Interactive nursing control method and system for tracheotomy patient

Abstract

The invention relates to the field of medical data analysis, in particular to an interactive nursing control method and system for a tracheotomy patient, and the method comprises the steps: drawing a simulated feeding scene for a user subjected to tracheotomy based on a virtual reality technology, and stimulating the user to generate a swallowing action through virtual food; collecting bio-electricity signals generated by muscle group activities of multiple parts of the neck, the lower jaw, the tongue and the throat when the user performs swallowing actions, and inputting the bio-electricity signals into the trained neural network model; whether the swallowing action of the user is abnormal or not is judged through the neural network model, and the abnormal part is recognized when it is judged that the swallowing action of the user is abnormal. According to the method, the enthusiasm and participation degree of the patient on rehabilitation training are remarkably improved through interestingness and interactivity characteristics of the virtual reality technology, the recognition efficiency and accuracy of the swallowing dysfunction are remarkably improved, and the rehabilitation cycle of the patient is shortened.

Description

Technical Field

[0001] The present invention relates to the field of medical data analysis, and more specifically, to an interactive nursing control method and system for tracheotomy patients. Background Art

[0002] Tracheotomy is a common surgical procedure used to treat dyspnea or protect the airway from foreign bodies, especially for patients who rely on mechanical ventilation for a long time. However, this surgery is often accompanied by the risk of swallowing disorders, namely dysphagia, because the surgery directly affects the muscle control and sensory feedback of the throat, making it difficult for patients to swallow food and liquids safely and effectively after surgery, increasing the risk of aspiration and lung infection.

[0003] Traditional swallowing function rehabilitation methods mainly rely on face-to-face training under the guidance of physical therapists, including oral and facial exercises, swallowing muscle exercises, etc. However, such methods are limited by uneven resource allocation, lack of personalization, and difficulty in simulating real eating scenes, making it difficult to efficiently and systematically promote the full recovery of patients' swallowing function.

[0004] In recent years, the development of virtual reality (VR) technology and deep learning has brought new opportunities to the field of medical rehabilitation. VR technology can create an immersive simulation environment, allowing patients to feel as if they are in a real eating scene. It not only enhances the fun and interactivity of rehabilitation training, but also provides rich training conditions by simulating foods of different textures and temperatures. Deep learning algorithms can automatically learn and analyze patients' swallowing patterns based on a large amount of data, accurately identify the type and severity of swallowing disorders, and further guide the formulation of personalized rehabilitation plans.

[0005] Currently, there is a lack of a comprehensive system on the market that can seamlessly integrate VR technology, deep learning algorithms, bioelectric signal monitoring, and intelligent feedback regulation mechanisms to accurately monitor and adaptively adjust virtual eating scenes, thereby effectively promoting the recovery of swallowing function in patients after tracheotomy. Summary of the invention

[0006] In order to solve the above-mentioned technical problems, the present application is proposed to provide an interactive nursing control method and system for tracheotomy patients that can seamlessly integrate VR technology, deep learning algorithms, bioelectric signal monitoring, and intelligent feedback regulation mechanisms to accurately monitor and adaptively adjust virtual eating scenes, thereby effectively promoting the recovery of swallowing function in patients after tracheotomy.

[0007] In a first aspect, the present invention provides an interactive nursing control method for patients undergoing tracheotomy, comprising: based on virtual reality technology, drawing a simulated eating scene for a user who has undergone tracheotomy, displaying virtual food with preset attributes to the user in the simulated eating scene, and stimulating the user to swallow through the virtual food; collecting bioelectric signals generated by the activities of muscle groups in multiple parts of the user's neck, mandible, tongue and larynx when the user swallows; inputting the attributes of the virtual food and the bioelectric signals corresponding to the user's swallowing action into a trained neural network model, the neural network model is trained based on the attributes of the real food when the sample user swallows the real food and the bioelectric signals generated by the muscle groups of the neck, mandible, tongue and larynx when the sample user swallows, and marking the parts with abnormal muscle group movements when the sample user swallows abnormally; judging whether the user's swallowing action is abnormal through the neural network model, and identifying the abnormal parts when judging that the user's swallowing action is abnormal.

[0008] Optionally, the aforementioned interactive nursing control method for tracheotomy patients, before inputting the bioelectric signal corresponding to the attribute of the virtual food and the swallowing action of the user into the trained neural network model, further comprises: training the neural network model, wherein the loss function used in the training process of the neural network model is , wherein the number of parts of the plurality of parts is n, represents the result of whether the i-th part among the n parts judged by the neural network model is abnormal, Indicates the actual situation of whether the i-th part is abnormal, The weight coefficient representing the degree of harm caused by abnormal muscle group movement in the i-th part, is the bias value.

[0009] Optionally, the aforementioned interactive nursing control method for tracheotomy patients, before inputting the bioelectric signal corresponding to the attribute of the virtual food and the swallowing action of the user into the trained neural network model, further comprises: before training the neural network model, setting a learning rate for any parameter in the neural network model , whose initial value is ; After the After the training iterations, the learning rate of any parameter is adjusted according to the gradient direction of the loss function. Adjust the learning rate after adjustment , where the sqrt function is used to calculate the square root, , is the preset momentum decay rate, and t is the number of iterative training times.

[0010] Optionally, the aforementioned interactive nursing control method for tracheotomy patients, the collecting of bioelectric signals generated by the activities of muscle groups in multiple parts of the user's neck, mandible, tongue and throat when the user performs swallowing, also includes: calculating the deformation effect of the user's swallowing action on the virtual food according to the bioelectric signals corresponding to the user's swallowing action and the properties of the virtual food; calculating the graphics and sounds generated when the virtual food is deformed according to the deformation effect of the user's swallowing action on the virtual food; and providing visual feedback and auditory feedback to the user in the simulated eating scene according to the graphics and sounds generated when the virtual food is deformed, and stimulating the user to produce new swallowing actions through the visual feedback and the auditory feedback.

[0011] Optionally, in the aforementioned interactive nursing control method for tracheotomy patients, the attributes of the virtual food include the viscosity of the virtual food, and the display of virtual food with preset attributes to the user in the simulated eating scene includes: counting the time of multiple historical swallowing actions completed by the user, and calculating the average swallowing time T of the user; and calculating the viscosity of the virtual food according to the average swallowing time of the user. ,in, , is the preset empirical coefficient, the length of time that the tracheotomy operation has been performed for the user up to the present time, The average length of recovery time after performing the tracheotomy procedure for the sample users.

[0012] Optionally, in the aforementioned interactive nursing control method for tracheotomy patients, the attributes of the virtual food include the hardness of the virtual food, and the display of virtual food with preset attributes to the user in the simulated eating scene includes: counting the historical bioelectric signals generated by the activities of the muscle groups in the multiple parts when the user completes multiple historical swallowing actions, and calculating the average intensity values ​​of the historical bioelectric signals of the multiple parts of the user; calculating the average swallowing strength of the user's multiple historical swallowing actions based on the average intensity values ​​of the historical bioelectric signals of the multiple parts of the user; and setting the hardness of the virtual food based on the average swallowing strength of the user's multiple historical swallowing actions.

[0013] Optionally, the aforementioned interactive nursing control method for tracheotomy patients further includes: when an abnormal part is identified, detecting whether the strength value of the bioelectric signal corresponding to the abnormal part is within a preset interval; when the judgment result is no, generating abnormal warning information; when the judgment result is yes, collecting multiple basic anatomical structure images of the abnormal part at multiple time points; based on the multiple basic anatomical structure images of the abnormal part, identifying the direction and amplitude of muscle movement at the abnormal part, wherein the direction of the muscle movement , the amplitude of the muscle movement , is the inverse sine function, is the horizontal pixel distance of any anchor point of the muscle in the basic anatomical structure image at two adjacent time points, is the vertical pixel distance of any anchor point of the muscle in the basic anatomical structure image at two adjacent time points, and m is the number of anchor points pre-selected on the muscle; when the direction and amplitude of muscle movement at the abnormal site meet the preset conditions, the abnormal warning information is generated.

[0014] Optionally, the aforementioned interactive care control method for tracheotomy patients, before collecting the bioelectric signals generated by the activities of muscle groups in multiple parts of the user's neck, mandible, tongue and throat when the user performs swallowing, also includes: setting a plurality of self-adsorption electrodes at multiple parts of the user's neck, mandible, tongue and throat to collect bioelectric signals, the adsorption pressure of the plurality of self-adsorption electrodes affecting the intensity of the collected bioelectric signals; detecting the adsorption pressure of the plurality of self-adsorption electrodes, and detecting the intensity of the bioelectric signals collected by the plurality of self-adsorption electrodes; adjusting the adsorption pressure of the plurality of self-adsorption electrodes so that the ratio of the sum of the adsorption pressures of the plurality of self-adsorption electrodes to the sum of the intensities of the collected bioelectric signals is minimized.

[0015] In a second aspect, the present invention provides an interactive nursing control system for patients undergoing tracheotomy, comprising: a virtual reality module, which, based on virtual reality technology, draws a simulated eating scene for a user who has undergone tracheotomy, displays virtual food with preset attributes to the user in the simulated eating scene, and stimulates the user to swallow through the virtual food; a signal acquisition module, which acquires bioelectric signals generated by the activities of muscle groups in multiple parts of the user's neck, mandible, tongue and larynx when the user swallows; a signal input module, which inputs the attributes of the virtual food and the bioelectric signals corresponding to the user's swallowing action into a trained neural network model, the neural network model is trained based on the attributes of the real food when the sample user swallows the real food and the bioelectric signals generated by the muscle groups of the neck, mandible, tongue and larynx when the sample user swallows, and marks the parts where the muscle groups have abnormal movements when the sample user swallows abnormally; an abnormality recognition module, which determines whether the user's swallowing action is abnormal through the neural network model, and identifies the parts where the abnormality exists when it is determined that the user's swallowing action is abnormal.

[0016] The above one or more technical solutions of the present invention have at least one or more of the following beneficial effects: The present invention provides a realistic eating scene through virtual reality technology, and monitors the patient's physiological reactions during the swallowing process in real time, that is, the bioelectric signals of the muscle group activities in the neck, mandible, tongue and throat, etc., and analyzes and processes a large number of bioelectric signals through a neural network to identify the parts where swallowing abnormalities exist. Compared with traditional rehabilitation methods, the fun and interactive features of virtual reality technology significantly enhance the patient's enthusiasm and participation in rehabilitation training, significantly improve the recognition efficiency and accuracy of swallowing dysfunction, and help shorten the patient's rehabilitation cycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1 A flowchart of an interactive nursing control method for tracheotomy patients according to an embodiment of the present application; Figure 2 A partial flow chart of an interactive nursing control method for tracheotomy patients according to an embodiment of the present application; Figure 3Another partial flow chart of the interactive nursing control method for tracheotomy patients according to an embodiment of the present application; Figure 4 It is another partial flow chart of the interactive nursing control method for tracheotomy patients according to the embodiment of the present application; Figure 5 It is another partial flow chart of the interactive nursing control method for tracheotomy patients according to the embodiment of the present application; Figure 6 The block diagram is a control system for interactive nursing care of patients undergoing tracheotomy surgery according to an embodiment of the present application. DETAILED DESCRIPTION

[0019] Some embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0020] like Figure 1 As shown, in one embodiment of the present invention, a method for interactive nursing control of tracheotomy patients is provided, comprising: Step S110, based on virtual reality technology, a simulated eating scene is drawn for the user who has undergone tracheotomy surgery, virtual food with preset attributes is displayed to the user in the simulated eating scene, and the user is stimulated to swallow through the virtual food.

[0021] In this embodiment, the attributes of virtual food include but are not limited to food type, hardness, color, etc. In this embodiment, a realistic eating scene is provided through virtual reality technology, providing a highly personalized and immersive rehabilitation training environment for patients after tracheotomy. This personalized training can customize the training content according to the actual situation of each patient, significantly improving the pertinence and effectiveness of rehabilitation training.

[0022] Step S120, collecting bioelectric signals generated by the activities of muscle groups in multiple parts of the user's neck, jaw, tongue and throat when the user swallows.

[0023] In this embodiment, the bioelectric signals of muscle group activities in the neck, mandible, tongue and larynx are actually the patient's physiological reactions during swallowing, such as muscle activity intensity and coordination, reflecting the patient's current swallowing function level.

[0024] Specifically, bioelectric signals can be collected in the following ways: (1) Multiple self-adsorption electrodes are arranged at multiple locations of the user's neck, mandible, tongue and throat to collect bioelectric signals, and the adsorption pressure of the multiple self-adsorption electrodes affects the intensity of the collected bioelectric signals. (2) The adsorption pressure of the multiple self-adsorption electrodes is detected, and the intensity of the bioelectric signals collected by the multiple self-adsorption electrodes is detected. (3) The adsorption pressure of the multiple self-adsorption electrodes is adjusted to minimize the ratio of the sum of the adsorption pressures of the multiple self-adsorption electrodes to the sum of the intensities of the collected bioelectric signals.

[0025] In this embodiment, the adsorption pressure of the electrode is adaptively adjusted to ensure good contact between the electrode and the skin, thereby reducing signal attenuation or interference caused by poor contact, and significantly improving the acquisition quality of bioelectric signals (such as muscle electrical signals). The signal obtained in this way is more accurate and stable, providing a more reliable data basis for subsequent analysis and diagnosis. Traditional fixed-pressure electrodes may be too tight or too loose due to individual differences. The former may cause discomfort to the user or even skin damage, while the latter affects the signal quality. The technical solution of this embodiment adjusts the electrode adsorption pressure in a personalized manner so that it is sufficient to maintain stable contact without causing unnecessary pressure to the user.

[0026] Step S130, the properties of the virtual food and the bioelectric signals corresponding to the user's swallowing action are input into a trained neural network model. The neural network model is trained based on the properties of the real food when the sample user swallows the real food and the bioelectric signals generated by the muscle groups of the neck, jaw, tongue and throat when the sample user swallows. When the sample user swallows abnormally, the parts with abnormal muscle group movements are marked.

[0027] In this embodiment, the application of neural networks in the field of medical health has gradually matured, and experiments have shown that the trained neural network model can efficiently and accurately process a large number of bioelectric signals generated by the activities of muscle groups in the neck, jaw, tongue, throat and other parts of the patient when swallowing.

[0028] Step S140, determining whether there is any abnormality in the swallowing action of the user through a neural network model, and identifying the abnormal part when determining that there is any abnormality in the swallowing action of the user.

[0029] Furthermore, this step may also include: (1) When an abnormal part is identified, detect whether the intensity value of the bioelectric signal corresponding to the abnormal part is within a preset range.

[0030] (2) When the judgment result is negative, an abnormal warning message is generated.

[0031] (3) When the judgment result is yes, multiple basic anatomical structure images of the abnormal part at multiple time points are collected. Based on the multiple basic anatomical structure images of the abnormal part, the direction and amplitude of muscle movement at the abnormal part are identified, wherein the direction of muscle movement , the amplitude of the muscle movement , is the inverse sine function, is the horizontal pixel distance of any anchor point of the muscle in the basic anatomical structure image at two adjacent time points, is the vertical pixel distance of any anchor point of the muscle in the basic anatomical structure image at two adjacent time points, and m is the number of anchor points pre-selected on the muscle. When the direction and amplitude of muscle movement at the abnormal part meet the preset conditions, the abnormal warning information is generated.

[0032] In this embodiment, for the abnormal part identified by the neural network, it is further determined whether the bioelectric signal strength value is within a reasonable range. If it is outside the reasonable range, a warning message is generated to prompt medical staff to further ensure the accuracy of the prompt information; if it is within the reasonable range, a more specific analysis is performed, and the basic anatomical structure images in time sequence are accurately collected according to the positioning of the abnormal part. Then, multiple continuous basic anatomical structure images show the direction and amplitude of muscle movement at the abnormal part, which can reflect the working condition of the muscles at the abnormal part. If the direction and amplitude of muscle movement at the abnormal part are consistent with the symptoms of dysphagia, a warning message is generated to prompt medical staff.

[0033] According to the technical solution of this embodiment, a realistic eating scene is provided by virtual reality technology, and the patient's physiological reactions during swallowing, that is, the bioelectric signals of the muscle group activities in the neck, mandible, tongue and throat are monitored in real time. A large number of bioelectric signals are analyzed and processed through a neural network to identify the parts where swallowing abnormalities exist. Compared with traditional rehabilitation methods, the fun and interactive features of virtual reality technology have significantly enhanced the enthusiasm and participation of patients in rehabilitation training, significantly improved the recognition efficiency and accuracy of swallowing dysfunction, and helped to shorten the patient's rehabilitation cycle.

[0034] An embodiment of the present invention provides an interactive nursing control method for patients undergoing tracheotomy surgery. Compared with the aforementioned embodiment, the interactive nursing control method for patients undergoing tracheotomy surgery in this embodiment, before step S130, further includes: Train the neural network model. The loss function used by the neural network model during training is , where the number of parts of the multiple parts is n, It indicates whether the i-th part of the n parts is abnormal as judged by the neural network model. Indicates the actual situation of whether the i-th part is abnormal. The weight coefficient that indicates the degree of harm caused by abnormal muscle group movement in the i-th part, is the bias value.

[0035] According to the technical solution of this embodiment, the traditional loss function design method is optimized according to the application scenario of the neural network, further amplifying the impact of incorrect identification of parts that cause greater harm when muscle groups move abnormally, thereby ensuring that the trained neural network can more accurately identify the actual situation of parts that cause greater harm when muscle groups move abnormally during swallowing, so as to protect the patient's life safety.

[0036] like Figure 2 As shown, an embodiment of the present invention provides an interactive nursing control method for patients undergoing tracheotomy surgery. Compared with the aforementioned embodiment, the interactive nursing control method for patients undergoing tracheotomy surgery in this embodiment, before step S130, further includes: Step S210, before training the neural network model, setting a learning rate for any parameter in the neural network model , whose initial value is .

[0037] Step S220: after the tth iteration of training, the learning rate of any parameter is adjusted according to the gradient direction of the aforementioned loss function. Adjust the learning rate after adjustment , where the sqrt function is used to calculate the square root, , is the preset momentum decay rate, and t is the number of iterative training times.

[0038] According to the technical solution of this embodiment, a new learning rate adaptive adjustment formula for neural network parameters is designed. According to this formula, the parameter adjustment amplitude of the neural network gradually decreases with the increase of the number of training iterations. It is beneficial to speed up the training efficiency of the neural network in the early stage and to accurately determine the parameters of the neural network in the later stage.

[0039] like Figure 3 As shown, an embodiment of the present invention provides an interactive nursing control method for patients undergoing tracheotomy surgery. Compared with the aforementioned embodiment, the interactive nursing control method for patients undergoing tracheotomy surgery in this embodiment further includes: Step S310 , calculating the deformation effect of the user's swallowing action on the virtual food according to the bioelectric signal corresponding to the user's swallowing action and the properties of the virtual food.

[0040] In this embodiment, the changes in virtual food under the patient's swallowing action are calculated based on the patient's swallowing force and speed reflected by the bioelectric signal, as well as the type and hardness of the virtual food. The degree of deformation is referred to the situation of swallowing real food. For example, when swallowing virtual bread, the bread will undergo different degrees of deformation based on the patient's swallowing force and speed, as well as the hardness and viscosity of the bread.

[0041] Step S320 , calculating the graphics and sounds generated when the virtual food is deformed according to the deformation effect of the user's swallowing action on the virtual food.

[0042] In this embodiment, the graphics and sounds generated when the virtual food is deformed may be the changes in the appearance of the food and the swallowing sounds when the patient swallows the food. For example, when the virtual bread is deformed, the shape and color of its surface will change, and at the same time, the sound effects of the bread being swallowed will be generated. Specifically, the color, shape, color, and sound effects of the virtual bread will change depending on the different bioelectric signals of the patient's swallowing and the different properties of the bread.

[0043] Step S330, providing visual feedback and auditory feedback to the user in a simulated eating scene based on the graphics and sounds generated when the virtual food is deformed, and stimulating the user to generate a new swallowing action through the visual feedback and auditory feedback.

[0044] According to the technical solution of this embodiment, based on the bioelectric signals corresponding to the patient's swallowing movements and the properties of the virtual food, the changes in the virtual food when the patient swallows the food are calculated, and the changes are fed back to the patient in visual and auditory forms to guide the patient to continue to generate new swallowing movements.

[0045] like Figure 4 As shown, an embodiment of the present invention provides an interactive nursing control method for patients undergoing tracheotomy surgery. Compared with the aforementioned embodiment, in the interactive nursing control method for patients undergoing tracheotomy surgery of this embodiment, the attributes of the virtual food include the viscosity of the virtual food, and step S110 includes: Step S410 , counting the time of multiple historical swallowing actions completed by the user, and calculating the average swallowing time T of the user.

[0046] In this embodiment, the swallowing time of the user reflects the swallowing ability of the user to a certain extent.

[0047] Step S420, calculating the viscosity of the virtual food according to the average swallowing time of the user ,in, , is the preset empirical coefficient, The length of time since the user has had a tracheotomy, Average length of recovery time after tracheotomy surgery for sample users.

[0048] According to the technical solution of this embodiment, the viscosity of the virtual food is adjusted according to the swallowing ability of the patient to guide the patient to swallow with an appropriate force to avoid harm to the patient due to excessive or insufficient swallowing force.

[0049] like Figure 5 As shown, an embodiment of the present invention provides an interactive nursing control method for patients undergoing tracheotomy surgery. Compared with the aforementioned embodiment, in the interactive nursing control method for patients undergoing tracheotomy surgery of this embodiment, the attribute of the virtual food includes the hardness of the virtual food, and step S510 includes: Step S510, counting historical bioelectric signals generated by muscle group activities of multiple parts of the user when the user completes multiple historical swallowing actions, and calculating the average intensity values ​​of the historical bioelectric signals of multiple parts of the user.

[0050] Step S520, calculating the average swallowing strength of multiple historical swallowing actions of the user according to the historical average strength values ​​of the bioelectric signals of multiple parts of the user.

[0051] In this embodiment, the average intensity value of the historical bioelectric signals of multiple parts of the patient actually reflects the swallowing strength of the patient.

[0052] Step S530, setting the hardness of the virtual food according to the average swallowing strength of multiple historical swallowing actions of the user.

[0053] According to the technical solution of this embodiment, the swallowing strength of the user is judged from the patient's historical swallowing data, so as to reasonably set the hardness of the virtual food to avoid harm to the patient due to excessively high or low swallowing strength.

[0054] like Figure 6 As shown, in one embodiment of the present invention, an interactive nursing control system for tracheotomy patients is provided, comprising: The virtual reality module 610 draws a simulated eating scene for the user who has undergone tracheotomy based on virtual reality technology, displays virtual food with preset attributes to the user in the simulated eating scene, and stimulates the user to swallow through the virtual food.

[0055] In this embodiment, the attributes of virtual food include but are not limited to food type, hardness, color, etc. In this embodiment, a realistic eating scene is provided through virtual reality technology, providing a highly personalized and immersive rehabilitation training environment for patients after tracheotomy. This personalized training can customize the training content according to the actual situation of each patient, significantly improving the pertinence and effectiveness of rehabilitation training.

[0056] The signal acquisition module 620 collects bioelectric signals generated by the activities of muscle groups in multiple parts of the user's neck, jaw, tongue and throat when the user swallows.

[0057] In this embodiment, the bioelectric signals of muscle group activities in the neck, mandible, tongue and larynx are actually the patient's physiological reactions during swallowing, such as muscle activity intensity and coordination, reflecting the patient's current swallowing function level.

[0058] The signal input module 630 inputs the bioelectric signals corresponding to the properties of the virtual food and the user's swallowing action into a trained neural network model. The neural network model is trained based on the properties of the real food when the sample user swallows the real food and the bioelectric signals generated by the muscle groups of the neck, jaw, tongue and throat when the sample user swallows. When the sample user swallows abnormally, the parts with abnormal muscle group movements are marked.

[0059] In this embodiment, the application of neural networks in the field of medical health has gradually matured, and experiments have shown that the trained neural network model can efficiently and accurately process a large number of bioelectric signals generated by the activities of muscle groups in the neck, jaw, tongue, throat and other parts of the patient when swallowing.

[0060] The abnormality identification module 640 determines whether there is an abnormality in the swallowing action of the user through a neural network model, and identifies the abnormal part when it is determined that there is an abnormality in the swallowing action of the user.

[0061] Furthermore, this step may also include: (1) When an abnormal part is identified, detect whether the intensity value of the bioelectric signal corresponding to the abnormal part is within a preset range.

[0062] (2) When the judgment result is negative, an abnormal warning message is generated.

[0063] (3) When the judgment result is yes, multiple basic anatomical structure images of the abnormal part at multiple time points are collected. Based on the multiple basic anatomical structure images of the abnormal part, the direction and amplitude of muscle movement at the abnormal part are identified, wherein the direction of muscle movement , the amplitude of the muscle movement , is the inverse sine function, is the horizontal pixel distance of any anchor point of the muscle in the basic anatomical structure image at two adjacent time points, is the vertical pixel distance of any anchor point of the muscle in the basic anatomical structure image at two adjacent time points, and m is the number of anchor points pre-selected on the muscle. When the direction and amplitude of muscle movement at the abnormal part meet the preset conditions, the abnormal warning information is generated.

[0064] In this embodiment, for the abnormal part identified by the neural network, it is further determined whether the bioelectric signal strength value is within a reasonable range. If it is outside the reasonable range, a warning message is generated to prompt medical staff to further ensure the accuracy of the prompt information; if it is within the reasonable range, a more specific analysis is performed, and the basic anatomical structure images in time sequence are accurately collected according to the positioning of the abnormal part. Then, multiple continuous basic anatomical structure images show the direction and amplitude of muscle movement at the abnormal part, which can reflect the working condition of the muscles at the abnormal part. If the direction and amplitude of muscle movement at the abnormal part are consistent with the symptoms of dysphagia, a warning message is generated to prompt medical staff.

[0065] According to the technical solution of this embodiment, a realistic eating scene is provided by virtual reality technology, and the patient's physiological reactions during swallowing, that is, the bioelectric signals of the muscle group activities in the neck, mandible, tongue and throat are monitored in real time. A large number of bioelectric signals are analyzed and processed through a neural network to identify the parts where swallowing abnormalities exist. Compared with traditional rehabilitation methods, the fun and interactive features of virtual reality technology have significantly enhanced the enthusiasm and participation of patients in rehabilitation training, significantly improved the recognition efficiency and accuracy of swallowing dysfunction, and helped to shorten the patient's rehabilitation cycle.

[0066] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0067] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0068] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0069] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0070] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. An interactive nursing control method for patients undergoing tracheotomy, characterized in that: include: Based on virtual reality technology, a simulated eating scene is drawn for a user who has undergone a tracheotomy operation, virtual food with preset attributes is displayed to the user in the simulated eating scene, and the user is stimulated to swallow through the virtual food; collecting bioelectric signals generated by muscle group activities at multiple locations of the user's neck, jaw, tongue, and throat when the user performs a swallowing action; Inputting the properties of the virtual food and the bioelectric signals corresponding to the swallowing action of the user into a trained neural network model, wherein the neural network model is trained based on the properties of the real food when the sample user swallows the real food and the bioelectric signals generated by the muscle groups of the neck, mandible, tongue and throat when the sample user swallows, and marking the parts of the muscle group with abnormal movements when the sample user swallows abnormally; The neural network model is used to determine whether the swallowing action of the user is abnormal, and when it is determined that the swallowing action of the user is abnormal, the abnormal part is identified.

2. The interactive nursing control method for tracheotomy patients according to claim 1, characterized in that: Before inputting the bioelectric signal corresponding to the attribute of the virtual food and the swallowing action of the user into the trained neural network model, the method further includes: Train the neural network model, and the loss function used in the training process of the neural network model is , wherein the number of parts of the plurality of parts is n, represents the result of whether the i-th part among the n parts judged by the neural network model is abnormal, Indicates the actual situation of whether the i-th part is abnormal, The weight coefficient representing the degree of harm caused by abnormal muscle group movement in the i-th part, is the bias value.

3. The interactive nursing control method for tracheotomy patients according to claim 2, characterized in that: Before inputting the bioelectric signal corresponding to the attribute of the virtual food and the swallowing action of the user into the trained neural network model, the method further includes: Before training the neural network model, set a learning rate for any parameter in the neural network model , whose initial value is ; After the tth iteration of training, the learning rate of any parameter is adjusted according to the gradient direction of the loss function. Adjust the learning rate after adjustment , where the sqrt function is used to calculate the square root, , is the preset momentum decay rate, and t is the number of iterative training times.

4. The interactive nursing control method for tracheotomy patients according to claim 1, characterized in that: Collecting bioelectric signals generated by muscle group activities at multiple locations of the user's neck, jaw, tongue, and throat when the user performs a swallowing action, further comprising: Calculating the deformation effect of the user's swallowing action on the virtual food according to the bioelectric signal corresponding to the user's swallowing action and the properties of the virtual food; Calculating graphics and sounds generated when the virtual food is deformed according to the deformation effect of the swallowing action of the user on the virtual food; According to the graphics and sounds generated when the virtual food is deformed, visual feedback and auditory feedback are provided to the user in the simulated eating scene, and the user is stimulated to produce a new swallowing action through the visual feedback and the auditory feedback.

5. The interactive nursing control method for tracheotomy patients according to claim 1, characterized in that: The attributes of the virtual food include the viscosity of the virtual food, and displaying the virtual food with preset attributes to the user in the simulated eating scene includes: Counting the duration of multiple historical swallowing actions completed by the user, and calculating the average swallowing time T of the user; Calculate the viscosity of the virtual food based on the average swallowing time of the user ,in, , is the preset empirical coefficient, the length of time that the tracheotomy operation has been performed for the user up to the present time, The average length of recovery time after performing the tracheotomy procedure for the sample users.

6. The interactive nursing control method for tracheotomy patients according to claim 1, characterized in that: The attribute of the virtual food includes the hardness of the virtual food, and displaying the virtual food with preset attributes to the user in the simulated eating scene includes: Counting historical bioelectric signals generated by muscle group activities of the multiple parts when the user completes multiple historical swallowing actions, and calculating average intensity values ​​of the historical bioelectric signals of the multiple parts of the user; Calculating an average swallowing strength of multiple historical swallowing actions of the user according to the historical average strength values ​​of the bioelectric signals of the multiple parts of the user; The hardness of the virtual food is set according to an average swallowing strength of multiple historical swallowing actions of the user.

7. The interactive nursing control method for tracheotomy patients according to claim 1, characterized in that: Also includes: When an abnormal part is identified, detecting whether the strength value of the bioelectric signal corresponding to the abnormal part is within a preset range; When the judgment result is no, an abnormal warning message is generated; When the judgment result is yes, a plurality of basic anatomical structure images of the abnormal part at a plurality of time points are collected; According to a plurality of basic anatomical structure images of the abnormal part, the direction and amplitude of muscle movement at the abnormal part are identified, wherein the direction of the muscle movement , the amplitude of the muscle movement , is the inverse sine function, is the horizontal pixel distance of any anchor point of the muscle in the basic anatomical structure image at two adjacent time points, is the vertical pixel distance of any anchor point of the muscle in the basic anatomical structure image at two adjacent time points, and m is the number of anchor points pre-selected on the muscle; When the direction and amplitude of the muscle movement at the abnormal part meet the preset conditions, the abnormal warning information is generated.

8. The interactive nursing control method for tracheotomy patients according to claim 1, characterized in that: Before collecting the bioelectric signals generated by the muscle group activities at multiple parts of the user's neck, jaw, tongue and throat when the user performs a swallowing action, the method further includes: A plurality of self-adsorption electrodes are arranged at a plurality of locations on the neck, mandible, tongue and throat of the user to collect bioelectric signals, wherein the adsorption pressure of the plurality of self-adsorption electrodes affects the intensity of the collected bioelectric signals; detecting the adsorption pressure of the plurality of self-adsorption electrodes, and detecting the intensity of bioelectric signals collected by the plurality of self-adsorption electrodes; The adsorption pressures of the plurality of self-adsorption electrodes are adjusted so that the ratio of the sum of the adsorption pressures of the plurality of self-adsorption electrodes to the sum of the strengths of the collected bioelectric signals is minimized.

9. An interactive nursing control system for patients undergoing tracheotomy, characterized in that: include: A virtual reality module, based on virtual reality technology, draws a simulated eating scene for a user who has undergone a tracheotomy, displays virtual food with preset attributes to the user in the simulated eating scene, and stimulates the user to swallow through the virtual food; A signal acquisition module collects bioelectric signals generated by muscle group activities at multiple locations of the user's neck, jaw, tongue, and throat when the user performs a swallowing action; a signal input module, inputting the properties of the virtual food and the bioelectric signals corresponding to the swallowing action of the user into a trained neural network model, wherein the neural network model is trained based on the properties of the real food when the sample user swallows the real food and the bioelectric signals generated by the muscle groups of the neck, mandible, tongue and larynx when the sample user swallows, and marking the parts of the muscle groups with abnormal movements when the sample user swallows abnormally; The abnormality identification module determines whether the swallowing action of the user is abnormal through the neural network model, and identifies the abnormal part when it is determined that the swallowing action of the user is abnormal.

Images