Body position adjusting method based on coma patient body position fixing device

Abstract

The embodiment of the invention relates to the field of body position adjustment, in particular to a body position adjustment method based on an unconscious patient body position fixing device. A specific embodiment of the method comprises the following steps: controlling a body position camera device to collect body three-dimensional data of a target user; controlling an angle adjusting device to adjust the body posture of the target user; controlling a pressure sensor to collect the adjusted body pressure value set of the target user; controlling a body position camera device to detect body position information of the target user in real time; inputting the body pose information into a pre-trained pose state recognition model to obtain a pose state recognition result; controlling a voice playing device to play the abnormal pose voice; controlling a physical sign detection device to detect physical sign data of the target user; performing identification processing on the physical sign data to generate a physical sign identification result; and controlling the voice playing device to play the sign recognition result. According to the embodiment, the body position angle of the target user can be accurately controlled, and the stability of the target user can be kept.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of body position adjustment, and in particular to a body position adjustment method based on a body position fixing device for a comatose patient. Background Art

[0002] Since comatose patients in neurosurgery are in a coma for a long time and have no autonomous consciousness, their treatment and care are completely dependent on medical staff. Such patients are often accompanied by symptoms such as hypoproteinemia and systemic edema, which increases the risk of bedsores and may cause pleural effusion, ascites and pulmonary dysfunction. Therefore, it is necessary to adjust the body posture of the target user (coma patient) on time. At present, when adjusting the body posture, the usual method is: medical staff manually adjust the body posture of the target user.

[0003] However, it is found in practice that when the above method is used to adjust the body position, the following technical problems often occur: First, it is difficult to accurately control the body angle of the target user and to keep the target user stable by manually adjusting the body posture; Second, the target user's body position is not recognized after the adjustment, so it is difficult to determine whether the target user is in a side-lying state after the body position adjustment. When the target user is not in a side-lying state, it is easy to increase the target user's cardiac burden; In the process of adopting technical solutions to solve the above-mentioned technical problem one, the following problems often occur: the target user's pressure sore condition is not detected, which makes it difficult to treat the target user's existing pressure sores. In response to the above problems, the conventional solution is generally: medical staff observe the pressure sore condition of the target user through observation. However, the above-mentioned conventional solution still has the following technical problem three: the observation method relies on the experience of medical staff, making it difficult to observe potential pressure sore lesions, and it is difficult to judge the severity of the pressure sore site, resulting in low accuracy of the observed pressure sore condition, which is not conducive to the treatment and recovery of the target user. Summary of the invention

[0004] 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.

[0005] Some embodiments of the present application propose a body position adjustment method based on a body position fixation device for a comatose patient, a computer device, and a computer-readable storage medium to solve one or more of the technical problems mentioned in the above background technology section.

[0006] In a first aspect, some embodiments of the present application provide a method for adjusting the body position based on a body position fixing device for a comatose patient, which is applied to the body position fixing device for a comatose patient. The body position fixing device for a comatose patient comprises: a body position camera device, a pressure sensor, an angle adjustment device, a vital sign detection device, and a voice playback device. The method comprises: controlling the body position camera device to collect three-dimensional body data of a target user; based on the three-dimensional body data, controlling the angle adjustment device to adjust the body posture of the target user; controlling the pressure sensor to collect a body pressure value set of the target user after adjustment; and in response to determining that the body pressure value set meets a preset abnormality, Under normal pressure conditions, the body posture camera device is controlled to detect the body posture information of the target user in real time; the body posture information is input into a pre-trained posture state recognition model to obtain a posture state recognition result; in response to determining that the posture state recognition result meets a preset abnormal posture state condition, the voice playback device is controlled to play abnormal posture voice; the vital sign detection device is controlled to detect the vital sign data of the target user; the vital sign data is recognized and processed to generate a vital sign recognition result; in response to determining that the vital sign recognition result meets the preset abnormal vital sign condition, the voice playback device is controlled to play the vital sign recognition result.

[0007] 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.

[0008] 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.

[0009] The above-mentioned embodiments of the present application have the following beneficial effects: through the body position adjustment method based on the coma patient body position fixing device of some embodiments of the present application, the body position angle of the target user can be accurately controlled and the target user can be kept stable. Specifically, the reason why it is difficult to accurately control the body position angle of the target user and to keep the target user stable is that it is difficult to accurately control the body position angle of the target user and to keep the target user stable by manually adjusting the body posture. Based on this, the body position adjustment method based on the coma patient body position fixing device of some embodiments of the present application, first, controls the above-mentioned body position camera device to collect the three-dimensional body data of the target user. Thus, the three-dimensional body data can be collected so that the body posture of the target user can be adjusted by the angle adjustment device later. Secondly, based on the above-mentioned three-dimensional body data, the above-mentioned angle adjustment device is controlled to adjust the body posture of the above-mentioned target user. Thus, the target user can be adjusted from a lying state to a side-lying state by the angle adjustment device. Next, the above-mentioned pressure sensor is controlled to collect the adjusted body pressure value set of the target user. Thus, the body pressure values ​​of each contact point between the target user and the angle adjustment device can be collected. Next, in response to determining that the above-mentioned body pressure value set meets the preset abnormal pressure condition, the above-mentioned body position camera device is controlled to detect the body posture information of the above-mentioned target user in real time. Therefore, when the body pressure value set meets the preset abnormal pressure condition, it indicates that the body position of the target user is not correct, and it is necessary to further detect the body posture information through the body position camera device in order to subsequently determine whether an abnormal body position occurs. Then, the above-mentioned body posture information is input into the pre-trained posture state recognition model to obtain the posture state recognition result. Therefore, the posture state recognition result that indicates whether the target user is in a normal body position (side-lying state) can be identified through the posture state recognition model. Then, in response to determining that the above-mentioned posture state recognition result meets the preset abnormal posture state condition, the above-mentioned voice playback device is controlled to play abnormal posture voice. Therefore, when the posture state recognition result indicates that the target user is in an abnormal body position, the voice playback device can be controlled to play so as to adjust the body posture of the target user again. Afterwards, the above-mentioned physical sign detection device is controlled to detect the physical sign data of the above-mentioned target user. Therefore, the physical sign data of the target user can be detected by the physical sign detection device so as to subsequently identify whether the target user has pressure sores. After that, the above-mentioned physical sign data is processed for identification to generate a physical sign identification result. Thus, the pressure sore information present in the physical sign data can be identified through grayscale processing, equalization processing, noise reduction processing and threshold segmentation processing. Finally, in response to determining that the above-mentioned physical sign identification result meets the preset abnormal physical sign condition, the above-mentioned voice playback device is controlled to play the above-mentioned physical sign identification result. Thus, the physical sign identification result representing the pressure sore information can be played so as to subsequently treat the pressure sore.Therefore, by adjusting the body posture of the target user through the angle adjustment device included in the vital sign adjustment system instead of manual adjustment, the body angle of the target user can be accurately controlled and the stability of the target user can be maintained. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] 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 reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0011] Figure 1 is a flow chart of some embodiments of a body position adjustment method based on a body position fixing device for a comatose patient according to the present application; Figure 2 It is a structural schematic diagram of some embodiments of the turning over adjustment device of the body position fixing device for a comatose patient according to the present application; Figure 3 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 process 100 of some embodiments of the method for adjusting the body position of a comatose patient based on the body position fixing device according to the present application is shown. The method for adjusting the body position of a comatose patient based on the body position fixing device is applied to the body position fixing device for a comatose patient, and the body position fixing device for a comatose patient includes: a body position camera device, a pressure sensor, an angle adjustment device, a vital sign detection device, and a voice playback device, including the following steps: Step 101: Control a body camera device to collect three-dimensional body data of a target user.

[0019] In some embodiments, the execution subject of the domain name anomaly detection method based on the domain name detection model can control the above-mentioned body position camera device to collect the three-dimensional body data of the target user. Among them, the target user may be a comatose patient who has no autonomous consciousness but needs to turn over. The three-dimensional body data may include but is not limited to at least one of the following: three-dimensional coordinates of the head (the coordinates of the head of the target user in the three-dimensional coordinate system), three-dimensional coordinates of the left shoulder (the coordinates of the left shoulder of the target user in the three-dimensional coordinate system), and three-dimensional coordinates of the right shoulder (the coordinates of the right shoulder of the target user in the three-dimensional coordinate system). The body position camera device may be a camera for collecting the three-dimensional body data of the target user. The pressure sensor may be a sensor for collecting the pressure value when the target user lies down. The angle adjustment device may be a device for adjusting the body posture of the target user. The physical sign detection device may be a device for detecting the physical sign data of the target user. The voice playback device may be a speaker for playing voice. Among them, the above-mentioned physical sign data may include but is not limited to: a three-dimensional physical sign picture set and a physical sign infrared picture. Here, the three-dimensional physical sign picture in the three-dimensional physical sign picture set may be a picture representing the body part of the target user. The physical sign infrared picture may be an infrared picture representing the temperature.

[0020] Step 102: Based on the three-dimensional body data, control the angle adjustment device to adjust the body posture of the target user.

[0021] In some embodiments, based on the three-dimensional body data, the execution entity may control the angle adjustment device to adjust the body posture of the target user. The angle adjustment device may include: a turning-over adjustment device, a soft pillow placement device, and a head fixing device. Here, the turning-over adjustment device may be a device for adjusting the target user from a lying state to a side-lying state. The soft pillow placement device may be a device (e.g., a robotic arm) for placing a soft pillow behind and in front of the chest of the target user. The head fixing device may be a device for fixing the head of the target user. For example, the turning-over adjustment device may refer to Figure 2The structure schematic diagram of some embodiments of the turning adjustment device of the comatose patient position fixing device according to the present application is shown. Figure 2 As shown, the turning over adjustment device includes: a first motor 21, a second motor 22, a third motor 23, a first rotating shaft 24, a second rotating shaft 25, a first turning over board 26, a second turning over board 27, and a third turning over board 28. Among them, the first motor 21 is used to raise the first turning over board 26. The second motor 22 is used to raise the second turning over board 27. The third motor 23 is used to raise the third turning over board 27. The first rotating shaft 24 is used to connect the first turning over board 26 and the second turning over board 27. The second rotating shaft 25 is used to connect the second turning over board 27 and the third turning over board 28. For example, the turning over adjustment device can control the first motor 21 and the second motor 22 to start, so as to raise the first turning over board 26 and the second turning over board 27, so that the target user completes the turning over from left to right. The turning over adjustment device can also control the second motor 22 and the third motor 23 to start, so as to raise the second turning over board 27 and the third turning over board 28, so that the target user completes the turning over from right to left.

[0022] In practice, based on the three-dimensional body data, the execution subject may control the angle adjustment device to adjust the body posture of the target user through the following steps: The first step is to control the turning over adjustment device to adjust the body posture of the target user based on the above-mentioned three-dimensional body data. In practice, in response to determining that the above-mentioned three-dimensional body data meets the preset adjustment conditions, the turning over adjustment device is controlled to adjust the body posture of the target user from the first state to the second state. For example, the first state may be a lying state. The second state may be a side-lying state. Here, the preset adjustment condition may be that the above-mentioned three-dimensional body data represents that the target user is in a lying state. For example, the target user being in a lying state may be represented by: the ordinate of the three-dimensional coordinate of the head, the ordinate of the three-dimensional coordinate of the left shoulder, and the ordinate of the three-dimensional coordinate of the right shoulder included in the three-dimensional body data are all the same.

[0023] The second step is to control the soft pillow placement device to place the first target soft pillow at the first designated position of the target user. The first target soft pillow may be a soft pillow placed behind the target user. For example, the first designated position may be behind the target user.

[0024] Therefore, a soft pillow is placed behind the target user to facilitate the placement of the arm at a suitable angle and ensure the integrity of the skin.

[0025] The third step is to control the soft pillow placement device to place the second target soft pillow at the second designated position of the target user. The second target soft pillow may be a soft pillow for placement on the chest of the target user. For example, the second designated position may be on the chest of the target user.

[0026] Therefore, placing a soft pillow on the target user's chest can disperse pressure, increase comfort, and assist recovery.

[0027] The fourth step is to control the head fixing device to fix the head of the target user.

[0028] In this way, the head of the target user can be fixed to protect the brain nerves.

[0029] Step 103: Control the pressure sensor to collect the adjusted body pressure value set of the target user.

[0030] In some embodiments, the execution subject may control the pressure sensor to collect the adjusted body pressure value set of the target user, wherein the body pressure value in the body pressure value set may represent the pressure value of the contact point between the target user and the body position adjustment device.

[0031] Step 104 : in response to determining that the body pressure value set meets the preset abnormal pressure condition, controlling the body position camera device to detect the body position information of the target user in real time.

[0032] In some embodiments, the execution subject may control the body position camera device to detect the body position information of the target user in real time in response to determining that the body pressure value set satisfies a preset abnormal pressure condition. The preset abnormal pressure condition may be that the body pressure value set contains a body pressure value greater than a first preset pressure value, or the body pressure value set contains a body pressure value less than a second preset pressure value. For example, the first preset pressure value may be 100. The second preset pressure value may be 0.

[0033] In practice, in response to determining that the body pressure value set meets the preset abnormal pressure condition, the execution subject may control the body position camera device to detect the body posture information of the target user in real time through the following steps: The first step is to control the above-mentioned body position camera device to take a sequence of initial body state pictures of the above-mentioned target user. In practice, the above-mentioned execution subject can control the above-mentioned body position camera device to take the initial body state pictures of the above-mentioned target user at each time granularity within a preset time period to obtain a sequence of initial body state pictures. Among them, the initial body state pictures in the sequence of initial body state pictures can represent the pictures of the target user taken by the body position camera device. For example, the preset time period can be a time period from one minute before the current time to the current time. The time granularity can be one second.

[0034] In the second step, each initial body state picture in the above-mentioned initial body state picture sequence is preprocessed to generate a body state picture, thereby obtaining a body state picture sequence. In practice, the above-mentioned execution subject may preprocess each initial body state picture in the above-mentioned initial body state picture sequence by a preset noise reduction algorithm to generate a body state picture, thereby obtaining a body state picture sequence. For example, the preset noise reduction algorithm may be: a Gaussian filtering algorithm.

[0035] The third step is to perform feature extraction processing on each body state picture in the above body state picture sequence to generate a body feature information group and obtain a body feature information group sequence. In practice, the above execution subject can perform feature extraction processing on each body state picture in the above body state picture sequence through a preset feature extraction algorithm to generate a body feature information group and obtain a body feature information group sequence. For example, the preset feature extraction algorithm can be: SIFT (Scale-invariant feature transform, scale-invariant feature transformation algorithm).

[0036] The fourth step is to vectorize each body feature information in the above body feature information group sequence to generate a body feature vector and obtain a body feature vector group sequence. In practice, the above execution subject can vectorize each body feature information in the above body feature information group sequence by using a preset vectorization algorithm to generate a body feature vector and obtain a body feature vector group sequence. For example, the preset vectorization algorithm can be: One-Hot encoding algorithm.

[0037] Step 5: For each body feature vector group in the body feature vector group sequence, in response to determining that the body feature vector group satisfies a first preset vector condition, a target body feature vector group is determined based on the body feature vector group. The first preset vector condition may be that the body feature vector group is the first body feature vector group in the body feature vector group sequence.

[0038] In the sixth step, each generated target body feature vector group is determined as a target body feature vector group sequence.

[0039] In the seventh step, the last target body feature vector group in the above target body feature vector group sequence is determined as the body posture information.

[0040] In practice, based on the above body feature vector group, the above execution subject can determine the target body feature vector group through the following steps: In the first step, in response to determining that the above-mentioned body feature vector group satisfies the second preset vector condition, a first projection processing is performed on the first body feature vector group in the above-mentioned body feature vector group sequence to generate a target body feature vector group. In practice, in response to determining that the above-mentioned body feature vector group satisfies the second preset vector condition, the above-mentioned execution subject may perform a first projection processing on the first body feature vector group in the above-mentioned body feature vector group sequence by a preset projection algorithm to generate a target body feature vector group. The second preset vector condition may be: the body feature vector group is the second body feature vector group in the body feature vector group sequence. For example, the preset projection algorithm may be a map projection algorithm.

[0041] In the second step, in response to determining that the body feature vector group does not satisfy the second preset vector condition, a second projection process is performed on the body feature vector group to generate a target body feature vector group.

[0042] In practice, in response to determining that the body feature vector group does not satisfy the second preset vector condition, the execution subject may perform a second projection process on the body feature vector group through the following sub-steps to generate a target body feature vector group: In a first sub-step, the previous body feature vector group of the body feature vector group in the body feature vector group sequence is determined as an initial body feature vector group.

[0043] The second sub-step is to screen the above-mentioned initial body feature vector group to generate a body feature screening vector group. In practice, the above-mentioned execution entity can determine each initial body feature vector in the above-mentioned initial body feature vector group that meets the preset screening condition as the body feature screening vector group. Among them, the preset screening condition can be that the body feature confidence corresponding to the initial body feature vector in the initial body feature vector group is greater than the preset confidence. For example, the preset confidence can be 0.8. Here, the confidence corresponding to the initial body feature vector in the initial body feature vector group can be: obtained by inputting the initial body feature vector in the initial body feature vector group into the classification model. For example, the classification model can be: SVM (Support Vector Machine) model.

[0044] The third sub-step is to project the body feature screening vector group to generate a body feature projection vector group. In practice, the execution subject may project the body feature screening vector group to generate a body feature projection vector group using the preset projection algorithm.

[0045] The fourth sub-step is to construct an initial three-dimensional query vector group based on the above-mentioned body feature projection vector group. In practice, the above-mentioned execution subject can randomly generate an initial three-dimensional query vector group with the same number of body feature projection vectors as the above-mentioned body feature projection vector group. For example, the initial three-dimensional query vector in the initial three-dimensional query vector group can be a sparse matrix. Here, the initial three-dimensional query vector in the initial three-dimensional query vector group can correspond to the body feature projection vector in the body feature projection vector group.

[0046] The fifth sub-step is to merge the body feature projection vector group and the initial three-dimensional query vector group to generate a three-dimensional query vector group. In practice, first, for each body feature projection vector in the body feature projection vector group, the execution subject may merge the body feature projection vector and the initial three-dimensional query vector corresponding to the body feature projection vector in the initial three-dimensional query vector group to generate a three-dimensional query vector, thereby obtaining a three-dimensional query vector group.

[0047] The sixth sub-step is to perform deduplication processing on the above three-dimensional query vector group to generate a target three-dimensional query vector group. In practice, the above execution subject can input the above three-dimensional query vector group into a self-attention model to obtain a target three-dimensional query vector group. For example, the self-attention model can be a Transformer (neural network model based on attention mechanism) model.

[0048] The seventh sub-step is to perform spatial projection processing on the above body feature vector group to generate a body three-dimensional feature vector group. In practice, the above execution subject can perform spatial projection processing on the above body feature vector group through the above preset projection algorithm to generate a body three-dimensional feature vector group.

[0049] In the eighth sub-step, the target three-dimensional query vector group and the body three-dimensional feature vector group are input into an interactive attention model to obtain a target body feature vector group. For example, the interactive attention model may be a cross-attention model.

[0050] Step 105, inputting the body posture information into a pre-trained posture state recognition model to obtain a posture state recognition result.

[0051] In some embodiments, the execution subject may input the body posture information into a pre-trained posture state recognition model to obtain a posture state recognition result. The posture state recognition model may be a pre-trained neural network model that takes body posture information as input and takes posture state recognition results as output. The posture state recognition result may be, but is not limited to, a first recognition result and a second recognition result. For example, the first recognition result may indicate that the target user is in a side-lying state. The second recognition result may indicate that the target user is not in a side-lying state.

[0052] Optionally, before inputting the body posture information into a pre-trained posture state recognition model to obtain a posture state recognition result, the method further includes: The first step is to obtain the training sample set.

[0053] In some embodiments, the execution subject may obtain a training sample set, wherein the training samples in the training sample set may include: sample body posture information and sample posture state recognition results.

[0054] The third step is to determine the initial posture state recognition model.

[0055] In some embodiments, the execution subject may determine an initial posture state recognition model, wherein the initial posture state recognition model may include: an initial classification model, an initial regression model, and an initial recognition model.

[0056] The initial classification model may be a neural network model that takes the sample body posture information as input and takes the initial posture category information as output. For example, the initial classification model may be an SVM (Support Vector Machine) model. The initial posture category information may be first category information and second category information. For example, the first category information may indicate that the target user is in a lying state. The second category information may indicate that the target user is not in a lying state.

[0057] The initial regression model may be a neural network model that takes the sample body posture information as input and the initial posture regression information as output. For example, the initial regression model may be a logistic regression model. The initial posture regression information may represent the target user's three-dimensional coordinate information. The target user's three-dimensional coordinate information may include but is not limited to at least one of the following: left shoulder three-dimensional coordinates (the coordinates of the target user's left shoulder in the three-dimensional coordinate system), right shoulder three-dimensional coordinates (the coordinates of the target user's right shoulder in the three-dimensional coordinate system) The initial recognition model may be a model that takes the initial posture category information and the initial posture regression information as input and takes the initial posture state recognition result as output. For example, the initial recognition model may be used to: first, in response to determining that the initial posture regression information satisfies the preset regression condition, determine the first preset regression information as the regression classification information. Wherein, the preset regression condition may be: the ordinate of the left shoulder three-dimensional coordinate included in the initial posture regression information is the same as the ordinate of the right shoulder three-dimensional coordinate, and the abscissa of the left shoulder three-dimensional coordinate included in the initial posture regression information is the same as the abscissa of the right shoulder three-dimensional coordinate. The first preset regression information may characterize that the side lying angle of the target user is 90 degrees. Secondly, in response to determining that the initial posture regression information does not meet the above preset regression condition, determine the second preset regression information as the regression classification information. Wherein, the second preset regression information may characterize that the side lying angle of the target user is not 90 degrees. Next, in response to determining that the initial posture category information and the initial posture regression information meet the preset posture condition, determine the first recognition result as the initial posture state recognition result. The preset posture condition may be: the initial posture category information is the second category information and the initial posture regression information is the first preset regression information. Then, in response to determining that the initial posture category information and the initial posture regression information do not meet the preset posture condition, the second recognition result is determined as the initial posture state recognition result. The fourth step is to select target training samples from the above training sample set.

[0058] In some embodiments, the execution entity may select a target training sample from the training sample set. In practice, the execution entity may randomly select a training sample from the training sample set as the target training sample.

[0059] The fifth step is to input the sample body posture information included in the selected target training sample into the initial classification model to obtain the initial posture category information.

[0060] In some embodiments, the execution entity may input the sample body posture information included in the selected target training sample into an initial classification model to obtain initial posture category information.

[0061] In the sixth step, the sample body posture information included in the selected target training sample is input into the initial regression model to obtain the initial posture regression information.

[0062] In some embodiments, the execution entity may input the sample body posture information included in the selected target training sample into the initial regression model to obtain initial posture regression information.

[0063] In the seventh step, the above-mentioned initial posture category information and the above-mentioned initial posture regression information are input into the initial recognition model to obtain the initial posture state recognition result.

[0064] In some embodiments, the execution entity may input the initial posture category information and the initial posture regression information into an initial recognition model to obtain an initial posture state recognition result.

[0065] In the eighth step, based on a preset loss function, a difference value between the initial posture state recognition result and the sample posture state recognition result included in the selected target training sample is determined.

[0066] In some embodiments, the execution subject may determine the difference between the initial posture state recognition result and the sample posture state recognition result included in the selected target training sample based on a preset loss function. The preset prediction loss function may be, but is not limited to, mean square error loss function (MSE), hinge loss function, cross entropy loss function (CrossEntropy), 0-1 loss function, absolute value loss function, log loss function, square loss function, exponential loss function, etc.

[0067] In a ninth step, in response to determining that the difference value is greater than or equal to a preset difference value, the network parameters of the initial posture state recognition model are adjusted.

[0068] In some embodiments, the execution subject may adjust the network parameters of the initial posture state recognition model in response to determining that the difference value is greater than or equal to the preset difference value. For example, the predicted difference value and the preset predicted difference value may be differentiated. On this basis, the parameters of the initial code scheduling information prediction model are adjusted using methods such as back propagation and gradient descent. For example, the preset predicted difference value may be 0.1.

[0069] Optionally, in response to determining that the difference value is less than the above-mentioned preset difference value, the initial posture state recognition model is determined as the trained posture state recognition model.

[0070] In some embodiments, the execution entity may determine the initial posture state recognition model as the trained posture state recognition model in response to determining that the difference value is less than the preset difference value.

[0071] The relevant content of step 105, as an inventive point of an embodiment of the present application, solves the technical problem 2 mentioned in the background technology, "it is easy to increase the cardiac burden of the target user". Among them, the factors that are easy to increase the cardiac burden of the target user are often as follows: the position of the adjusted target user is not recognized, and it is difficult to determine that the target user after the position adjustment is in a side-lying state. When the target user is not in a side-lying state, it is easy to increase the cardiac burden of the target user. If the above factors are solved, the accuracy of abnormal domain name alarms can be improved. In order to achieve this effect, the present application firstly identifies the relatively accurate posture category information of the target user through the initial classification model to identify whether the target user is currently in a lying state. Then, the coordinates of the target user in the three-dimensional coordinate system can be identified through the initial regression model. After that, a relatively accurate initial posture state recognition result that characterizes the target user in a side-lying state can be identified through the initial recognition model. Therefore, by training an initial posture state recognition model including an initial classification model, an initial regression model and an initial recognition model, a relatively accurate posture state recognition model can be trained to identify a relatively accurate posture state recognition result. Thus, the adjusted target user can be further identified to identify whether the target user is currently in a side-lying state, so that the abnormal posture voice can be played to the target user who is not in the side-lying state, so as to adjust the body posture of the target user again to make it in the side-lying state. Furthermore, the target user can be put in the side-lying state to reduce the target user's cardiac burden.

[0072] Step 106, in response to determining that the posture state recognition result meets the preset abnormal posture state condition, controlling the voice playback device to play the abnormal posture voice.

[0073] In some embodiments, the execution subject may control the voice playback device to play the abnormal posture voice in response to determining that the posture state recognition result satisfies a preset abnormal posture state condition. The preset abnormal posture state condition may be: the posture state recognition result is not the first recognition result. For example, the abnormal posture voice may be a voice indicating that "the current user is not in a side-lying state and needs to turn over again".

[0074] Step 107: Control the vital sign detection device to detect the vital sign data of the target user.

[0075] In some embodiments, the execution subject may control the vital sign detection device to detect the vital sign data of the target user. The vital sign data may include but are not limited to: a three-dimensional vital sign picture set and a vital sign infrared picture. Here, the three-dimensional vital sign picture in the three-dimensional vital sign picture set may be a picture representing a body part of the target user. The vital sign infrared picture may be an infrared picture representing temperature.

[0076] Step 108: Perform recognition processing on the vital sign data to generate a vital sign recognition result.

[0077] In some embodiments, the execution entity may perform recognition processing on the vital sign data to generate a vital sign recognition result.

[0078] Regarding the third technical problem mentioned above: the observation method relies on the experience of medical staff, making it difficult to observe potential pressure sore lesions and to judge the severity of the pressure sore site, resulting in low accuracy of the observed pressure sore conditions, which is not conducive to the treatment and recovery of the target users. In combination with the current technical status, we decided to adopt the following solution: In practice, the execution subject may perform recognition processing on the above-mentioned vital sign data through the following steps to generate a vital sign recognition result: The first step is to grayscale each three-dimensional physical sign picture in the three-dimensional physical sign picture set included in the physical sign data to generate a three-dimensional physical sign grayscale picture and obtain a three-dimensional physical sign grayscale picture set. In practice, the execution subject can grayscale each three-dimensional physical sign picture in the three-dimensional physical sign picture set included in the physical sign data by a preset grayscale algorithm to generate a three-dimensional physical sign grayscale picture and obtain a three-dimensional physical sign grayscale picture set. For example, the preset grayscale algorithm can be but is not limited to: component method, maximum value method, average value method, weighted average method.

[0079] The second step is to perform equalization processing on each 3D physical sign grayscale image in the above 3D physical sign grayscale image set to generate a 3D physical sign balanced image and obtain a 3D physical sign balanced image set. In practice, the above execution subject can perform equalization processing on each 3D physical sign grayscale image in the above 3D physical sign grayscale image set by using a preset equalization algorithm to generate a 3D physical sign balanced image and obtain a 3D physical sign balanced image set. For example, the preset equalization algorithm can be: histogram equalization algorithm.

[0080] The third step is to perform noise reduction processing on each 3D body sign balanced image in the above 3D body sign balanced image set to generate a 3D body sign noise reduction image and obtain a 3D body sign noise reduction image set. In practice, the above execution subject can perform noise reduction processing on each 3D body sign balanced image in the above 3D body sign balanced image set by using a preset noise reduction algorithm to generate a 3D body sign noise reduction image and obtain a 3D body sign noise reduction image set. For example, the preset noise reduction algorithm can be a Haar wavelet transform algorithm.

[0081] The fourth step is to perform threshold segmentation processing on each 3D body sign denoised image in the above 3D body sign denoised image set to generate a pressure sore image group and obtain a pressure sore image group set. In practice, the above execution subject can perform threshold segmentation processing on each 3D body sign denoised image in the above 3D body sign denoised image set by a preset threshold segmentation algorithm to generate a pressure sore image group and obtain a pressure sore image group set. For example, the preset threshold segmentation algorithm can be: valley minimum algorithm, gray average algorithm, Huang's fuzzy threshold algorithm, bimodal average algorithm.

[0082] The fifth step is to perform feature extraction processing on each pressure sore picture in the above pressure sore picture set to generate pressure sore feature information and obtain a pressure sore feature information set. In practice, the above execution entity can perform feature extraction processing on each pressure sore picture in the above pressure sore picture set through a preset pressure sore feature extraction algorithm to generate pressure sore feature information and obtain a pressure sore feature information set. For example, the preset pressure sore feature extraction algorithm can be FAST (features from acceleratedsegment test). Here, the pressure sore feature information in the pressure sore feature information set can include but is not limited to: pressure sore length, pressure sore width, and pressure sore coordinates.

[0083] The sixth step is to identify and process the vital sign infrared images included in the vital sign data based on the pressure sore characteristic information set to generate a pressure sore temperature set. The pressure sore temperatures in the pressure sore temperature set correspond to the pressure sore characteristic information in the pressure sore characteristic information set. In practice, the execution subject can identify the pressure sore temperature corresponding to each pressure sore characteristic information in the pressure sore characteristic information set from the vital sign infrared images to obtain the pressure sore temperature set.

[0084] Step 7: The pressure sore characteristic information set and the pressure sore temperature set are fused to generate a pressure sore information set. In practice, first, for each pressure sore characteristic information in the pressure sore characteristic information set, the execution subject may fuse the pressure sore characteristic information and the pressure sore temperature corresponding to the pressure sore characteristic information in the pressure sore temperature set to generate pressure sore information. Then, the execution subject may determine the generated individual pressure sore information as a pressure sore information set.

[0085] In step 8, in response to determining that the pressure sore information set is empty, the preset normal physical sign information is determined as the physical sign recognition result. For example, the preset normal physical sign information may be information indicating that "the current user has no pressure sores".

[0086] In a ninth step, in response to determining that the pressure sore information set is not empty, the pressure sore information set is determined as a physical sign identification result.

[0087] The relevant content of step 108, as an inventive point of the embodiment of the present application, solves the technical problem three mentioned in the background technology, "not conducive to the treatment and recovery of the target user". Among them, the factors that are not conducive to the treatment and recovery of the target user are often as follows: the observation method relies on the experience of medical staff, it is difficult to observe potential pressure sore lesions, and it is difficult to judge the severity of the pressure sore site, resulting in low accuracy of the observed pressure sore situation. If the above factors are solved, the effect of improving the accuracy of abnormal domain name alarm can be achieved. In order to achieve this effect, the present application can firstly grayscale, equalize, and denoise the three-dimensional sign picture set included in the sign data, so as to pre-process each three-dimensional sign picture in the three-dimensional sign picture set to eliminate irrelevant information in the three-dimensional sign picture, simplify the data, and obtain a three-dimensional sign denoised picture set. Then, each three-dimensional sign denoised picture in the three-dimensional sign denoised picture set can be threshold segmented to generate a pressure sore picture group representing the pressure sore of the target user, and obtain a pressure sore picture group set. After that, the pressure sore feature information set representing the length, width, and coordinates of the pressure sore can be identified through feature extraction processing. Then, the temperature of the pressure sore can be identified through the infrared image of the physical sign, and a pressure sore temperature set representing the severity of the pressure sore can be obtained (the higher the pressure sore temperature, the more severe the pressure sore). Finally, the pressure sore feature information set and the pressure sore temperature set can be fused to obtain a pressure sore information set including the size, coordinates, and severity of the pressure sore. Therefore, through the above processing flow, relatively accurate pressure sore information including the size, coordinates, and severity of the pressure sore can be identified, which is beneficial to the treatment and recovery of the target user.

[0088] Step 109, in response to determining that the physical sign recognition result meets the preset abnormal physical sign condition, controlling the voice playback device to play the physical sign recognition result.

[0089] In some embodiments, the execution subject may control the voice playback device to play the physical sign recognition result in response to determining that the physical sign recognition result meets the preset abnormal physical sign condition. The preset abnormal physical sign condition may be: the physical sign recognition result is not the preset normal physical sign information. For example, the preset normal physical sign information may be information indicating that "the current user does not have a pressure sore".

[0090] The above-mentioned embodiments of the present application have the following beneficial effects: through the body position adjustment method based on the coma patient body position fixing device of some embodiments of the present application, the body position angle of the target user can be accurately controlled and the target user can be kept stable. Specifically, the reason why it is difficult to accurately control the body position angle of the target user and to keep the target user stable is that it is difficult to accurately control the body position angle of the target user and to keep the target user stable by manually adjusting the body posture. Based on this, the body position adjustment method based on the coma patient body position fixing device of some embodiments of the present application, first, controls the above-mentioned body position camera device to collect the three-dimensional body data of the target user. Thus, the three-dimensional body data can be collected so that the body posture of the target user can be adjusted by the angle adjustment device later. Secondly, based on the above-mentioned three-dimensional body data, the above-mentioned angle adjustment device is controlled to adjust the body posture of the above-mentioned target user. Thus, the target user can be adjusted from a lying state to a side-lying state by the angle adjustment device. Next, the above-mentioned pressure sensor is controlled to collect the adjusted body pressure value set of the target user. Thus, the body pressure values ​​of each contact point between the target user and the angle adjustment device can be collected. Next, in response to determining that the above-mentioned body pressure value set meets the preset abnormal pressure condition, the above-mentioned body position camera device is controlled to detect the body posture information of the above-mentioned target user in real time. Therefore, when the body pressure value set meets the preset abnormal pressure condition, it indicates that the body position of the target user is not correct, and it is necessary to further detect the body posture information through the body position camera device in order to subsequently determine whether an abnormal body position occurs. Then, the above-mentioned body posture information is input into the pre-trained posture state recognition model to obtain the posture state recognition result. Therefore, the posture state recognition result that indicates whether the target user is in a normal body position (side-lying state) can be identified through the posture state recognition model. Then, in response to determining that the above-mentioned posture state recognition result meets the preset abnormal posture state condition, the above-mentioned voice playback device is controlled to play abnormal posture voice. Therefore, when the posture state recognition result indicates that the target user is in an abnormal body position, the voice playback device can be controlled to play so as to adjust the body posture of the target user again. Afterwards, the above-mentioned physical sign detection device is controlled to detect the physical sign data of the above-mentioned target user. Therefore, the physical sign data of the target user can be detected by the physical sign detection device so as to subsequently identify whether the target user has pressure sores. After that, the above-mentioned physical sign data is processed for identification to generate a physical sign identification result. Thus, the pressure sore information present in the physical sign data can be identified through grayscale processing, equalization processing, noise reduction processing and threshold segmentation processing. Finally, in response to determining that the above-mentioned physical sign identification result meets the preset abnormal physical sign condition, the above-mentioned voice playback device is controlled to play the above-mentioned physical sign identification result. Thus, the physical sign identification result representing the pressure sore information can be played so as to subsequently treat the pressure sore.Therefore, by adjusting the body posture of the target user through the angle adjustment device included in the vital sign adjustment system instead of manual adjustment, the body angle of the target user can be accurately controlled and the stability of the target user can be maintained.

[0091] The present application also provides a computer device 300. Figure 3 As shown, the computer device 300 includes: a bus 301, a processor 302, a memory 303 and a communication interface 304. The processor 302, the memory 303 and the communication interface 304 communicate with each other through the bus 301. The computer device 300 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 300.

[0092] The bus 301 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 3 The bus 301 is represented by only one line, but it does not mean that there is only one bus or one type of bus. The bus 301 may include a path for transmitting information between various components of the computer device 300 (for example, the memory 303, the processor 302, and the communication interface 304).

[0093] The processor 302 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).

[0094] The memory 303 may include a volatile memory, such as a random access memory (RAM). The memory 303 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).

[0095] The memory 303 stores executable program codes, and the processor 302 executes the executable program codes to respectively implement the functions of the acquisition module, the sampling module, the determination module, and the mixing module, thereby implementing the above-mentioned body position adjustment method based on the comatose patient body position fixing device. That is, the memory 303 stores instructions for executing the above-mentioned body position adjustment.

[0096] The communication interface 304 uses a transceiver module such as, but not limited to, a network interface card or a transceiver to implement communication between the computer device 300 and other devices or a communication network.

[0097] 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 body position adjustment method based on the comatose patient body position fixing device.

[0098] 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 that contains 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 body position adjustment method based on the comatose patient body position fixation device.

[0099] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0100] 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 method for adjusting the body position of a comatose patient based on a body position fixing device, applied to the body position fixing device for a comatose patient, the body position fixing device for a comatose patient comprising: A body position camera device, a pressure sensor, an angle adjustment device, a vital sign detection device, and a voice playback device. The method includes: Controlling the body camera device to collect three-dimensional body data of the target user; Based on the three-dimensional body data, controlling the angle adjustment device to adjust the body posture of the target user; Controlling the pressure sensor to collect the adjusted body pressure value set of the target user; In response to determining that the body pressure value set meets a preset abnormal pressure condition, controlling the body position camera device to detect the body posture information of the target user in real time; Inputting the body posture information into a pre-trained posture state recognition model to obtain a posture state recognition result; In response to determining that the posture state recognition result meets a preset abnormal posture state condition, controlling the voice playback device to play abnormal posture voice; Controlling the vital sign detection device to detect the vital sign data of the target user; Performing recognition processing on the vital sign data to generate a vital sign recognition result; In response to determining that the physical sign recognition result meets a preset abnormal physical sign condition, the voice playback device is controlled to play the physical sign recognition result.

2. The method for adjusting the body position of a comatose patient based on the body position fixing device according to claim 1, wherein: The angle adjustment device includes: a turning over adjustment device, a soft pillow placement device, and a head fixing device; and the controlling the angle adjustment device to adjust the body posture of the target user based on the three-dimensional body data includes: Based on the three-dimensional body data, controlling the turning adjustment device to adjust the body posture of the target user; Controlling the soft pillow placement device to place a first target soft pillow at a first designated position of the target user; Controlling the soft pillow placement device to place a second target soft pillow at a second designated position of the target user; The head fixing device is controlled to fix the head of the target user.

3. The method for adjusting the body position of a comatose patient based on the body position fixing device according to claim 1, wherein: The controlling the body position camera device to detect the body position information of the target user in real time includes: Controlling the body position camera device to capture a sequence of initial body state pictures of the target user; Preprocessing each initial body state picture in the initial body state picture sequence to generate a body state picture, thereby obtaining a body state picture sequence; Performing feature extraction processing on each body state picture in the body state picture sequence to generate a body feature information group, thereby obtaining a body feature information group sequence; Performing vectorization processing on each body feature information in the body feature information group sequence to generate a body feature vector, thereby obtaining a body feature vector group sequence; For each body feature vector group in the body feature vector group sequence, in response to determining that the body feature vector group satisfies a first preset vector condition, determining a target body feature vector group based on the body feature vector group; Determining each generated target body feature vector group as a target body feature vector group sequence; The last target body feature vector group in the target body feature vector group sequence is determined as body posture information.

4. The method for adjusting the body position of a comatose patient based on the body position fixing device according to claim 3, wherein: The step of determining a target body feature vector group based on the body feature vector group includes: In response to determining that the body feature vector group satisfies a second preset vector condition, performing a first projection process on a first body feature vector group in the body feature vector group sequence to generate a target body feature vector group; In response to determining that the body feature vector group does not satisfy the second preset vector condition, a second projection process is performed on the body feature vector group to generate a target body feature vector group.

5. The method for adjusting the body position of a comatose patient based on the body position fixing device according to claim 4, wherein: The performing a second projection process on the body feature vector group to generate a target body feature vector group includes: Determining a previous body feature vector group of the body feature vector group in the body feature vector group sequence as an initial body feature vector group; Performing screening processing on the initial body feature vector group to generate a body feature screening vector group; Performing projection processing on the body feature screening vector group to generate a body feature projection vector group; Based on the body feature projection vector group, construct an initial three-dimensional query vector group; Merging the body feature projection vector group and the initial three-dimensional query vector group to generate a three-dimensional query vector group; Performing deduplication processing on the three-dimensional query vector group to generate a target three-dimensional query vector group; Performing spatial projection processing on the body feature vector group to generate a body three-dimensional feature vector group; The target three-dimensional query vector group and the body three-dimensional feature vector group are input into an interactive attention model to obtain a target body feature vector group.

6. 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 when the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 5 are implemented.

7. 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 any one of claims 1 to 5 are implemented.

Images