding to the medical staff and the violence probability of each type of violence event are analyzed by a violence trend analysis strategy; When there is a violence probability greater than a violence probability threshold, based on the target violence probability greater than the violence probability threshold and the target violence event type corresponding to the target violence probability, generate current initial alarm information, and when the keyword segment information of the current audio data is not 0, based on the keyword segment information, generate current initial alarm information through a keyword segment alarm strategy; The current location information of the medical staff is collected, and based on the current location information and the current initial alarm information, emergency alarm information of the medical staff is generated.

2. The method according to claim 1, characterized in that The identifying, based on the current video data, the violence trend information corresponding to the medical staff through a behavior trend identification strategy includes: Splitting the current video data into image data of each frame, and identifying the facial range of each person and the limb range of each person in each image data by an image edge recognition algorithm; The body movement information of the body image of each person corresponding to the body range of each person is identified through the action recognition network, and the facial expression information of the facial image of each person corresponding to the facial range of each person is identified through the expression recognition network; According to the time sequence of each of the image data, the body movement information of each person's image data is sorted and processed to obtain the body movement change information of each person, and according to the time sequence of each of the image data, the facial expression information of each person's image data is sorted and processed to obtain the facial expression change information of each person; Based on the body movement change information of each person and the facial expression change information of each person, predict the body movement change trend of each person and the facial expression change trend of each person, and based on the body movement change trend of each person and the facial expression change trend of each person, identify the sub-violence trend information of each person through the violence trend identification strategy; The sub-violence trend information of all personnel is used as the violence trend information corresponding to the medical staff.

3. The method according to claim 1, characterized in that The identifying, based on the current audio data, the semantic information of the current audio data and the key segment information of the current audio data by using a semantic recognition strategy includes: The current audio data is converted into a current audio text through a text recognition network, and a keyword segment search is performed in the current audio text to identify the keyword segment information in the current audio text; The text semantic content of the current audio text is identified through a semantic recognition network, and the text semantic content is used as the semantic information of the current audio data.

4. The method according to claim 1, characterized in that: The analyzing, based on the violence trend information and the semantic information, the types of violence events corresponding to the medical staff and the violence probability of each type of violence event through a violence trend analysis strategy includes: Based on the semantic information, identifying each violent semantic content in the semantic information through a semantic screening network, and identifying the recurrence frequency of each violent word corresponding to each violent semantic content in the current audio text; Identify the type of violent event corresponding to each violent semantic content, and identify the initial violent probability of each violent event type based on the recurrence frequency of each violent word corresponding to each violent semantic content; Based on the sub-violence trend information of each person, the violence information corresponding to each person and the violence risk corresponding to each violence information are identified, and the initial violence probability of each type of violence event is adjusted to obtain the violence probability of each type of violence event.

5. The method according to claim 4, characterized in that The generating of current initial alarm information based on the target violence probability greater than the violence probability threshold and the target violence event type corresponding to the target violence probability includes: Based on the target violent event type, identifying the violent information corresponding to the target violent event type, and based on the target violence probability of the target violent event type and the violence risk level of the violent information corresponding to each person, screening each target violent information and the target person corresponding to each target violent information; A preset alarm information template is obtained, and each target violence information, a target person corresponding to each target violence information, and a violence risk level of each target violence information are filled into the alarm information template to obtain the current initial alarm information.

6. The method according to claim 1, characterized in that The generating of the current initial alarm information based on the keyword segment information and the keyword segment alarm strategy includes: Based on the keyword segment information, identifying each keyword contained in the keyword segment information, and querying the alarm information corresponding to each keyword in the database; The alarm information corresponding to all keywords is used as the current initial alarm information.

7. An emergency intelligent alarm system for emergency treatment, characterized in that: The system comprises: An acquisition module, used to acquire current video data collected by the smart device of the emergency medical staff and current audio data collected by the smart device, and identify the violence trend information corresponding to the medical staff through a behavior trend recognition strategy based on the current video data; an identification module, configured to identify semantic information of the current audio data and key segment information of the current audio data through a semantic identification strategy based on the current audio data, and to analyze the types of violence events corresponding to the medical staff and the violence probability of each type of violence event through a violence trend analysis strategy based on the violence trend information and the semantic information when the key segment information of the current audio data is 0; A generating module, configured to generate current initial alarm information based on a target violence probability greater than the violence probability threshold and a target violence event type corresponding to the target violence probability when there is a violence probability greater than the violence probability threshold, and to generate current initial alarm information based on the keyword segment information and through a keyword segment alarm strategy when the keyword segment information of the current audio data is not 0; The alarm module is used to collect the current location information of the medical staff and generate emergency alarm information of the medical staff based on the current location information and the current initial alarm information.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images