Wound nursing intelligent decision-making method and system based on high-precision image recognition

Abstract

The invention provides a wound nursing intelligent decision-making method and system based on high-precision image recognition. The method comprises the following steps: in response to a received wound monitoring request, extracting a patient identifier, wound image information corresponding to a multi-angle high-precision image, wound background information and multi-modal sensor information from the wound monitoring request; a pre-trained three-dimensional convolutional neural network model is used for automatically recognizing the wound state, and a wound recognition result is obtained; based on the wound image information and the wound background information, assisting in positioning through an augmented reality technology; performing wound healing environment assessment based on the information of the multi-modal sensor; retrieving matched nursing strategy information from a preset nursing scheme database based on patient historical information and personalized treatment preference information corresponding to the patient identifier, the wound identification result, the wound positioning information and the wound healing environment evaluation information; and pushing the nursing strategy information. According to the scheme, the wound nursing quality is improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of image processing technology, and in particular to a wound care intelligent decision-making method and system based on high-precision image recognition. Background Art

[0002] With the development of medical technology, especially in the field of telemedicine and smart medicine, higher requirements are placed on wound care. Traditional wound care relies on the experience and manual operation of medical staff, but in the face of complex and changeable wound conditions, this model is difficult to meet the needs of high precision and personalization. Therefore, there is an urgent need for an intelligent decision-making method for wound care based on high-precision image recognition to achieve automated, intelligent management and care of wounds.

[0003] The existing intelligent decision-making methods for wound care are based on the patient's wound image information to identify the corresponding wound location, wound area, wound severity, etc., and then determine the nursing strategy information.

[0004] However, factors such as low image accuracy and angle differences can easily lead to inaccurate wound image information, which in turn leads to low recognition accuracy of wound location, wound area, wound severity, etc., and because the wound healing environment is unclear, the nursing strategy information is biased. In addition, existing technologies cannot provide personalized services for different patients, and the patient experience is poor. Summary of the invention

[0005] The embodiments of the present application provide a wound care intelligent decision-making method and system based on high-precision image recognition, which are used to solve the problem of poor wound care quality in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a wound care intelligent decision-making method based on high-precision image recognition, comprising:

[0007] In response to the received wound monitoring request, extract key information from the wound monitoring request, the key information including: a patient identifier, wound image information corresponding to the multi-angle high-precision image, wound background information, and multimodal sensor information;

[0008] Acquire the patient history information and personalized treatment preference information corresponding to the patient identifier from a preset medical information database; the personalized treatment preference information includes: the patient's preference for drugs and treatment methods, and the treatment methods include drug therapy, surgical treatment or psychotherapy;

[0009] Based on the wound image information and the wound background information, a pre-trained three-dimensional convolutional neural network model is used to automatically identify the wound state to obtain a wound recognition result, wherein the convolutional neural network model is obtained by training the wound image information corresponding to multi-scene high-precision images, and the multi-scene high-precision images include high-precision images with different lighting conditions, angles and / or backgrounds;

[0010] Based on the wound image information and the wound background information, wound positioning information is obtained by using augmented reality technology to assist positioning;

[0011] Performing wound healing environment assessment based on multimodal sensor information to obtain wound healing environment assessment information; the wound healing environment assessment information includes: first assessment information indicating whether the temperature is suitable for wound healing, second assessment information indicating whether the humidity is suitable for wound healing, third assessment information indicating whether the gas concentration is suitable for wound healing, and fourth assessment information indicating whether biological activity affects wound healing;

[0012] Retrieving matching nursing strategy information from a preset nursing program database based on the patient history information, personalized treatment preference information, wound identification results, wound location information, and wound healing environment assessment information;

[0013] Push the nursing strategy information.

[0014] Optionally, the wound positioning information includes a target wound boundary and a target wound size;

[0015] Based on the wound image information and the wound background information, wound positioning information is obtained by using augmented reality technology to assist positioning, including:

[0016] Based on the wound image information, construct a three-dimensional depth image using a depth estimation algorithm;

[0017] Setting an augmented reality mark on the wound area of ​​the three-dimensional depth image, wherein the augmented reality mark is a preset pattern or a preset QR code;

[0018] Segmenting the wound area using a semantic segmentation algorithm to obtain an initial wound boundary;

[0019] Based on the pixel-level segmentation result of the initial wound boundary, a contour detection algorithm is used to calculate and obtain the initial wound size;

[0020] Based on the wound background information, the initial wound boundary and the initial wound size are adjusted using a background perception algorithm to obtain a target wound boundary and a target wound size.

[0021] Optionally, adjusting the initial wound boundary and the initial wound size using a background perception algorithm based on the wound background information to obtain a target wound boundary and a target wound size includes:

[0022] Extracting feature information based on the wound background information, wherein the feature information includes: color distribution information, texture pattern information and skin texture information;

[0023] Based on the feature information, a normal skin area within the initial wound boundary is removed, and a wound area outside the initial wound boundary is filled to obtain a target wound boundary;

[0024] A size correction is performed based on the target wound boundary to obtain a target wound size.

[0025] Optionally, removing the normal skin area within the initial wound boundary comprises:

[0026] By fusing the feature information and using an anomaly detection algorithm to identify a normal skin area within the initial wound boundary;

[0027] The normal skin area within the initial wound boundary is marked to remove the normal skin area within the initial wound boundary.

[0028] Optionally, filling the wound area outside the initial wound boundary comprises:

[0029] The wound area outside the initial wound boundary is filled using a generative adversarial network technique or a multi-scale context-aware filling technique.

[0030] Optionally, the multimodal sensor information includes temperature sensor information, humidity sensor information, gas sensor information and biosensor information;

[0031] The wound healing environment assessment is performed based on the multimodal sensor information to obtain wound healing environment assessment information, including:

[0032] Building a temperature impact model based on the temperature sensor information;

[0033] Building a humidity impact model based on the humidity sensor information;

[0034] Building a gas concentration impact model based on the gas sensor information;

[0035] constructing a biological impact model based on the biosensor information;

[0036] The wound healing environment is evaluated based on the temperature influence model, humidity influence model, gas concentration influence model and biological influence model to obtain wound healing environment evaluation information.

[0037] Optionally, after performing wound healing environment assessment based on the temperature impact model, the humidity impact model, the gas concentration impact model and the biological impact model to obtain wound healing environment assessment information, the method further includes:

[0038] quantifying the wound healing environment assessment information to obtain an assessment value;

[0039] If the evaluation value is less than the preset evaluation threshold, the wound healing environment processing information is pushed.

[0040] Optionally, the key information also includes the sending time corresponding to the wound monitoring request and the geographical coordinates of the patient;

[0041] After pushing the nursing strategy information, the method further includes:

[0042] In response to the received medical consultation guidance request, based on the nursing strategy information, the sending time and the patient's geographical coordinates, determine and push pharmacy information or hospital information, the pharmacy information includes the pharmacy name and the pharmacy's geographical coordinates, and the hospital information includes the hospital name, the hospital's geographical coordinates and the waiting time of the department corresponding to the wound;

[0043] In response to the received pharmacy information or hospital information, a medical consultation route is determined and pushed.

[0044] Optionally, after receiving the pharmacy information or hospital information, the method further includes:

[0045] In response to the received nursing feedback information, the key information, the nursing strategy information and the nursing feedback information are sent to a terminal device of a department corresponding to the wound in a target pharmacy or a target hospital;

[0046] When it is identified that the distance between the patient's geographical coordinates and the geographical coordinates of the target pharmacy or target hospital is less than a preset distance threshold, the medical consultation preparation information is sent to the terminal device of the department corresponding to the wound in the target pharmacy or target hospital.

[0047] In a second aspect, the embodiment of the present application provides a wound care intelligent decision-making system based on high-precision image recognition, including:

[0048] an extraction module, configured to extract key information from the wound monitoring request in response to the received wound monitoring request, the key information comprising: a patient identifier, wound image information corresponding to the multi-angle high-precision image, wound background information, and multimodal sensor information;

[0049] an acquisition module, configured to acquire the patient history information and personalized treatment preference information corresponding to the patient identifier from a preset medical information database; the personalized treatment preference information includes: the patient's preference for drugs and treatment methods, the treatment methods including drug therapy, surgical treatment or psychotherapy;

[0050] A recognition module, configured to automatically recognize the wound state based on the wound image information and the wound background information using a pre-trained three-dimensional convolutional neural network model to obtain a wound recognition result, wherein the convolutional neural network model is obtained by training wound image information corresponding to multi-scene high-precision images, and the multi-scene high-precision images include high-precision images with different lighting conditions, angles and / or backgrounds;

[0051] A positioning module, used to obtain wound positioning information by using augmented reality technology to assist positioning based on the wound image information and the wound background information;

[0052] An environmental assessment module, configured to perform wound healing environment assessment based on multimodal sensor information to obtain wound healing environment assessment information; the wound healing environment assessment information includes: first assessment information indicating whether the temperature is suitable for wound healing, second assessment information indicating whether the humidity is suitable for wound healing, third assessment information indicating whether the gas concentration is suitable for wound healing, and fourth assessment information indicating whether biological activity affects wound healing;

[0053] A retrieval module, configured to retrieve matching nursing strategy information from a preset nursing program database based on the patient history information, personalized treatment preference information, wound identification results, wound location information, and wound healing environment assessment information;

[0054] A push module is used to push the nursing strategy information.

[0055] In an embodiment of the present application, a wound care intelligent decision-making method based on high-precision image recognition includes: in response to a received wound monitoring request, extracting key information from the wound monitoring request, the key information including: a patient identifier, wound image information corresponding to multi-angle high-precision images, wound background information, and multimodal sensor information; obtaining patient history information and personalized treatment preference information corresponding to the patient identifier from a preset medical information library; the personalized treatment preference information includes: the patient's preference for drugs and treatment methods, the treatment methods including drug therapy, surgical treatment or psychotherapy; based on the wound image information and wound background information, using a pre-trained three-dimensional convolutional neural network model to automatically identify the wound state to obtain a wound recognition result, the convolutional neural network model is through the wound image information corresponding to the multi-scene high-precision images The multi-scene high-precision images obtained through training include high-precision images with different lighting conditions, angles and / or backgrounds; based on the wound image information and the wound background information, the wound positioning information is obtained by assisting positioning with augmented reality technology; based on the multimodal sensor information, the wound healing environment is evaluated to obtain the wound healing environment evaluation information; the wound healing environment evaluation information includes: first evaluation information indicating whether the temperature is suitable for wound healing, second evaluation information indicating whether the humidity is suitable for wound healing, third evaluation information indicating whether the gas concentration is suitable for wound healing, and fourth evaluation information indicating whether biological activity affects wound healing; based on the patient's historical information, personalized treatment preference information, wound identification results, wound positioning information and wound healing environment evaluation information, matching nursing strategy information is retrieved from the preset nursing plan database; and the nursing strategy information is pushed.

[0056] Compared with the prior art, the embodiments of the present application have the following effects:

[0057] (1) Automatically identify wound status using a pre-trained 3D convolutional neural network model, which is trained based on high-precision images of multiple scenes, covering high-precision images under different lighting conditions, angles, and backgrounds. This means that accurate wound recognition results can be obtained regardless of the environment in which the high-precision images are taken, thereby improving the accuracy of wound status recognition.

[0058] (2) The introduction of augmented reality technology to assist in positioning can more accurately determine wound location information and reduce errors caused by image shooting angles or background interference. The application of this technology makes wound location information more reliable and provides accurate data support for subsequent nursing measures.

[0059] (3) Comprehensive wound healing environment assessments such as temperature, humidity, gas concentration, and biological activity can be combined with multimodal sensor information to help determine whether the patient's environment meets the appropriate conditions for wound healing. In this way, it can be ensured that wound healing is promoted in the best environment.

[0060] (4) Based on the patient's historical information, including the patient's personalized treatment preference information on drug preferences and treatment methods, wound identification results, wound location information, and healing environment assessment information, the most suitable nursing strategy information for the patient can be retrieved from the preset nursing plan database. This not only improves the pertinence of the nursing plan, but also enhances the effectiveness of the nursing measures. The automatic push of nursing strategy information enables patients or their families to receive professional nursing advice in a timely manner, which helps to take appropriate nursing measures in a timely manner and accelerate the wound healing process. In addition, this method provides a data-driven wound care intelligent decision-making system for wound care by integrating multi-source heterogeneous data. Such a wound care intelligent decision-making system can better support clinical decision-making and improve the quality of medical services.

[0061] In summary, the embodiments of the present application provide an intelligent and precise wound care decision support method by integrating high-precision image recognition technology, augmented reality positioning technology, multimodal sensor data fusion, and personalized nursing strategy recommendation functions, which helps to improve the quality and efficiency of wound care.

[0062] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0064] Figure 1 A flowchart of an intelligent decision-making method for wound care based on high-precision image recognition provided in an embodiment of the present application;

[0065] Figure 2 A schematic diagram of the structure of an intelligent decision-making system for wound care based on high-precision image recognition provided in an embodiment of the present application;

[0066] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0067] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0068] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0069] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0070] Figure 1 A flowchart of a wound care intelligent decision-making method based on high-precision image recognition is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0071] S11. In response to the received wound monitoring request, extract key information from the wound monitoring request, where the key information includes: a patient identifier, wound image information corresponding to multi-angle high-precision images, wound background information, and multimodal sensor information.

[0072] Among them, key information can be understood as necessary information required to execute the intelligent decision-making method for wound care, including but not limited to patient identifiers (such as patient name or patient number), information related to the wound, and information related to the wound healing environment. Optionally, multi-angle high-precision images are wound images taken by smartphones or dedicated medical equipment, and these wound images capture the overall view of the wound from different angles. Optionally, the wound background information includes wound color information, background color information, etc., wherein the background color information can be the skin color around the wound or the color of the area outside the body of the patient. Multimodal sensor information includes temperature sensor information, humidity sensor information, gas sensor information, biosensor information, etc. In addition, key information can also include timestamps, the shooting location of the high-precision image, etc., wherein the timestamp corresponds to the moment of high-precision image acquisition.

[0073] For example, the patient's arm is scratched by glass, and the patient or family members take pictures of the patient's arm through the application in the terminal device to obtain wound images at different distances and angles (for example: 30 degrees, 45 degrees, 90 degrees, 135 degrees). The distance can refer to the distance between the wound and the terminal device, for example: 20 cm, 50 cm, etc. The angle can be the angle between the terminal device plane and the approximate plane where the wound is located. When the angle is 90 degrees, the corresponding wound image can be used as a front view of the wound. Alternatively, the terminal device can shoot a video around the wound. After the photo is taken, when the patient or family members submit a wound monitoring request through the application, the wound care intelligent decision-making system based on high-precision image recognition will automatically identify key information such as the patient number in the wound monitoring request. It should be understood that the wound care intelligent decision-making system based on high-precision image recognition can be a computing device. Terminal devices and computing devices can refer to smart phones, computers, tablets, special medical equipment and other devices.

[0074] S12. Obtaining patient history information and personalized treatment preference information corresponding to the patient identifier from a preset medical information database. The personalized treatment preference information includes: the patient's preference for medication and treatment method, and the treatment method includes medication, surgery or psychotherapy.

[0075] Optionally, the preset medical information library stores the electronic health record of each patient, and the electronic health records of different patients are represented by patient identifiers. The electronic health record stores patient historical information and personalized treatment preference information. Among them, the patient historical information can be historical information related to the current wound, such as: the historical treatment experience of the current wound. It can also be historical information unrelated to the current wound, such as the historical treatment experience of wounds in other locations, allergy history, surgical history, etc. Personalized treatment preference information can refer to the patient's preference for drugs, whether he is willing to accept a certain treatment method, etc.

[0076] S13. Based on the wound image information and the wound background information, a pre-trained three-dimensional convolutional neural network model is used to automatically identify the wound status to obtain a wound recognition result. The convolutional neural network model is obtained by training the wound image information corresponding to the multi-scene high-precision images. The multi-scene high-precision images include high-precision images with different lighting conditions, angles and / or backgrounds.

[0077] Optionally, in order to obtain a three-dimensional convolutional neural network model with high recognition accuracy, this embodiment trains the model before applying it, and the training samples are different from the wound images of a single angle in the related art, but are high-precision wound images under a large number of different lighting conditions, angles and backgrounds. The pre-trained three-dimensional convolutional neural network model can automatically analyze the wound image information and the wound background information to obtain the wound recognition result. The wound recognition result can be the classification result of the wound, such as: abrasion, cut, contusion, etc.

[0078] S14. Based on the wound image information and wound background information, the wound positioning information is obtained by using augmented reality technology to assist positioning.

[0079] Optionally, based on the wound image information and wound background information, the wound care intelligent decision-making system based on high-precision image recognition can use augmented reality technology to more accurately determine the wound location information. The wound location information includes the boundary, shape, length, width, area, etc. of the wound.

[0080] S15. Perform wound healing environment assessment based on the multimodal sensor information to obtain wound healing environment assessment information.

[0081] The wound healing environment assessment information includes: first assessment information indicating whether the temperature is suitable for wound healing, second assessment information indicating whether the humidity is suitable for wound healing, third assessment information indicating whether the gas concentration is suitable for wound healing, and fourth assessment information indicating whether biological activity affects wound healing.

[0082] For example, by evaluating the wound healing environment, it is possible to monitor whether the temperature and humidity around the wound are suitable for healing, and it is also possible to detect whether the ratio of oxygen and carbon dioxide in the air is conducive to wound recovery.

[0083] S16. Based on the patient's historical information, personalized treatment preference information, wound identification results, wound location information, and wound healing environment assessment information, matching nursing strategy information is retrieved from a preset nursing plan database.

[0084] Optionally, patient history information, personalized treatment preference information, wound identification results, wound location information, and wound healing environment assessment information can all reflect the individual conditions of the patient. The intelligent decision-making system for wound care based on high-precision image recognition can recommend personalized care plans based on the individual conditions of the patient. For example, based on the specific conditions of patient A, the most suitable nursing measures such as disinfection methods, dressing selection, and drug use are recommended for patient A. Based on the specific conditions of patient B, the most suitable nursing measures such as disinfection methods, dressing selection, and drug use are recommended for patient A.

[0085] S17. Push nursing strategy information.

[0086] The intelligent decision-making system for wound care based on high-precision image recognition can be pushed to patients, family members or medical staff through applications, SMS notifications, voice prompts, etc. The nursing strategy information can be a detailed nursing guide document to guide the next treatment steps.

[0087] Compared with the prior art, the embodiments of the present application have the following effects:

[0088] (1) In S13, a pre-trained 3D convolutional neural network model is used to automatically identify wound status. The model is trained based on high-precision images of multiple scenes, covering high-precision images under different lighting conditions, angles and backgrounds. This means that no matter what environment the high-precision images are taken in, accurate wound recognition results can be obtained, thereby improving the accuracy of wound status recognition.

[0089] (2) The introduction of augmented reality technology to assist positioning in S14 can more accurately determine wound positioning information and reduce errors caused by image shooting angles or background interference. The application of this technology makes wound positioning information more reliable and provides accurate data support for subsequent nursing measures.

[0090] (3) In S15, a comprehensive wound healing environment assessment including temperature, humidity, gas concentration, and biological activity is performed by combining multimodal sensor information to help determine whether the patient's environment meets the appropriate conditions for wound healing. In this way, wound healing can be promoted in the best environment.

[0091] (4) In S16-S17, based on the patient's historical information, including the patient's personalized treatment preference information for drug preferences and treatment methods, wound identification results, wound location information, and healing environment assessment information, the nursing strategy information that best suits the patient can be retrieved from the preset nursing plan database. This not only improves the pertinence of the nursing plan, but also enhances the effectiveness of the nursing measures. The automatic push of nursing strategy information enables patients or their families to receive professional nursing advice in a timely manner, which helps to take appropriate nursing measures in a timely manner and accelerate the wound healing process. In addition, this method provides a data-driven wound care intelligent decision-making system for wound care by integrating multi-source heterogeneous data. Such a wound care intelligent decision-making system can better support clinical decision-making and improve the quality of medical services.

[0092] In summary, the embodiments of the present application provide an intelligent and precise wound care decision support method by integrating high-precision image recognition technology, augmented reality positioning technology, multimodal sensor data fusion, and personalized nursing strategy recommendation functions, which helps to improve the quality and efficiency of wound care.

[0093] In a possible embodiment, the wound location information may include but is not limited to: target wound boundary, target wound size. S14, based on the wound image information and wound background information, using augmented reality technology to assist in positioning, to obtain wound location information, including:

[0094] Step 141: Based on the wound image information, a three-dimensional depth image is constructed using a depth estimation algorithm.

[0095] Optionally, the depth estimation algorithm is a computer vision technology. In this embodiment, the depth of the wound can be predicted by analyzing multiple high-precision images. The depth estimation algorithm includes a stereo matching algorithm, an optical flow method, etc. The value of each pixel in the three-dimensional depth image is the three-dimensional depth information of the point. In step 141, the three-dimensional depth information of the wound area can be inferred from the two-dimensional wound image by the depth estimation algorithm.

[0096] Step 142: setting an augmented reality mark on the wound area of ​​the three-dimensional depth image, where the augmented reality mark is a preset pattern or a preset QR code.

[0097] Optionally, the augmented reality mark may include relevant information of the wound, such as the location, etc. Using a preset pattern or preset QR code in augmented reality, it is possible to identify and map the location of the wound in the positioning image to the location of the wound in the real world. In step 142, in the constructed three-dimensional depth image, the wound area is selected and the augmented reality mark is set to assist in positioning and subsequent processing.

[0098] Step 143: Segment the wound area using a semantic segmentation algorithm to obtain an initial wound boundary.

[0099] Optionally, in this embodiment, the three-dimensional depth image is segmented into a wound area and a background area (ie, a non-wound area), the object in the wound area is a wound, and a semantic segmentation algorithm is used to identify the pixel-level position of the wound.

[0100] Exemplarily, the three-dimensional depth image is segmented to obtain a binary image, in which the white part represents the wound area and the black part is the background area.

[0101] Step 144: Based on the pixel-level segmentation result of the initial wound boundary, a contour detection algorithm is used to perform calculations to obtain the initial wound size.

[0102] Optionally, a contour detection algorithm is used to detect the edge of the wound from the three-dimensional depth image, and then calculate the plane size of the wound in two-dimensional space, the plane size of the wound in three-dimensional space, the three-dimensional size, etc. The plane size includes information such as size, shape, area or surface area, perimeter, etc. The three-dimensional size includes volume, depth, three-dimensional shape, etc.

[0103] Step 145: Based on the wound background information, use a background perception algorithm to adjust the initial wound boundary and the initial wound size to obtain a target wound boundary and a target wound size.

[0104] Optionally, considering the influence of wound background information, the initially obtained initial wound boundary and initial wound size are corrected by a background perception algorithm to obtain a more accurate result.

[0105] For example, the skin color and texture information around the wound is analyzed, and the parameters of the semantic segmentation algorithm are adjusted through the background perception algorithm to make the segmentation result more accurate, and finally determine the precise boundary and size of the wound. For example, if it is found that the color of the wound edge is significantly different from that of the surrounding skin, the segmentation accuracy can be improved by adjusting the segmentation threshold.

[0106] Through the above steps, the wound care intelligent decision-making system based on high-precision image recognition can use augmented reality technology to assist in positioning based on wound image information and wound background information, and ultimately obtain accurate target wound positioning information.

[0107] As an optional implementation, step 145, based on the wound background information, using a background perception algorithm to adjust the initial wound boundary and the initial wound size to obtain a target wound boundary and a target wound size, includes:

[0108] Step 1451: extract feature information based on the wound background information, the feature information including: color distribution information, texture pattern information and skin texture information.

[0109] Optionally, the color distribution information may refer to the color composition of the wound area in the three-dimensional depth image, the color composition of the non-wound area, etc. The texture pattern information may refer to the repetitive or regular arrangement of the color or brightness in the wound area in the three-dimensional depth image, the repetitive or regular arrangement of the color or brightness in the non-wound area, etc. Skin texture information may refer to the detailed features of the patient's skin surface, such as wrinkles, pores, etc., which are useful for distinguishing healthy skin from damaged skin. Damaged skin can be understood as skin with wounds.

[0110] In this embodiment, by extracting features of wound background information, including color distribution information, texture pattern information, and skin texture information, the wound area and the surrounding normal skin area can be distinguished more accurately. This helps to improve the recognition accuracy of the wound boundary in subsequent processing steps, ensuring that normal skin is not mistaken for part of the wound, thereby avoiding erroneous analysis results caused by misidentification.

[0111] Step 1452: Based on the feature information, remove the normal skin area within the initial wound boundary, and fill the wound area outside the initial wound boundary to obtain the target wound boundary.

[0112] Optionally, removing the normal skin area within the initial wound boundary includes: step a1, identifying the normal skin area within the initial wound boundary by fusing feature information and using an anomaly detection algorithm; step a2, marking the normal skin area within the initial wound boundary to remove the normal skin area within the initial wound boundary.

[0113] Optionally, filling the wound area outside the initial wound boundary includes: using a generative adversarial network technology or a multi-scale context-aware filling technology to fill the wound area outside the initial wound boundary.

[0114] Alternatively, multi-scale context-aware filling techniques restore the missing parts by considering contextual information at different scales of the image.

[0115] In this embodiment, by removing the normal skin area within the initial wound boundary and filling the wound area outside the boundary, the possible errors in the initial identification process can be corrected, so that the actual boundary of the wound is closer to the real wound condition. This not only improves the accuracy of wound analysis, but also reduces the need for manual intervention and enhances the automation level of the system.

[0116] Step 1453: Perform size correction based on the target wound boundary to obtain the target wound size.

[0117] Optionally, after adjusting the boundaries, recalculate the wound size to obtain the precise target wound size. This can be done by counting the number of pixels within the wound area or using an area measurement tool. For example, in the adjusted binary image, the number of all white pixels represents the area of ​​the wound.

[0118] In this embodiment, the size correction is performed based on the adjusted wound boundary to ensure that the target wound size is accurate. This is crucial for the intelligent decision-making method for wound care, because the wound size is an important basis for evaluating the progress of treatment and deciding the next treatment plan. Correct size measurement can prevent treatment delays or improper treatment caused by inaccurate size.

[0119] Under the combined effect of steps S1451 to S1453, this embodiment can significantly improve the quality of high-precision image analysis, provide more reliable data support for clinicians, and thus optimize the nursing decision-making process. In addition, this series of steps can also help reduce the waste of medical resources and improve the efficiency of medical services.

[0120] In a possible embodiment, the multimodal sensor information includes but is not limited to: temperature sensor information, humidity sensor information, gas sensor information, and biosensor information. Therefore, S15, performing wound healing environment assessment based on the multimodal sensor information to obtain wound healing environment assessment information includes:

[0121] Step 151: construct a temperature impact model based on temperature sensor information.

[0122] Optionally, the temperature impact model is a relationship model between temperature and wound healing rate, and the temperature impact model is used to evaluate the impact of ambient temperature on wound healing. By analyzing the temperature sensor information, it is possible to learn how temperature changes affect the wound healing process.

[0123] Exemplarily, this embodiment uses a temperature sensor to continuously monitor the temperature near the wound, records the temperature values ​​in different time periods, and associates them with the wound healing status. Then, a machine learning method (such as linear regression, decision tree, etc.) is used to establish a relationship model between temperature and wound healing rate, and the wound healing rate under different temperature conditions is predicted based on the relationship model.

[0124] Step 152: construct a humidity impact model based on the humidity sensor information.

[0125] Optionally, the humidity impact model is a relationship model between humidity and wound healing rate, and the humidity impact model is used to evaluate the impact of environmental humidity on wound healing. By analyzing the humidity sensor information, it can be known how humidity changes affect the wound healing process.

[0126] Exemplarily, this embodiment uses a humidity sensor to continuously monitor the humidity near the wound, records the humidity values ​​in different time periods, and associates them with the wound healing status. Then, a machine learning method (such as support vector machine, random forest, etc.) is used to establish a relationship model between humidity and wound healing rate. The relationship model is used to predict the wound healing rate under different humidity conditions.

[0127] Step 153: construct a gas concentration impact model based on the gas sensor information.

[0128] Optionally, the gas concentration effect model is used to evaluate the effect of gas components (such as oxygen, carbon dioxide, etc.) in the environment on wound healing. By analyzing the gas sensor information, it is possible to learn how different gas concentrations affect the wound healing process.

[0129] Exemplarily, this embodiment uses a gas sensor to continuously monitor the oxygen and carbon dioxide concentrations near the wound. The gas concentration values ​​at different time periods are recorded and associated with the wound healing status. A relationship model between gas concentration and wound healing rate is established using machine learning methods (such as neural networks, Bayesian networks, etc.). The speed of wound healing under different gas concentration conditions is predicted based on the relationship model.

[0130] Step 154: construct a biological impact model based on the biosensor information.

[0131] Optionally, the biological influence model is used to evaluate the influence of biological factors (such as bacteria, viruses, etc.) on wound healing. By analyzing the biosensor information, it is possible to learn how biological factors influence the wound healing process.

[0132] Exemplarily, this embodiment uses a biosensor (such as a microbial sensor) to continuously monitor the biological activity near the wound. The biological activity values ​​of different time periods are recorded and associated with the wound healing status. A relationship model between biological activity and wound healing rate is established using machine learning methods (such as logistic regression, gradient boosting tree, etc.). The relationship model predicts the speed of wound healing under different biological activity conditions.

[0133] Step 155 : Perform wound healing environment assessment based on the temperature impact model, the humidity impact model, the gas concentration impact model, and the biological impact model to obtain wound healing environment assessment information.

[0134] By integrating the results of the above four models, the wound healing environment can be comprehensively evaluated and a scientific basis can be provided for wound care.

[0135] In practical applications, the results of the temperature influence model, humidity influence model, gas concentration influence model and biological influence model can be integrated. Based on the integrated data, it is evaluated whether the current environment is conducive to wound healing. Wound healing environment evaluation information containing the following contents is generated: first evaluation information indicating whether the temperature is suitable for wound healing, second evaluation information indicating whether the humidity is suitable for wound healing, third evaluation information indicating whether the gas concentration is suitable for wound healing, fourth evaluation information indicating whether biological activity affects wound healing, etc.

[0136] Through the above steps, the wound care intelligent decision-making system based on high-precision image recognition can comprehensively evaluate the wound healing environment based on multimodal sensor information and provide scientific decision-making support for medical staff, thereby optimizing the wound care process.

[0137] Optionally, this embodiment can provide specific environmental improvement suggestions based on the evaluation results, such as adjusting temperature, humidity, increasing oxygen supply, or taking measures to inhibit the growth of harmful microorganisms. In a possible embodiment, after performing wound healing environment evaluation based on the temperature influence model, humidity influence model, gas concentration influence model, and biological influence model in S155 to obtain wound healing environment evaluation information, the method further includes:

[0138] Step 156: quantify the wound healing environment assessment information to obtain an assessment value.

[0139] Optionally, the assessment value is a quantitative indicator of the quality of the wound healing environment, which reflects the overall degree to which the environmental conditions affect wound healing. By quantifying this, it is easier to compare different environmental conditions and determine when action is needed to improve the environment.

[0140] Exemplarily, in a hospital's intensive care unit, in order to promote wound healing in patients, doctors and nurses need to monitor the microenvironmental conditions around the wound. The following describes how to implement the process from environmental assessment to wound healing environment processing information push in this scenario. This embodiment assigns a weight to each environmental factor (temperature, humidity, gas concentration, biological activity) to reflect its importance to wound healing. Given the wound healing environment assessment information in step 155, each factor is scored (e.g., 0-100 points, where 0 represents the most unfavorable healing and 100 represents the best healing condition). The scores of each factor are multiplied by their corresponding weights and then summed to obtain the final evaluation value. Optionally, this embodiment can standardize the evaluation value so that it falls into a specific range, such as between 0-100, for ease of understanding and comparison.

[0141] For example, assuming that the temperature weight is 0.3, the humidity weight is 0.3, the gas concentration weight is 0.2, and the biological activity weight is 0.2. The temperature score is 80, the humidity score is 70, the gas concentration score is 90, and the biological activity score is 60. Evaluation value = (80×0.3)+(70×0.3)+(90×0.2)+(60×0.2)=24+21+18+12=75.

[0142] Step 157: If the evaluation value is less than the preset evaluation threshold, the wound healing environment processing information is pushed.

[0143] Optionally, the preset evaluation threshold is a set standard, such as 80 points, 90 points, etc. When the evaluation value is lower than this standard, it indicates that the environmental conditions are not ideal and measures need to be taken to improve them. The wound healing environment processing information is a specific environmental processing recommendation generated based on the wound healing environment evaluation information, such as adjusting the indoor temperature, increasing ventilation times, using disinfectants to reduce biological contamination, etc. The wound healing environment processing information is pushed to relevant medical staff via email, text message or terminal device application. The purpose of pushing processing information is to remind medical staff to take timely action.

[0144] For example, the preset assessment threshold is 80 points, and the assessment value is 75 points, which is lower than the preset assessment threshold. The wound healing environment processing information may include: "The current temperature in the ward is too high, it is recommended to reduce it to 22 degrees Celsius; the humidity is low, please consider using a humidifier; high biological activity is detected, please strengthen local cleaning."

[0145] By implementing the above process, hospitals can monitor and improve the wound healing environment in real time, thereby improving the speed and quality of patients' recovery.

[0146] In a possible embodiment, the key information further includes the sending time corresponding to the wound monitoring request and the geographical coordinates of the patient; after S17, pushing the nursing strategy information, the method further includes:

[0147] Step 18. In response to the received medical consultation guidance request, determine and push the pharmacy information or hospital information based on the nursing strategy information, the sending time and the patient's geographic coordinates. The pharmacy information includes the pharmacy name and the pharmacy's geographic coordinates. The hospital information includes the hospital name, the hospital's geographic coordinates and the waiting time of the department corresponding to the wound.

[0148] Step 19: In response to the received pharmacy information or hospital information, determine and push the medical treatment route.

[0149] For example, patient A is accidentally injured at home and has a wound on his leg. When the surrounding conditions make it impossible to provide care based on the nursing strategy information (for example, the nursing strategy information is to use medical disinfectant B for disinfection, but patient A does not have medical disinfectant B at home), patient A sends a medical consultation guidance request through the application. When the nursing strategy information contains information such as the need for surgery, hospital information is pushed; when the nursing strategy information contains information such as no surgery is needed, pharmacy information or hospital information is pushed; or, when the nursing strategy information contains information such as no surgery is needed and there are a large number of people queuing in the hospital surgery department, pharmacy information is pushed.

[0150] It should be understood that this embodiment can recommend hospital information of one or more hospitals and / or recommend hospital information of one or more pharmacies. This embodiment does not specifically limit the recommended content.

[0151] Patient A's home is located in the central area of ​​the city, near many hospitals and pharmacies. Patient A cuts at 3 o'clock and sends a wound monitoring request. The nursing strategy information includes the following: Hospital name: Hospital A1, hospital geographic coordinates: longitude and latitude coordinates; the waiting time for the wound corresponding to the department (i.e., surgery) is 20 minutes. Hospital A2, hospital geographic coordinates: longitude and latitude coordinates; the waiting time for the wound corresponding to the department (i.e., surgery) is 10 minutes.

[0152] After patient A selects hospital A2, this embodiment plans the optimal medical treatment route according to the current location of patient A and the geographical coordinates of hospital A2, and pushes it to the patient.

[0153] In this way, the wound care intelligent decision-making system based on high-precision image recognition not only provides immediate care guidance, but also helps patients quickly find suitable medical institutions and provides convenient navigation services, greatly improving the efficiency of medical services and patient satisfaction.

[0154] In a possible embodiment, after receiving the pharmacy information or hospital information, the method further includes:

[0155] Step 201: In response to the received nursing feedback information, key information, nursing strategy information and nursing feedback information are sent to the terminal device of the wound corresponding department in the target pharmacy or the target hospital.

[0156] Step 202: When it is identified that the distance between the patient's geographical coordinates and the geographical coordinates of the target pharmacy or the target hospital is less than a preset distance threshold, a consultation preparation information is sent to a terminal device of a department corresponding to the wound in the target pharmacy or the target hospital.

[0157] The preset distance threshold may be 500 meters, 800 meters, etc.

[0158] Exemplarily, after patient A receives preliminary care based on the nursing strategy information, he sends nursing feedback information to the intelligent decision-making system for wound care based on high-precision image recognition, for example: preliminary disinfection has been performed, but bandaging has not been performed according to the nursing strategy information due to lack of gauze or other reasons. Sending key information, nursing strategy information, and nursing feedback information to the terminal device of the surgical department of hospital A2 accordingly enables medical staff to learn the latest condition of patient A's wound, and continuously update the patient's geographic coordinates when the distance between patient A and hospital A2 is greater than 500 meters. When the distance between patient A and hospital A2 is equal to or less than 500 meters, the consultation preparation information is sent to the terminal device of the surgical department of hospital A2, so that medical staff can prepare gauze and other supplies in advance based on the consultation preparation information.

[0159] In this way, the wound care intelligent decision-making system based on high-precision image recognition not only ensures that the patient's relevant information can be delivered to the hospital in a timely manner, but also notifies the hospital to be prepared when the patient approaches the hospital, thereby improving the efficiency of medical services and the patient's medical experience.

[0160] Figure 2 A structural diagram of a wound care intelligent decision-making system based on high-precision image recognition is provided for an embodiment of the present application, such as Figure 2 As shown, the system includes:

[0161] The extraction module 21 is used to extract key information from the wound monitoring request in response to the received wound monitoring request, where the key information includes: patient identifier, wound image information corresponding to multi-angle high-precision images, wound background information, and multimodal sensor information.

[0162] The acquisition module 22 is used to obtain the patient history information and personalized treatment preference information corresponding to the patient identifier from the preset medical information library; the personalized treatment preference information includes: the patient's preference for drugs and treatment methods, and the treatment methods include drug therapy, surgical treatment or psychotherapy;.

[0163] The recognition module 23 is used to automatically identify the wound status based on the wound image information and the wound background information using a pre-trained three-dimensional convolutional neural network model to obtain a wound recognition result. The convolutional neural network model is obtained by training the wound image information corresponding to the multi-scene high-precision images. The multi-scene high-precision images include high-precision images with different lighting conditions, angles and / or backgrounds.

[0164] The positioning module 24 is used to obtain wound positioning information by using augmented reality technology to assist positioning based on wound image information and wound background information.

[0165] The environmental assessment module 25 is used to perform wound healing environment assessment based on multimodal sensor information to obtain wound healing environment assessment information; the wound healing environment assessment information includes: first assessment information indicating whether the temperature is suitable for wound healing, second assessment information indicating whether the humidity is suitable for wound healing, third assessment information indicating whether the gas concentration is suitable for wound healing, and fourth assessment information indicating whether biological activity affects wound healing.

[0166] The retrieval module 26 is used to retrieve matching nursing strategy information from a preset nursing program database based on patient history information, personalized treatment preference information, wound identification results, wound location information, and wound healing environment assessment information.

[0167] The push module 27 is used to push nursing strategy information.

[0168] Figure 2 The wound care intelligent decision-making system based on high-precision image recognition can perform Figure 1 The implementation principle and technical effect of the intelligent decision-making method for wound care based on high-precision image recognition described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the intelligent decision-making system for wound care based on high-precision image recognition in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.

[0169] In one possible design, Figure 2 The wound care intelligent decision-making system based on high-precision image recognition of the illustrated embodiment can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32 .

[0170] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0171] The processing component 32 is used to: in response to a received wound monitoring request, extract key information from the wound monitoring request, the key information including: a patient identifier, wound image information corresponding to multi-angle high-precision images, wound background information, and multimodal sensor information; obtain patient history information and personalized treatment preference information corresponding to the patient identifier from a preset medical information library; the personalized treatment preference information includes: the patient's preference for drugs and treatment methods, the treatment methods include drug therapy, surgical treatment or psychotherapy; based on the wound image information and wound background information, use a pre-trained three-dimensional convolutional neural network model to automatically identify the wound state to obtain a wound recognition result, the convolutional neural network model is obtained by training the wound image information corresponding to multi-scene high-precision images, and the multi-scene high-precision The images include high-precision images of different lighting conditions, angles and / or backgrounds; based on the wound image information and the wound background information, the wound positioning information is obtained by assisting positioning with augmented reality technology; the wound healing environment is evaluated based on the multimodal sensor information to obtain the wound healing environment evaluation information; the wound healing environment evaluation information includes: first evaluation information indicating whether the temperature is suitable for wound healing, second evaluation information indicating whether the humidity is suitable for wound healing, third evaluation information indicating whether the gas concentration is suitable for wound healing, and fourth evaluation information indicating whether biological activity affects wound healing; based on the patient's historical information, personalized treatment preference information, wound identification results, wound positioning information and wound healing environment evaluation information, matching nursing strategy information is retrieved from a preset nursing plan database; and the nursing strategy information is pushed.

[0172] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0173] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as random access memory (RAM), static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0174] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0175] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0176] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0177] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0178] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment is an intelligent decision-making method for wound care based on high-precision image recognition.

[0179] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0180] The system embodiment described above is merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0181] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0182] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A wound care intelligent decision-making method based on high-precision image recognition, characterized in that: include: In response to the received wound monitoring request, extract key information from the wound monitoring request, the key information including: a patient identifier, wound image information corresponding to the multi-angle high-precision image, wound background information, and multimodal sensor information; Acquire the patient history information and personalized treatment preference information corresponding to the patient identifier from a preset medical information database; the personalized treatment preference information includes: the patient's preference for drugs and treatment methods, and the treatment methods include drug therapy, surgical treatment or psychotherapy; Based on the wound image information and the wound background information, a pre-trained three-dimensional convolutional neural network model is used to automatically identify the wound state to obtain a wound recognition result, wherein the convolutional neural network model is obtained by training the wound image information corresponding to multi-scene high-precision images, and the multi-scene high-precision images include high-precision images with different lighting conditions, angles and / or backgrounds; Based on the wound image information and the wound background information, wound positioning information is obtained by using augmented reality technology to assist positioning; Performing wound healing environment assessment based on multimodal sensor information to obtain wound healing environment assessment information; the wound healing environment assessment information includes: first assessment information indicating whether the temperature is suitable for wound healing, second assessment information indicating whether the humidity is suitable for wound healing, third assessment information indicating whether the gas concentration is suitable for wound healing, and fourth assessment information indicating whether biological activity affects wound healing; Retrieving matching nursing strategy information from a preset nursing program database based on the patient history information, personalized treatment preference information, wound identification results, wound location information, and wound healing environment assessment information; Push the nursing strategy information.

2. The method according to claim 1, characterized in that The wound positioning information includes a target wound boundary and a target wound size; Based on the wound image information and the wound background information, wound positioning information is obtained by using augmented reality technology to assist positioning, including: Based on the wound image information, construct a three-dimensional depth image using a depth estimation algorithm; Setting an augmented reality mark on the wound area of ​​the three-dimensional depth image, wherein the augmented reality mark is a preset pattern or a preset QR code; Segmenting the wound area using a semantic segmentation algorithm to obtain an initial wound boundary; Based on the pixel-level segmentation result of the initial wound boundary, a contour detection algorithm is used to calculate and obtain the initial wound size; Based on the wound background information, the initial wound boundary and the initial wound size are adjusted using a background perception algorithm to obtain a target wound boundary and a target wound size.

3. The method according to claim 2, characterized in that The adjusting the initial wound boundary and the initial wound size using a background perception algorithm based on the wound background information to obtain a target wound boundary and a target wound size includes: Extracting feature information based on the wound background information, wherein the feature information includes: color distribution information, texture pattern information and skin texture information; Based on the feature information, a normal skin area within the initial wound boundary is removed, and a wound area outside the initial wound boundary is filled to obtain a target wound boundary; A size correction is performed based on the target wound boundary to obtain a target wound size.

4. The method according to claim 3, characterized in that The removing of the normal skin area within the initial wound boundary comprises: By fusing the feature information and using an anomaly detection algorithm to identify a normal skin area within the initial wound boundary; The normal skin area within the initial wound boundary is marked to remove the normal skin area within the initial wound boundary.

5. The method according to claim 3, characterized in that: Filling the wound area outside the initial wound boundary comprises: The wound area outside the initial wound boundary is filled using a generative adversarial network technique or a multi-scale context-aware filling technique.

6. The method according to any one of claims 1 to 5, characterized in that: The multimodal sensor information includes temperature sensor information, humidity sensor information, gas sensor information and biosensor information; The wound healing environment assessment is performed based on the multimodal sensor information to obtain wound healing environment assessment information, including: Building a temperature impact model based on the temperature sensor information; Building a humidity impact model based on the humidity sensor information; Building a gas concentration impact model based on the gas sensor information; constructing a biological impact model based on the biosensor information; The wound healing environment is evaluated based on the temperature influence model, humidity influence model, gas concentration influence model and biological influence model to obtain wound healing environment evaluation information.

7. The method according to claim 6, characterized in that After performing wound healing environment assessment based on the temperature impact model, the humidity impact model, the gas concentration impact model and the biological impact model to obtain wound healing environment assessment information, the method further includes: quantifying the wound healing environment assessment information to obtain an assessment value; If the evaluation value is less than the preset evaluation threshold, the wound healing environment processing information is pushed.

8. The method according to any one of claims 1 to 5, characterized in that: The key information also includes the sending time corresponding to the wound monitoring request and the geographical coordinates of the patient; After pushing the nursing strategy information, the method further includes: In response to the received medical consultation guidance request, based on the nursing strategy information, the sending time and the patient's geographical coordinates, determine and push pharmacy information or hospital information, the pharmacy information includes the pharmacy name and the pharmacy's geographical coordinates, and the hospital information includes the hospital name, the hospital's geographical coordinates and the waiting time of the department corresponding to the wound; In response to the received pharmacy information or hospital information, a medical consultation route is determined and pushed.

9. The method according to any one of claims 1 to 5, characterized in that: After receiving the pharmacy information or hospital information, the method further includes: In response to the received nursing feedback information, the key information, the nursing strategy information and the nursing feedback information are sent to a terminal device of a department corresponding to the wound in a target pharmacy or a target hospital; When it is identified that the distance between the patient's geographical coordinates and the geographical coordinates of the target pharmacy or target hospital is less than a preset distance threshold, the medical consultation preparation information is sent to the terminal device of the department corresponding to the wound in the target pharmacy or target hospital.

10. A wound care intelligent decision-making system based on high-precision image recognition, characterized in that: include: an extraction module, configured to extract key information from the wound monitoring request in response to the received wound monitoring request, the key information comprising: a patient identifier, wound image information corresponding to the multi-angle high-precision image, wound background information, and multimodal sensor information; an acquisition module, configured to acquire the patient history information and personalized treatment preference information corresponding to the patient identifier from a preset medical information database; the personalized treatment preference information includes: the patient's preference for drugs and treatment methods, the treatment methods including drug therapy, surgical treatment or psychotherapy; A recognition module, configured to automatically recognize the wound state based on the wound image information and the wound background information using a pre-trained three-dimensional convolutional neural network model to obtain a wound recognition result, wherein the convolutional neural network model is obtained by training wound image information corresponding to multi-scene high-precision images, and the multi-scene high-precision images include high-precision images with different lighting conditions, angles and / or backgrounds; A positioning module, used to obtain wound positioning information by using augmented reality technology to assist positioning based on the wound image information and the wound background information; An environmental assessment module, configured to perform wound healing environment assessment based on multimodal sensor information to obtain wound healing environment assessment information; the wound healing environment assessment information includes: first assessment information indicating whether the temperature is suitable for wound healing, second assessment information indicating whether the humidity is suitable for wound healing, third assessment information indicating whether the gas concentration is suitable for wound healing, and fourth assessment information indicating whether biological activity affects wound healing; A retrieval module, configured to retrieve matching nursing strategy information from a preset nursing program database based on the patient history information, personalized treatment preference information, wound identification results, wound location information, and wound healing environment assessment information; A push module is used to push the nursing strategy information.

Images