Early warning method and computer device for preventing pressure injuries

By combining body temperature imaging and real-time data acquisition of pressure sensors in the pressure injury early warning system, pre-trained models are used to perform early warnings, and accurate responses are provided through language generation technology, the shortcomings of pressure injury prevention and early warning in the prior art are solved. , effectively monitor and remind patients with long-term bedridden.

CN118749920BActive Publication Date: 2024-11-22CHINA JAPAN FRIENDSHIP HOSPITAL
View PDF 10 Cites 0 Cited by

Patent Information

Application Number
CN202411244200.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-06
Publication Date
2024-11-22
Estimated Expiration
2044-09-06

AI Technical Summary

Technical Problem

The prior art is difficult to effectively prevent and early warning of stress injuries, especially in patients who are bedridden for a long time. The lack of real-time and accurate monitoring and reminding mechanisms leads to an increase in the incidence of stress injuries.

Method used

Using a pressure damage warning system including body temperature camera devices, pressure sensors and voice playback equipment, determine the correlation between the device and the user through the association diagram, collect body temperature and pressure data in real time, input a pre-trained model for early warning, and prompt the user to change in a timely manner pose and provide accurate replies through language generation and knowledge base extraction techniques.

Benefits of technology

It improves the timeliness and accuracy of preventing pressure injuries, ensures that patients promptly remind and monitor during sleep, and reduces the risk of pressure injuries.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118749920B_ABST
    Figure CN118749920B_ABST
Patent Text Reader

Abstract

The present application relates to the field of early warning for preventing pressure injuries, and specifically to an early warning method and computer device for preventing pressure injuries. A specific implementation of the method includes: in response to receiving a pressure injury prevention monitoring instruction for a target user, determining whether each device is associated with the target user according to a pressure injury early warning system association diagram; in response to determining that each device is associated with the target user, collecting the target user's body index information; collecting the target user's body temperature data in real time; controlling the pressure sensor to detect the target user's pressure data in real time; inputting the body index information, the collected body temperature data sequence, and the collected pressure data sequence into a pressure injury information early warning model; in response to determining that the pressure injury information early warning result meets the early warning condition, controlling the prompt device to play a body posture change voice. This implementation can improve the timeliness and accuracy of preventing pressure injuries.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of early warning for preventing pressure injuries, and specifically to an early warning method and computer device for preventing pressure injuries. Background Art

[0002] Pressure injury is a localized injury to the skin and / or subcutaneous soft tissue caused by pressure or pressure combined with shear force. It is one of the most common complications in patients who need to stay in bed for a long time due to illness. With the aging trend of the population and the rising incidence of cardiovascular diseases, the incidence of pressure injuries has also been on the rise in recent years. At present, the prevention of pressure injuries usually involves changing body positions regularly, reminding patients to change their lying positions regularly, or having medical staff regularly help patients change their body positions. Summary of the invention

[0003] The content of this application is used to introduce concepts in a brief form, which will be described in detail in the detailed implementation section below. The content of this application is not intended to identify the key features or essential features of the technical solution claimed for protection, nor is it intended to limit the scope of the technical solution claimed for protection.

[0004] Some embodiments of the present application propose an early warning method, a computer device, and a computer-readable storage medium for preventing pressure injuries to solve the technical problems mentioned in the above background technology section.

[0005] In a first aspect, some embodiments of the present application provide an early warning method for preventing pressure injuries, which is applied to a pressure injury early warning system, wherein the pressure injury early warning system includes: a body temperature camera device, a pressure sensor, a prompt device and a voice playback device, and the method includes: in response to receiving a pressure injury prevention monitoring instruction for a target user, determining whether each device is associated with the target user according to a pressure injury early warning system association relationship diagram, wherein the pressure injury early warning system association relationship diagram represents a diagram of the association relationship between login identity information and each device, and each device refers to the body temperature camera device, pressure sensor, prompt device and voice playback device included in the pressure injury early warning system; in response to determining that each device is associated with the target user, collecting body index information of the target user; controlling the associated body temperature camera device to collect body temperature data of the target user in real time; controlling the associated pressure sensor to detect the pressure data of the target user in real time; inputting the body index information, the collected body temperature data sequence and the collected pressure data sequence into a pre-trained pressure In the pressure injury information warning model, a pressure injury information warning result is obtained; in response to determining that the above pressure injury information warning result meets the warning condition, the associated prompt device is controlled to play a body posture change voice to prompt the above target user to change the body posture; in response to receiving the pressure injury consultation voice issued by the above target user, the above pressure injury consultation voice is converted into text to obtain the pressure injury consultation voice text; the above pressure injury consultation voice text is vectorized to obtain the pressure injury consultation voice text vector; according to the above pressure injury consultation voice text vector, a target pressure injury consultation knowledge fragment matching the above pressure injury consultation voice text is extracted from a preset pressure injury consultation knowledge base; according to the above pressure injury consultation voice text and the above target pressure injury consultation knowledge fragment, a language prompt-based method is used to perform data splicing to generate pressure injury consultation reply information, and the above pressure injury consultation reply information is converted into pressure injury consultation reply voice, and the associated voice playback device is controlled to play it.

[0006] In a second aspect, the present application also provides a computer device, comprising a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, the method described in any implementation of the first aspect is implemented.

[0007] In a third aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, wherein when the computer program is executed by a processor, the method described in any implementation of the first aspect is implemented.

[0008] The above-mentioned embodiments of the present application have the following beneficial effects: the timeliness and accuracy of preventing pressure injuries are improved, and while ensuring the patient user's sleep, the user can be warned more accurately. First, in response to receiving a pressure injury prevention monitoring instruction for a target user, according to the pressure injury warning system association diagram, it is determined whether each device is associated with the above-mentioned target user, wherein the above-mentioned pressure injury warning system association diagram represents a diagram of the association relationship between the login identity information and each device, and each device refers to the body temperature camera device, pressure sensor, prompt device and voice playback device included in the pressure injury warning system. Thus, the user's body temperature can be collected in real time by the body temperature camera device, so as to determine whether the user may sweat according to the body temperature. The pressure data generated when the user lies down / sits down can be monitored in real time by the pressure sensor, so as to determine the maximum lying time of the user. The prompt device and the voice playback device can remind the user to turn over in time. In addition, it can be determined whether each device has logged into the target user's account through the pressure injury warning system association diagram to ensure the accuracy of information collection. Secondly, in response to determining that each device is associated with the target user, the body index information of the target user is collected; the associated body temperature camera is controlled to collect the body temperature data of the target user in real time; and the associated pressure sensor is controlled to detect the pressure data of the target user in real time. Thus, data support is provided for preventing pressure injuries. Then, the body index information, the collected body temperature data sequence and the collected pressure data sequence are input into the pre-trained pressure injury information warning model to obtain the pressure injury information warning result. Thus, the pressure injury warning can be performed through the pre-trained pressure injury information warning model to prompt the user to turn over. Afterwards, in response to determining that the pressure injury information warning result meets the warning condition, the associated prompt device is controlled to play the body posture change voice to prompt the target user to change the body posture. Then, in response to receiving the pressure injury consultation voice issued by the target user, the pressure injury consultation voice is converted into text to obtain the pressure injury consultation voice text; the pressure injury consultation voice text is vectorized to obtain the pressure injury consultation voice text vector. Finally, based on the above-mentioned pressure injury consultation voice text vector, a target pressure injury consultation knowledge fragment matching the above-mentioned pressure injury consultation voice text is extracted from a preset pressure injury consultation knowledge base; based on the above-mentioned pressure injury consultation voice text and the above-mentioned target pressure injury consultation knowledge fragment, a language prompt-based method is used to perform data splicing to generate pressure injury consultation reply information, and the above-mentioned pressure injury consultation reply information is converted into pressure injury consultation reply voice, and the associated voice playback device is controlled to play it.Therefore, based on the retrieval enhancement and generation capabilities of the large model, pressure injury consultation knowledge fragments can be extracted from the pre-built pressure injury consultation knowledge base. Then, through the prompt method splicing, they can be rewritten into more accurate, controllable, and anthropomorphic replies. Since the knowledge fragment vectors of pressure injury knowledge are pre-stored in the pressure injury consultation knowledge base, when extracting pressure injury consultation knowledge fragments, only similarity calculation is needed to complete the recall of pressure injury consultation knowledge fragments. Thereby, the accuracy of reply information generation is improved. It is convenient to guide users to turn over or stand up and walk, so as to improve the timeliness and accuracy of preventing pressure injuries. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar figure labels represent the same or similar elements. It should be understood that the drawings are schematic and that the components and elements are not necessarily drawn to scale.

[0010] Figure 1 is a flow chart of some embodiments of the early warning method for preventing pressure injuries according to the present application;

[0011] Figure 2 It is a schematic diagram of the structure of a computer device suitable for implementing some embodiments of the present application. DETAILED DESCRIPTION

[0012] Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.

[0013] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other.

[0014] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0015] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0016] The names of the messages or information exchanged between multiple devices in the embodiments of the present application are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0017] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0018] Figure 1 The flowchart of some embodiments of the early warning method for preventing pressure injuries according to the present application is shown. The flowchart of some embodiments of the early warning method for preventing pressure injuries according to the present application is shown. The early warning method for preventing pressure injuries is applied to a pressure injury early warning system, which includes: a body temperature camera device, a pressure sensor, a prompt device and a voice playback device, and includes the following steps:

[0019] Step 101, in response to receiving a pressure injury prevention monitoring instruction for a target user, determining whether each device is associated with the target user according to a pressure injury early warning system association relationship diagram.

[0020] In some embodiments, the execution subject (e.g., a computer device) of the early warning method for preventing pressure injuries may determine whether each device is associated with the target user in response to receiving a pressure injury prevention monitoring instruction for the target user according to the pressure injury early warning system association relationship diagram. The pressure injury early warning system association relationship diagram represents a diagram of the association relationship between the login identity information and each device, and each device refers to the body temperature camera device, pressure sensor, prompt device and voice playback device included in the pressure injury early warning system. The body temperature camera device may refer to a body temperature camera for collecting the user's body temperature. The pressure sensor may refer to a pressure sensor set on the target user or on the object on which the target user lies. For example, the pressure sensor may be set on the back or buttocks of the target user. Thus, the pressure of the target user when sitting or lying down can be detected. For another example, the pressure sensor may be set on a bed or a sofa to detect the pressure data of the target user after lying down or sitting down. The target user may be a user with a pressure injury to be detected. The prompt device may be a voice device used to prompt the user to change posture in time to avoid pressure injuries to the user. The voice playback device may be a speaker. The pressure injury warning system association diagram may be an association diagram generated after information association statistics are performed on the user identity information of each device in the pressure injury warning system during a target historical time period. The identity information may be the user's account information. The association relationship may be a login relationship. The association diagram is an association diagram for multiple device information. For example, determine whether the login information of each device is the account information of the target user. The pressure injury prevention monitoring instruction may refer to an instruction to monitor the target user for pressure injuries. The pressure injury may refer to a pressure sore.

[0021] In an actual application scenario, the execution subject can determine whether each device is associated with the target user through the following steps:

[0022] The first step is to generate user identification information corresponding to each device information in the pressure injury warning system association diagram using a preset user identification rule model. The user identification information is information that characterizes whether the corresponding device-associated users are the same user. The user identification rule model can be a rule model for identifying whether each device information corresponds to each device-associated user. There is a one-to-one correspondence between the device-associated user in each device-associated user and the device information in each device information. The device-associated user can be a user who logs in to the account of the device corresponding to the device information.

[0023] As an example, the execution subject may determine whether there is a corresponding device change time for each two device information of the same login identity information according to the user identification rule model. In response to determining the device change time, it is determined that the two device information correspond to the same user. In response to determining the device change time, it is determined that the two device information do not correspond to the same user.

[0024] In a second step, in response to the user information represented by the user identification information not being the target user, adjusting the login association relationship between the respective devices and the target user.

[0025] The execution subject may generate user identification information corresponding to each device information in the pressure injury early warning system association diagram through the following steps:

[0026] First, determine every two target device information in the above-mentioned device information as a target device information group, and obtain at least one target device information group. Among them, at least two target device information in the target device information group have corresponding identical login identity information. At least two target device information in the target device information group have corresponding identical Pin account information.

[0027] Second, for each target device information group in the at least one target device information group, perform the following generation steps:

[0028] 1. Determine the user behavior overlap corresponding to the two target device information in the target device information group. The user behavior overlap can be numerical information, and the larger the numerical value, the more similar the behaviors of the device-associated users between the two target device information are. The user behavior overlap can represent the degree of similarity between the behaviors of the two device-associated users corresponding to the two target device information.

[0029] 2. Generate user identification sub-information for the above two target device information based on the above user behavior overlap. The above user identification sub-information represents information on whether the two target device-associated users are the same user, and the above two target device-associated users correspond to the above two target device information. There is a one-to-one correspondence between the target device-associated users in the two target device-associated users and the target device information in the two target device information. As an example, in response to determining that the user behavior overlap is less than or equal to the preset overlap, generate user identification sub-information representing information that the two target device-associated users are the same user. In response to determining that the user behavior overlap is greater than the preset overlap, generate user identification sub-information representing information that the two target device-associated users are not the same user.

[0030] Third, the user identification information is generated according to the obtained at least one user identification sub-information. The execution subject may combine the information of each user identification sub-information in the at least one user identification sub-information to obtain the user identification information.

[0031] Wherein, determining the user behavior overlap degree corresponding to two target device information in the target device information group includes:

[0032] 1. Determine the target behavior information of the first target device information within the target historical time period as the first target behavior information. The target behavior information may be the behavior content corresponding to the target behavior, and the target behavior may be a pre-set behavior. For example, the target behavior may be an account login behavior. The target historical time period may be a pre-set historical time period.

[0033] 2. Determine the target behavior information of the second target device information within the target historical time period as the second target behavior information. Similarly, the second target behavior information may be the number of login times of the device corresponding to the second target device information performing the login account behavior within the target historical time period.

[0034] 3. Generate target behavior feature similarity information between the first target behavior information and the second target behavior information. As an example, first, vectorize the first target behavior information and the second target behavior information to generate a first target behavior vector and a second target behavior vector. Then, perform vector dot product processing on the first target behavior vector and the second target behavior vector to generate target behavior feature similarity information.

[0035] 4. Determine the start time of the above-mentioned first target device information under the above-mentioned target behavior information as the first start time, and determine the start time of the above-mentioned second target device information under the above-mentioned target behavior information as the second start time. Among them, the first start time includes: a first start time and a first end time. The first start time can be the start time when the device corresponding to the first target device information performs the target behavior. The first end time can be the end time when the device corresponding to the first target device information performs the target behavior. The second start time includes: a second start time and a second end time. The second start time can be the start time when the device corresponding to the second target device information performs the target behavior. The second end time can be the end time when the device corresponding to the second target device information performs the target behavior.

[0036] 5. Generate time overlap information based on the first start time and the second start time. The time overlap information may be the overlap duration of the behavior cycle of the device corresponding to the first target device information and the device corresponding to the second target device information. As an example, first, the execution subject may determine the minimum time between the first end time and the second end time. Then, determine the maximum time between the first start time and the second start time. Next, determine the time difference between the minimum time and the maximum time. Finally, determine the maximum value between the time difference and the value 0 as the time overlap information.

[0037] 6. Generate the user behavior overlap degree based on the target behavior feature similarity information and the time overlap information. For example, the target behavior feature similarity information and the time overlap information may be weighted and summed to generate the user behavior overlap degree.

[0038] Therefore, by determining the user behavior overlap between each pair of target device information, the degree of behavior similarity between the two device-associated users corresponding to the two target device information can be accurately measured, so as to accurately generate user identification information.

[0039] Step 102: In response to determining that each device is associated with the target user, body index information of the target user is collected.

[0040] In some embodiments, the execution subject may collect body index information of the target user in response to determining that each device is associated with the target user. The body index information may be height, weight, and disease information of the target user.

[0041] Step 103: Control the associated body temperature camera device to collect the body temperature data of the target user in real time.

[0042] In some embodiments, the execution subject may control the associated body temperature camera to collect the body temperature data of the target user in real time. For example, the execution subject may control the associated body temperature camera to collect the body temperature data sequence of the target user in a preset time period in real time. One body temperature data corresponds to one time granularity.

[0043] Step 104: Control the associated pressure sensor to detect the pressure data of the target user in real time.

[0044] In some embodiments, the execution subject may control the associated pressure sensor to detect the pressure data of the target user in real time. For example, the associated pressure sensor may be controlled to collect the pressure data of the target user in a preset time period in real time. One pressure data corresponds to one time granularity. One time granularity may be 1 minute. The preset time period may be a pre-set detection duration. For example, 30 minutes. It should be noted that the preset time period should not be too long to avoid delaying the detection.

[0045] Step 105, input the above body index information, the collected body temperature data sequence and the collected pressure data sequence into a pre-trained pressure injury information warning model to obtain a pressure injury information warning result.

[0046] In some embodiments, the execution subject may input the body index information, the collected temperature data sequence, and the collected pressure data sequence into a pre-trained pressure injury information warning model to obtain a pressure injury information warning result. The pressure injury information warning model may be a pre-trained neural network model that takes the body index information, the collected temperature data sequence, and the collected pressure data sequence as input and takes the pressure injury information warning result as output. The pressure injury information warning result may be a result (probability) indicating the possibility of the target user causing pressure injury within a preset time period.

[0047] The pre-trained pressure injury information warning model can be obtained by training through the following steps:

[0048] First, obtain a user pressure injury information training sample set. Among them, one user pressure injury information training sample corresponds to one user. The user pressure injury information training sample may include: user pressure injury information sample and sample label. The user pressure injury information sample includes: body index information, body temperature data sequence and pressure data sequence.

[0049] Second, based on the user's pressure injury information training sample set, perform the following training steps:

[0050] 1. Inputting at least one user pressure injury information training sample in the user pressure injury information training sample set into the initial pressure injury information warning model to obtain an initial pressure injury information warning result corresponding to the at least one user pressure injury information training sample. One initial pressure injury information warning result corresponds to one user pressure injury information training sample.

[0051] 2. Determine whether the initial pressure injury information warning model has reached a preset optimization target based on the initial pressure injury information warning result corresponding to the at least one user pressure injury information training sample and the sample label corresponding to the at least one user pressure injury information training sample. The preset optimization target may be that the error value between the initial pressure injury information warning result of the at least one user pressure injury information training sample and the sample label of the at least one user pressure injury information training sample is less than a preset value.

[0052] 3. In response to determining that the initial pressure injury information warning model has achieved the above-mentioned preset optimization goal, the initial pressure injury information warning model is determined as the trained pressure injury information warning model.

[0053] Third, in response to determining that the initial pressure injury information warning model has not reached the above-mentioned preset optimization target, the network parameters of the initial pressure injury information warning model are adjusted, and the user pressure injury information training sample set composed of unused user pressure injury information training samples is used to perform the above-mentioned training steps again. For example, the difference between the above-mentioned error value and the preset error value can be calculated. On this basis, the network parameters of the above-mentioned initial pressure injury information warning model are adjusted using methods such as back propagation and gradient descent. It should be noted that the back propagation algorithm and the gradient descent method are currently well-known technologies that are widely studied and applied, and will not be repeated here. Among them, there is no limitation on the setting of the preset difference value. For example, the preset difference value can be 0.1.

[0054] Step 106, in response to determining that the pressure injury information warning result meets the warning condition, controlling the associated prompt device to play a body posture change voice to prompt the target user to change the body posture.

[0055] In some embodiments, the execution subject may control the associated prompt device to play a body posture change voice in response to determining that the pressure injury information warning result meets the warning condition, so as to prompt the target user to change the body posture. The warning condition may be: the probability represented by the pressure injury information warning result is greater than or equal to the preset probability. The body posture change voice may be a voice that prompts the target user to change the body posture to avoid a certain part of the user being compressed for a long time.

[0056] Step 107, in response to receiving the pressure injury consultation voice from the target user, converting the pressure injury consultation voice into text to obtain the pressure injury consultation voice text.

[0057] In some embodiments, the execution subject may, in response to receiving the pressure injury consultation voice issued by the target user, convert the pressure injury consultation voice into text to obtain the pressure injury consultation voice text. For example, the pressure injury consultation voice may be converted into text by voice conversion technology to obtain the pressure injury consultation voice text. Voice conversion technology may refer to a technology for converting voice into text.

[0058] Step 108, vectorizing the pressure injury consultation voice text to obtain a pressure injury consultation voice text vector.

[0059] In some embodiments, the execution subject may vectorize the pressure injury consultation voice text to obtain a pressure injury consultation voice text vector. As an example, the execution subject may input the pressure injury consultation voice text into a large model, thereby outputting a vector representation of the pressure injury consultation voice text. The large model here may adopt various language model structures, such as GPT-3 (artificial intelligence language model), LLM (Large Language Model), etc.

[0060] Step 109: based on the pressure injury consultation voice text vector, extract a target pressure injury consultation knowledge segment that matches the pressure injury consultation voice text from a preset pressure injury consultation knowledge base.

[0061] In some embodiments, the execution subject may extract a target pressure injury consultation knowledge segment that matches the pressure injury consultation voice text from a preset pressure injury consultation knowledge base based on the pressure injury consultation voice text vector. The pressure injury consultation knowledge base may be a database that pre-stores various pressure injury consultation knowledge segments. For example, a pressure injury consultation knowledge segment may be a response segment to a certain pressure injury consultation question.

[0062] In an actual application scenario, the execution subject may extract a target pressure injury consultation knowledge segment that matches the pressure injury consultation voice text from a preset pressure injury consultation knowledge base through the following steps:

[0063] The first step is to determine the similarity value between each pressure injury consultation knowledge segment vector in the preset pressure injury consultation knowledge base and the pressure injury consultation voice text vector. For example, the similarity value between each pressure injury consultation knowledge segment vector in the preset pressure injury consultation knowledge base and the pressure injury consultation voice text vector can be determined by the cosine similarity formula.

[0064] In the second step, a preset number of pressure injury consultation knowledge fragments are extracted from the above-mentioned pressure injury consultation knowledge base in order of similarity values ​​from high to low, as a group of alternative pressure injury consultation knowledge fragments.

[0065] The third step is to adjust the order of the candidate pressure injury consultation knowledge fragments in the above-mentioned candidate pressure injury consultation knowledge fragment group according to the preset rules, and extract the target number of candidate pressure injury consultation knowledge fragments from the adjusted sequence of candidate pressure injury consultation knowledge fragments as target pressure injury consultation knowledge fragments. Here, the preset rules can be set according to actual needs. In this way, the pressure injury consultation knowledge fragments that meet the requirements can be further screened out to improve the accuracy of the extraction results.

[0066] The second step may include the following sub-steps:

[0067] Sub-step 1, readjusting the order of each candidate pressure injury consultation knowledge fragment according to the similarity value order of each candidate pressure injury consultation knowledge fragment and the time difference between the time of writing into the above pressure injury consultation knowledge base and the current time.

[0068] Sub-step 2, traverse each alternative pressure injury consultation knowledge segment in the above-mentioned alternative pressure injury consultation knowledge segment group, and adjust the order of each alternative pressure injury consultation knowledge segment using the maximum margin correlation algorithm based on the similarity value between the alternative pressure injury consultation knowledge segment and the above-mentioned pressure injury consultation voice text vector, as well as the similarity value between the alternative pressure injury consultation knowledge segment and the traversed alternative pressure injury consultation knowledge segment.

[0069] Step 110, based on the above-mentioned pressure injury consultation voice text and the above-mentioned target pressure injury consultation knowledge fragment, a language prompt-based method is used to perform data splicing, generate pressure injury consultation response information, and convert the above-mentioned pressure injury consultation response information into pressure injury consultation response voice, and control the associated voice playback device to play it.

[0070] In some embodiments, the execution subject may perform data splicing based on the pressure injury consultation voice text and the target pressure injury consultation knowledge fragment using a language prompt-based method (prompt) to generate pressure injury consultation reply information, and convert the pressure injury consultation reply information into pressure injury consultation reply voice, and control the associated voice playback device to play it. Among them, the prompt method is generally a technology that provides a text prompt (prompt) to the model to guide it to generate text of a specific style or content. Specifically, a language prompt-based method may be used to splice the pressure injury consultation voice text and the target pressure injury consultation knowledge fragment in a preset format. Afterwards, the spliced ​​data may be input into a preset pressure injury consultation reply model to generate reply information for the information consultation request.

[0071] In an actual application scenario, the above execution entity can generate pressure injury consultation response information through the following steps:

[0072] The first step is to verify the target pressure injury consultation knowledge fragment according to the preset conditions. The preset conditions here can also be set by yourself. For example, the time can be set to be less than half a year and the correlation can be greater than 0.8.

[0073] In the second step, in response to determining that the target pressure injury consultation knowledge segment has passed the verification, the pressure injury consultation voice text and the target pressure injury consultation knowledge segment are spliced ​​according to a preset format using a language prompt-based method. The preset format is not limited.

[0074] The third step is to input the spliced ​​data into a preset pressure injury consultation response model to generate pressure injury consultation response information. The preset pressure injury consultation response model may refer to a model that converts the spliced ​​data into standard response information.

[0075] In the fourth step, in response to determining that the target pressure injury consultation knowledge segment fails the verification, the pressure injury consultation voice text is sent to a manual reply terminal for manual reply.

[0076] As an example, when a knowledge fragment of pressure injury consultation fails to pass the preset condition (such as relevance) verification, the fallback logic will generally be triggered. That is, the user can be informed that this request (question) cannot be answered temporarily according to a unified reply template, and the request can be submitted asynchronously to the manual customer service to generate a reply text for the request. When the knowledge fragment can pass the preset condition verification, the sorted target pressure injury consultation knowledge fragment and the pressure injury consultation voice text will be spliced ​​together to form a complete prompt, which will be rewritten through the large model. Rewriting can make the text generated by the large model as close to the meaning of the original text as possible, which can make the request response controllable.

[0077] Furthermore, in order to improve the accuracy of the information reply, after the reply information is generated, the reply information can be further verified. Specifically, a pre-trained verification model can be used to verify the factuality of the generated reply information. The verification model here can also use various machine learning structures. The verification model is generally obtained by training the training data.

[0078] The present application also provides a computer device 200. Figure 2As shown, the computer device 200 includes: a bus 201, a processor 202, a memory 203 and a communication interface 204. The processor 202, the memory 203 and the communication interface 204 communicate with each other through the bus 201. The computer device 200 can be a server or a terminal device. It should be understood that the present application does not limit the number of processors and memories in the computer device 200.

[0079] The bus 201 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 The bus 201 is represented by only one line, but it does not mean that there is only one bus or one type of bus. The bus 201 may include a path for transmitting information between various components of the computer device 200 (for example, the memory 203, the processor 202, and the communication interface 204).

[0080] The processor 202 may include any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0081] The memory 203 may include a volatile memory, such as a random access memory (RAM). The memory 203 may also include a non-volatile memory, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid state drive (SSD).

[0082] The memory 203 stores executable program codes, and the processor 202 executes the executable program codes to respectively implement the functions of the aforementioned acquisition module, sampling module, determination module, and mixing module, thereby implementing the aforementioned method for early warning of preventing pressure injuries. That is, the memory 203 stores instructions for executing the aforementioned method for early warning of preventing pressure injuries.

[0083] The communication interface 204 uses a transceiver module such as, but not limited to, a network interface card or a transceiver to implement communication between the computer device 200 and other devices or communication networks.

[0084] An embodiment of the present application also provides a chip, which includes a processor and a data interface. The processor reads instructions stored in a memory through the data interface to execute the above-mentioned early warning method for preventing pressure injuries.

[0085] The embodiment of the present application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that can be stored by a computing device or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk). The computer-readable storage medium includes instructions that instruct the computing device to execute the above-mentioned early warning method for preventing pressure injuries.

[0086] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described 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.

[0087] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the protection scope of the technical solutions of the embodiments of the present application.

Claims

1. A warning method for preventing pressure injuries, applied to a pressure injury warning system, the pressure injury warning system comprising: A body temperature camera device, a pressure sensor, a prompt device and a voice playback device, and the method includes: In response to receiving a pressure injury prevention monitoring instruction for a target user, determining whether each device is associated with the target user according to a pressure injury warning system association relationship diagram, wherein the pressure injury warning system association relationship diagram represents a diagram of association relationships between login identity information and each device, and each device refers to a body temperature camera device, a pressure sensor, a prompt device, and a voice playback device included in the pressure injury warning system; In response to determining that each device is associated with the target user, collecting body index information of the target user; Controlling the associated body temperature camera device to collect the body temperature data of the target user in real time; Controlling the associated pressure sensor to detect the pressure data of the target user in real time; Inputting the body index information, the collected temperature data sequence and the collected pressure data sequence into a pre-trained pressure injury information warning model to obtain a pressure injury information warning result; In response to determining that the pressure injury information warning result meets the warning condition, controlling the associated prompting device to play a body posture change voice to prompt the target user to change the body posture; In response to receiving the pressure injury consultation voice issued by the target user, converting the pressure injury consultation voice into text to obtain the pressure injury consultation voice text; Performing vectorization processing on the pressure injury consultation voice text to obtain a pressure injury consultation voice text vector; According to the pressure injury consultation voice text vector, extracting a target pressure injury consultation knowledge segment that matches the pressure injury consultation voice text from a preset pressure injury consultation knowledge base; According to the pressure injury consultation voice text and the target pressure injury consultation knowledge fragment, a language prompt-based method is used to perform data splicing to generate pressure injury consultation reply information, and the pressure injury consultation reply information is converted into pressure injury consultation reply voice, and the associated voice playback device is controlled to play; The step of extracting a target pressure injury consultation knowledge segment that matches the pressure injury consultation voice text from a preset pressure injury consultation knowledge base includes: Determine a similarity value between each pressure injury consultation knowledge segment vector in a preset pressure injury consultation knowledge base and the pressure injury consultation speech text vector; Extracting a preset number of pressure injury consultation knowledge fragments from the pressure injury consultation knowledge base in descending order of similarity values ​​as a candidate pressure injury consultation knowledge fragment group; According to a preset rule, adjusting the order of each candidate pressure injury consulting knowledge fragment in the candidate pressure injury consulting knowledge fragment group, and extracting a target number of candidate pressure injury consulting knowledge fragments from the adjusted sequence of candidate pressure injury consulting knowledge fragments as target pressure injury consulting knowledge fragments; The adjusting of the order of the candidate pressure injury consultation knowledge segments in the candidate pressure injury consultation knowledge segment group includes: Re-adjusting the order of the candidate pressure injury consultation knowledge fragments according to the similarity value order of the candidate pressure injury consultation knowledge fragments and the time difference between the time of writing into the pressure injury consultation knowledge base and the current time; Traversing each candidate pressure injury consultation knowledge segment in the candidate pressure injury consultation knowledge segment group, and adjusting the order of each candidate pressure injury consultation knowledge segment using a maximum margin correlation algorithm according to a similarity value between the candidate pressure injury consultation knowledge segment and the pressure injury consultation speech text vector, and a similarity value between the candidate pressure injury consultation knowledge segment and the traversed candidate pressure injury consultation knowledge segment; Among them, the method of splicing data based on the pressure injury consultation voice text and the target pressure injury consultation knowledge fragment is used to generate pressure injury consultation reply information by using a language prompt-based method, including: According to preset conditions, verifying the target pressure injury consultation knowledge fragment; In response to determining that the target pressure injury consultation knowledge segment has passed verification, splicing the pressure injury consultation voice text and the target pressure injury consultation knowledge segment using a language prompt-based method in accordance with a preset format; Input the spliced ​​data into a preset pressure injury consultation response model to generate pressure injury consultation response information; In response to determining that the target pressure injury consultation knowledge segment fails verification, sending the pressure injury consultation voice text to a manual reply terminal for manual reply; Wherein, determining whether each device is associated with the target user according to the pressure injury early warning system association relationship diagram includes: Using a preset user identification rule model, generating user identification information corresponding to each device information in the pressure injury early warning system association relationship diagram, wherein the user identification information is information indicating whether the users associated with each device are the same user; In response to the user information represented by the user identification information not being the target user, adjusting the login association relationship between each device and the target user; Wherein, the user identification information corresponding to each device information in the pressure injury early warning system association relationship diagram is generated by using a preset user identification rule model, including: Determine every two target device information in the respective device information as a target device information group, and obtain at least one target device information group, wherein at least two target device information in the target device information group have corresponding identical login identity information; For each target device information group in the at least one target device information group, the following generating steps are performed: Determine the user behavior overlap degree corresponding to two target device information in the target device information group; generating user identification sub-information for the two target device information according to the user behavior overlap, wherein the user identification sub-information represents information on whether the two target device-associated users are the same user, and the two target device-associated users correspond to the two target device information; generating user identification information according to the obtained at least one user identification sub-information; The step of determining the user behavior overlap between two pieces of target device information in the target device information group includes: Determine target behavior information of the first target device information within a target historical time period as first target behavior information; Determine target behavior information of the second target device information within the target historical time period as second target behavior information; generating target behavior feature similarity information between the first target behavior information and the second target behavior information; Determine a start time of the first target device information under the target behavior information as a first start time, and determine a start time of the second target device information under the target behavior information as a second start time; Generate time overlap information according to the first start time and the second start time; The user behavior overlap degree is generated according to the target behavior feature similarity information and the time overlap information.

2. A computer device, wherein: The computer device comprises a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein the steps of the method according to claim 1 are implemented when the computer program is executed by the processor.

3. A computer-readable storage medium, wherein: The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the method according to claim 1 are implemented.

Citation Information

Patent Citations

  • Pressure sore early-warning device

    CN111388236A

  • Pressure injury dynamic early warning system

    CN112185570A

  • Real-time dynamic prevention and detection system for pressure injury based on pressure sensor

    CN112370271A

  • Early warning method for pressure sore adverse events

    CN112382388A

  • Monitoring system, device and computer-implemented method for monitoring pressure ulcers

    CN113261945A