Image-based intelligent wound management method and system

Abstract

The invention relates to the technical field of intelligent medical treatment, and discloses an image-based intelligent wound management method and system, and the method comprises the steps that after an operation is completed, a doctor or a nurse uses a shooting device to shoot an image of an operation wound, and the image is preprocessed; automatically identifying a wound area by using a deep learning algorithm, and automatically marking the edge of the wound; based on the image data and a deep learning model, evaluating the healing progress of the wound, and generating a healing progress report; through image analysis and subjective data uploaded by a patient, the system intelligently detects whether a wound has an infection sign or other abnormal conditions; a patient regularly uploads wound images and provides subjective feeling data, and doctors and nurses monitor the healing condition of the wound in real time; after the patient is discharged from the hospital, the patient continues to photograph or record and upload wound images through the mobile device, and the system automatically provides home nursing suggestions and reminds the patient to pay attention to daily nursing. The method has the advantage of improving the wound treatment effect.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical technology, and particularly to an image-based intelligent wound management method and system. Background Technique

[0002] Wound management is a key task in clinical nursing and the medical field, especially in dealing with chronic wounds, diabetic ulcers, burns, and postoperative wounds, etc. Traditional wound assessment and treatment methods usually rely on the subjective judgment of medical staff. This method is not only easily affected by the differences in the experience of observers, but may also lead to diagnostic deviations, thereby affecting the stability and accuracy of the treatment effect. For example, different medical staff may have different judgments on the same wound, resulting in inconsistent treatment strategies. In addition, traditional methods may also face the problem that the change of wound status cannot be immediately feedback. Especially during the wound healing process, the lag of diagnosis and intervention may affect the rehabilitation process of patients. With the rapid development of medical technology, wound management is gradually moving towards digitalization and automation, especially through image processing and computer vision technology. These technologies provide a non-invasive, fast, and accurate assessment tool for wound management. By using image analysis, the type, size, morphological changes, and whether there is infection of the wound can be efficiently identified, thus providing a more accurate assessment basis for doctors. The introduction of an intelligent wound management system can not only reduce human errors, but also monitor the wound healing process in real time and provide personalized treatment plans according to the wound status.

[0003] However, traditional wound management methods often rely on regular inspections or manual records in the hospital, usually limited to the treatment during the patient's hospitalization. After the patient is discharged, especially during the home care stage, the continuous monitoring and management of the wound face huge challenges. Patients after discharge are difficult to obtain timely professional guidance, and the changes of the wound cannot be continuously tracked, which makes it easy to generate monitoring blanks in the wound healing process by traditional methods and difficult to detect potential problems in time. For example, wound infection or abnormal conditions may not be detected in time, resulting in the deterioration of the patient's condition or delayed healing. Therefore, it is necessary to design an image-based intelligent wound management method and system to improve the wound treatment effect. Summary of the Invention

[0004] (1) Technical Problems to be Solved

[0005] In view of the deficiencies of the prior art, the present invention provides an image-based intelligent wound management method and system, which has the advantage of improving the wound treatment effect and solves the problems in the above background technique.

[0006] (2) Technical Solutions

[0007] To achieve the above object of improving the wound treatment effect, the present invention provides the following technical solution: An image-based intelligent wound management method, comprising the following steps:

[0008] S1: After the surgery is completed, a doctor or a caregiver uses a photographing device to photograph the image of the surgical wound and preprocess the image.

[0009] Preferably, the S1 further includes multiple steps of optimizing the quality of the wound image through image preprocessing techniques, including denoising, brightness and contrast adjustment, local area enhancement, automatically segmenting and annotating the wound area using a deep learning algorithm, accurately identifying the boundary of the wound and different tissue layers, and detecting the clarity and exposure of the image through automated quality assessment. If the image quality does not meet the standard, the user will be automatically reminded to retake the photo.

[0010] S2: Use a deep learning algorithm to automatically identify the wound area and automatically annotate the wound edge.

[0011] Preferably, the S2 further includes combining a convolutional neural network and a U-Net structure for feature extraction and segmentation of the wound image, automatically identifying the wound edge, wound surface, and exudate area through a deep learning model. The U-Net architecture is particularly suitable for small sample data sets, and a deep learning-based edge detection algorithm is used, combined with multi-scale image input and multi-channel data.

[0012] S3: Based on the image data and the deep learning model, evaluate the healing progress of the wound and generate a healing progress report.

[0013] Preferably, the S3 further includes combining a deep learning model and a regression analysis method, evaluating the healing progress of the wound by analyzing the area, shape, and color in the wound image data, and comprehensively analyzing the wound image and the patient's physiological data through multi-modal learning. The model identifies different healing stages of the wound and predicts the healing trend through a quantitative scoring method.

[0014] S4: Through image analysis and the subjective data uploaded by the patient, the system intelligently detects whether there are signs of infection or other abnormal conditions in the wound.

[0015] Preferably, the S4 further includes using a deep learning algorithm for anomaly detection of the wound image, extracting the color and texture features of the wound through color space analysis and texture analysis to assist in judging whether there are abnormal conditions, combining the subjective data uploaded by the patient for multi-modal analysis, setting intelligent thresholds to monitor the size and color changes of the wound, automatically issuing an alarm and prompting potential infection or abnormality, and judging whether there are abnormalities during the wound healing process through long-term dynamic monitoring and trend analysis.

[0016] S5: The patient regularly uploads wound images and provides subjective feeling data, and doctors and nurses monitor the wound healing situation in real time.

[0017] Preferably, S5 further includes that the patient regularly uploads wound images and subjective feeling data through a mobile device, automatically synchronizes the data to the cloud or a local server, associates and stores the images and subjective data in the patient's electronic health record, and provides functions of real-time analysis, progress evaluation, infection detection and early warning.

[0018] S6: After discharge, the patient continues to take pictures or videos of the wound through a mobile device and uploads the wound images. The system automatically provides home care suggestions and reminds the patient to pay attention to daily care.

[0019] Preferably, S6 further includes combining a smart phone or a tablet device to support the patient to take pictures or videos and upload the wound images, automatically processing, quality evaluating and detecting the wound area through a cloud platform, building an automatic reminder mechanism based on deep learning and image analysis technologies, regularly sending care reminders to the patient, and providing specific care suggestions according to the patient's living habits. Doctors and nursing staff adjust the care plan and optimize the home care effect through real-time data monitoring, remote medical consultation and interactive communication.

[0020] An image-based intelligent wound management system includes an image acquisition module, a wound recognition module, a healing progress module, an anomaly detection module and a home reminder module;

[0021] The image acquisition module is responsible for, after the operation, doctors or nursing staff using a photographing device to take wound images and performing image preprocessing;

[0022] The wound recognition module uses a deep learning algorithm to automatically recognize the wound area, mark the edge of the wound, and automatically segment the wound area using a convolutional neural network;

[0023] The healing progress module is used to evaluate the healing progress of the wound and generate a healing progress report;

[0024] The anomaly detection module intelligently detects whether there are signs of infection or other abnormal conditions in the wound through image analysis and combining with the subjective data uploaded by the patient;

[0025] The home reminder module automatically generates personalized home care suggestions according to the wound images and the patient's healing progress, and sends reminders to the patient.

[0026] (III) Beneficial effects

[0027] Compared with the prior art, the present invention provides an image-based intelligent wound management method and system, which have the following beneficial effects:

[0028] Through deep learning algorithms, the present invention automatically identifies the wound area, evaluates the healing progress, and combines the subjective data of the patient. The system can intelligently detect whether there is an infection or abnormality in the wound, and promptly remind doctors or caregivers to intervene. After the patient is discharged, they can still continue to upload wound images and subjective data through a mobile device. The system provides personalized home care advice based on the patient's healing condition and automatically sends care reminders to ensure that the patient can perform necessary care operations on time. This method not only improves the accuracy and timeliness of wound treatment, but also reduces the monitoring gaps in traditional care methods, reduces the risk of potential complications, and further improves the patient's treatment effect and nursing satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a schematic diagram of the method of the present invention;

[0030] Figure 2 It is a schematic diagram of the structure of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0031] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0032] The present invention provides a technical solution: an image-based intelligent wound management method, including the following steps:

[0033] S1: After the operation is completed, the doctor or caregiver uses a photographing device to photograph the image of the surgical wound and preprocesses the image.

[0034] During image preprocessing, denoising technology is used to remove noise generated by the shooting equipment or environment. Denoising can reduce random noise or artifacts in the image, improve image clarity, and ensure that the true details of the wound can be accurately presented. By adjusting the brightness and contrast of the image, the details of the wound area are more prominent. Local area enhancement (such as local contrast optimization) is particularly helpful in emphasizing wound areas, such as wound edges and exudate areas, to ensure that these details can be fully identified and analyzed in subsequent healing progress assessments. Automated segmentation and annotation of wound areas are performed using deep learning algorithms. This technology can automatically identify and accurately delineate the boundaries of wounds and different tissue levels, such as wound surfaces, infected areas, or healing areas, reduce errors in manual annotation, and improve processing efficiency. Through automatic evaluation of indicators such as clarity and contrast, it is detected whether the image meets the quality requirements of subsequent analysis. If the image is unclear or the exposure is inappropriate, the system will automatically issue a reminder and suggest that the doctor or caregiver reshoot to ensure the quality and reliability of the image data.

[0035] Standardized shooting and lighting control ensure that the conditions for each wound image acquisition are consistent, thereby improving the repeatability and consistency of image analysis. This is very important for tracking wound healing over a long period of time, ensuring that images taken at different time points can be compared and tracked during analysis. After using denoising technology and image enhancement methods, the details of the wound image will be clearer and more accurate, especially when the shooting conditions are not ideal (such as low light or reflection effects). Preprocessing technology can significantly improve the usability of the image, thereby ensuring that the subsequent analysis algorithm can obtain accurate wound data. Automated region segmentation and annotation technology can accurately extract the boundaries and key features of the wound, helping doctors or systems to accurately locate and evaluate the scope of the wound. Especially in complex wounds or multi-layered wounds, deep learning technology can effectively avoid manual misjudgment and provide more accurate healing progress assessment. The image quality assessment and feedback mechanism can monitor the quality of the image in real time. When the image does not meet the analysis requirements, it automatically reminds the doctor or caregiver to reshoot to avoid analysis errors caused by image quality problems. This mechanism ensures the accuracy of the data and improves the reliability of the wound monitoring system. Through image preprocessing and accurate annotation of the wound area, the subsequent healing progress assessment will be more refined. Doctors can more accurately judge the healing of wounds, detect infection or other abnormal signs in a timely manner, and provide early intervention suggestions, which plays an important role in the early treatment and recovery of patients.

[0036] S2: Use deep learning algorithms to automatically identify wound areas and automatically mark wound edges.

[0037] Through convolutional neural networks for image feature extraction and recognition, deep learning models can extract local and global features from wound images and use them for precise wound area localization. The convolutional neural network learns the spatial features in the image through multiple convolutional layers and pooling layers, and identifies key areas such as wound edges, wound surfaces, and exudate areas. The U-Net structure is used for the segmentation of the wound area. U-Net is a deep learning architecture specifically designed for image segmentation tasks. It efficiently extracts image features through an encoder-decoder structure and gradually restores the spatial resolution in the decoding stage to accurately divide the wound area. U-Net is particularly suitable for small-sample datasets and can effectively handle complex wound image segmentation problems. A deep learning-based edge detection algorithm is adopted to automatically identify the edges of the wound and perform fine annotation. This process ensures that the boundaries of the wound can be accurately located, avoiding manual annotation errors, especially for wounds with complex shapes or difficult to identify. To improve the accuracy of wound recognition, multi-scale image inputs and multi-channel data (such as infrared images, ultraviolet images, etc.) are used, enabling the deep learning model to obtain a complete view of the wound from different levels and angles. This multi-scale input can effectively improve the recognition ability for different types and sizes of wounds. The model can be online trained and updated with real-time data, gradually improving the recognition ability for different patients and different types of wounds. With the accumulation of data, the model can continuously optimize and adapt to different types of surgical wounds, enhancing the generalization ability of recognition.

[0038] Using a convolutional neural network and a U-Net architecture, the deep learning model can accurately identify and segment the wound area. Compared with traditional methods, the deep learning algorithm can handle more complex image structures and details, achieve a higher-precision division of the wound area, and thus improve the reliability of subsequent healing assessment. Through edge detection and annotation techniques, the model can accurately locate and annotate the edges of the wound, ensuring that the shape, size, and position of the wound are accurately reflected in the image. This technology can avoid annotation deviations caused by shooting angles or image quality problems and provide more accurate wound information. Using multi-scale and multi-channel inputs, the deep learning model can comprehensively understand the wound image and obtain key features from different spectral information. Even in complex wound images, the model can improve the recognition and segmentation capabilities for different types and sizes of wounds, enhancing the applicability and generalization of the technology. Automatically identifying the wound area and performing edge annotation greatly reduces the need for manual intervention and improves work efficiency. The traditional manual annotation process is not only time-consuming but also prone to errors. Through deep learning automated identification and annotation, accurate wound data can be provided to doctors in a short time, improving the efficiency of clinical decision-making. Through real-time training and online learning, the deep learning model can continuously optimize with the accumulation of data, improving the accuracy and stability of recognition. This enables the system to adapt to the different situations of various patients and provide more accurate wound assessments over time, thereby supporting the formulation of personalized treatment plans.

[0039] S3: Based on the image data and the deep learning model, evaluate the healing progress of the wound and generate a healing progress report.

[0040] Based on image data, the changes in the wound area, including features such as wound area, shape, color, etc., are analyzed through a deep learning model. These changes can reflect the healing progress of the wound. The model will learn a large amount of labeled data to extract key features that can distinguish different healing stages. Using these features, the system can evaluate the healing status of the wound in real-time and determine which stage it is in (such as the inflammatory stage, proliferative stage, maturation stage). Combining the image features of the wound (such as area, depth, color change) with other auxiliary data (such as the patient's subjective feeling data, such as pain level, body temperature, etc.) for comprehensive analysis. Through multi-modal learning technology, the system can simultaneously consider image data and physiological data provided by the patient for multi-dimensional evaluation to ensure a more comprehensive and accurate assessment of the healing progress. Based on the deep learning model, a regression analysis method is used to quantitatively predict the healing progress of the wound. This method models the change trend of the wound to predict the healing situation of the wound in the future period, helping doctors identify the risk of healing delay or potential complications in advance. The deep learning model divides the wound healing process into multiple stages (such as the initial stage, acceleration stage, maturation stage, etc.), and defines different evaluation criteria for each stage. Each stage combines the appearance changes of the wound (such as wound area, granulation tissue formation, exudate reduction, etc.) for quantitative scoring, and the system automatically gives a comprehensive score of the wound healing according to these criteria, thus generating a detailed healing progress report.

[0041] Based on image data and a deep learning model, the system can accurately evaluate the healing progress of the wound at different time points. Through in-depth analysis of the wound change features (such as size, shape, color, etc.), the model can efficiently and accurately judge whether the wound conforms to the expected healing progress, providing data support for subsequent treatment. By fusing the image data with the patient's subjective feeling data, the system can achieve a more comprehensive assessment of the wound healing progress. For example, combining clinical information such as pain level and body temperature, not only can the healing situation be evaluated from the wound appearance, but also the risk of potential complications can be judged from the patient's overall condition to ensure a more refined assessment. Through the regression analysis method, the system can predict the future trend of wound healing, which is very important for timely adjusting the treatment plan. Doctors can intervene in advance according to the predicted healing time and possible healing delay to reduce the occurrence of complications. Dividing the wound healing process into multiple clear stages makes the evaluation criteria for each stage more specific and operable. Such stage-based evaluation can help doctors clearly understand each step of the wound healing, ensure the refinement and personalization of the treatment plan, and improve the healing effect.

[0042] S4: Through image analysis and the subjective data uploaded by the patient, the system intelligently detects whether there are signs of infection or other abnormal conditions in the wound.

[0043] Through deep learning algorithms, the system can perform anomaly detection on wound images. These algorithms can identify abnormal changes in wound images, such as signs of infection (such as redness, suppuration, color changes in secretions, etc.), necrotic tissue, etc. These deep learning models learn to recognize normal and abnormal wound characteristics through a large amount of training data, improving the accuracy and reliability of detection. In image processing, color space analysis (such as RGB, HSV and other color models) and texture analysis (such as Local Binary Pattern LBP, Gray Level Co-occurrence Matrix GLCM, etc.) are used to identify changes in the wound area. Infections and other abnormal conditions are usually accompanied by specific colors (such as redness, yellow pus, etc.) and textures (such as irregular wound surface structures, the appearance of necrotic tissue). These features help to accurately judge whether there are abnormalities in the wound. The system combines the subjective data uploaded by the patient (such as pain score, body temperature, description of wound secretions, etc.) for comprehensive analysis. The patient's subjective feelings can provide more clinical information for the system, especially common infection symptoms such as pain and fever. By combining with the image analysis results, the system can more accurately assess whether there is an infection or other complications in the wound and timely remind medical staff to intervene. The system sets intelligent thresholds according to existing clinical standards to monitor indicators such as wound size changes and color changes. When these indicators exceed the preset threshold range, the system will automatically issue an alarm and prompt potential infections or abnormalities. This early warning mechanism can effectively improve the response speed of medical staff to wound abnormalities and avoid the aggravation of infection or the deterioration of the condition. The system can conduct long-term dynamic monitoring of wounds. By analyzing the trends of wound images and subjective data at different time points, it can judge whether there are signs of infection or other abnormalities during the wound healing process. For example, if the size of the wound changes rapidly in the short term, or the color remains red and there is pus secretion, the system will automatically identify the abnormality and remind the doctor to deal with it in time.

[0044] Through deep learning and image analysis, the system can accurately detect whether there are signs of infection in the wound and timely remind doctors to take appropriate intervention measures. This not only improves the early detection rate of infections but also reduces the risk of infection spread and patient condition deterioration. By combining image analysis with patients' subjective data, the system can provide a more comprehensive wound assessment, taking into account both the visible physical changes in the image and the patients' subjective symptoms. This multi-dimensional assessment method can more accurately judge the true situation of the wound, thus improving the accuracy of diagnosis and the pertinence of treatment. Through automated detection and intelligent threshold setting, the system can detect subtle abnormal changes in the wound image, which may be easily overlooked in traditional manual examinations. The system can remind doctors through an automatic warning mechanism, detect potential infections or abnormalities in advance, reduce the lag time of treatment, and improve medical efficiency. Dynamic monitoring and trend analysis enable the system to track the wound healing situation over a long period. As time goes by, the healing patterns and trends of the wound are automatically recorded and compared, enabling the timely detection of any abnormalities in the wound progress, especially the early signs of infection, deterioration, or other complications, providing key decision-making support for doctors.

[0045] S5: Patients regularly upload wound images and provide subjective feeling data, and doctors and nurses monitor the wound healing situation in real time.

[0046] The system supports patients to regularly upload wound images through mobile devices. The images can be uploaded in the form of taking photos or videos. The uploaded image data is automatically synchronized to the cloud or local server to ensure that doctors and nurses can view and monitor the healing status of the patient's wound at any time. In addition, the system will record the timestamp of each upload to ensure the accuracy of the wound healing data in chronological order. Patients can input subjective perception data through the mobile application, such as pain score, body temperature, description of wound secretions, etc. The system automatically associates these subjective data with the wound images and generates standardized records. These data are submitted through input methods such as questionnaires, text boxes or sliders to ensure that patients can conveniently provide accurate feedback while reducing the complexity of data processing. Doctors and nurses can view the wound healing images and subjective data through the system's real-time monitoring platform. The platform integrates functions such as wound image analysis, progress assessment, infection detection, etc., provides real-time feedback and issues warnings. Doctors can view the wound healing situation through the platform and make treatment adjustments based on the patient's feedback and image analysis results. The system can dynamically generate feedback and suggestions according to real-time data. All uploaded wound images and subjective perception data will be automatically stored in the patient's electronic health record. The system conducts comparison and trend analysis on the historical records through data analysis technology to identify potential problems during the healing process and provide treatment decision support for doctors. Data analysis also includes the patient's healing progress, pain changes, body temperature changes, etc., to help doctors obtain a more comprehensive wound recovery situation. The system allows multiple doctors and nurses to share the patient's wound healing data to ensure multi-role teamwork. Each medical staff can access the patient's wound data through the permission management system and conduct real-time discussions, and put forward suggestions for different stages of the healing situation. This collaboration function improves the accuracy of treatment decisions and ensures that patients receive multi-faceted professional care.

[0047] Regularly uploading wound images and subjective data enables doctors and nurses to track the wound healing progress of patients in real time, avoiding the risk of failure to detect wound healing delays in a timely manner in traditional methods. The combination of real-time monitoring and data tracking improves the timeliness of medical intervention and adjusts treatment plans promptly. By combining the subjective data uploaded by patients with wound images, the system can provide doctors with comprehensive wound healing reports. These data help doctors better understand the pain level and healing status of patients, thereby formulating personalized treatment plans and improving treatment effects and patient comfort. By monitoring the wound healing progress in real time through the system and combining image analysis and subjective data, the system can issue an alarm immediately when abnormalities are detected. For example, when signs of wound infection, healing delays, or other abnormalities are detected, the system will automatically prompt the doctor and provide possible treatment suggestions. This real-time intervention and early warning function significantly improves the response speed to sudden problems and reduces the risk of patient condition deterioration. The system provides quantitative healing assessment reports and trend predictions by analyzing the uploaded wound images and subjective perception data, assisting doctors in making more accurate treatment decisions. The data-driven decision support system can reduce doctors' judgment errors, especially in cases of complex conditions or multiple treatments, improving treatment efficiency and effects.

[0048] S6: After discharge, the patient continues to upload wound images by taking photos or videos with a mobile device. The system automatically provides home care suggestions and reminds the patient to pay attention to daily care.

[0049] Patients can upload wound images by taking photos or videos using a smartphone or tablet device. The system automatically processes the uploaded images, including image preprocessing, quality assessment, wound area detection, etc., to ensure that the uploaded images are clear and accurate and are automatically optimized according to the image quality. The uploaded images will be stored on the cloud platform for doctors and caregivers to conduct remote monitoring and analysis. Based on deep learning and image analysis technologies, the system combines the wound images uploaded by patients and historical data to intelligently generate home care suggestions. The content of the suggestions includes wound cleaning, dressing change frequency, medications and dressings used, precautions, etc. The system will provide a personalized and dynamically adjusted care plan according to the wound healing progress and the specific situation of the patient. The system has a built-in automatic reminder mechanism that automatically sends care reminders to patients when the wound healing progress or health data of the patient reaches certain criteria. These reminders include scheduled dressing changes, wound cleaning, and regular upload of wound images. The system can also provide specific care precautions and reminders based on factors such as the patient's living environment and lifestyle habits to ensure that patients carry out home care in the best way. After patients regularly upload wound images and health data, the system will analyze the patient's healing progress in real time, generate a diagnosis report, and send the data to doctors and caregivers. Doctors can adjust the home care suggestions in real time based on the data and provide remote medical consultations or further treatment suggestions for patients. The system provides doctors with monitoring tools to provide personalized remote interventions when needed. The system allows for interactive communication between patients and caregivers. Patients can ask questions and provide feedback on their home care experiences through the platform. Caregivers can adjust the care suggestions in a timely manner based on the feedback from patients and provide professional guidance. The platform also supports patients in evaluating the practicality of the care suggestions, and the system is optimized and improved based on the feedback information to enhance the home care effect and satisfaction of patients.

[0050] After discharge, the patient continues to upload wound images through a mobile device to ensure continuous monitoring of the wound healing process. The system can evaluate the wound healing situation in real time and provide feedback, reducing potential nursing blind spots during home care, ensuring that the wound healing progress meets expectations, and reducing the risk of infection or other complications. Based on the patient's wound healing situation, lifestyle habits, as well as the uploaded images and subjective data, the system automatically generates personalized home care suggestions. These suggestions not only improve the accuracy of nursing but also make the patient's home care more scientific and systematic, reducing the occurrence of incorrect operations and helping the patient recover better. The automated reminder mechanism ensures that the patient performs wound care on time and follows medical advice, such as changing dressings regularly and keeping the wound clean. This automated nursing reminder system can significantly improve the patient's treatment compliance and reduce wound infections or delayed healing caused by negligence or non-standard nursing. The system can provide real-time feedback of the patient's wound healing data to doctors and nursing staff. Doctors can conduct remote medical interventions based on the data, timely adjust nursing suggestions or recommend further treatment. This remote medical support enhances the safety of home care and can intervene in a timely manner when the patient shows abnormal conditions to avoid the deterioration of the condition. By regularly uploading images and providing feedback on the effect of home care, the system increases the patient's sense of participation in the treatment process. The patient can see the actual progress of wound healing through the system and obtain professional advice from doctors, making the patient more confident during home treatment and at the same time enhancing the transparency of the treatment. The patient can more intuitively understand the treatment effect and enhance their cooperation with home care.

[0051] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0052] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made in these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An image-based intelligent wound management method, characterized in that: The following steps are involved: S1: After the operation is completed, the doctor or nurse uses a camera to take images of the surgical wound and pre-processes the images; S2: Use deep learning algorithms to automatically identify wound areas and automatically annotate wound edges; S3: Based on image data and deep learning models, evaluate the healing progress of wounds and generate healing progress reports; S4: Through image analysis and subjective data uploaded by patients, the system intelligently detects whether the wound has signs of infection or other abnormalities; S5: Patients regularly upload wound images and provide subjective feeling data, and doctors and nurses monitor the healing of wounds in real time; S6: After discharge, patients continue to upload wound images by taking photos or recording videos through mobile devices. The system automatically provides home care suggestions and reminds patients to pay attention to daily care.

2. The image-based intelligent wound management method according to claim 1, characterized in that: The S1 further includes multiple steps for optimizing the wound image quality through image preprocessing technology, including denoising, brightness and contrast adjustment, local area enhancement, automatic segmentation and labeling of wound areas using deep learning algorithms, accurate identification of wound boundaries and different tissue layers, and detection of image clarity and exposure through automated quality assessment. If the image quality does not meet the standards, the user will be automatically reminded to reshoot.

3. The image-based intelligent wound management method according to claim 1, characterized in that: The S2 further includes combining a convolutional neural network and a U-Net structure to extract and segment features of wound images, and automatically identifying wound edges, wound surfaces, and exudate areas through a deep learning model. The U-Net architecture is particularly suitable for small sample data sets, and uses a deep learning-based edge detection algorithm while combining multi-scale image input and multi-channel data.

4. The image-based intelligent wound management method according to claim 1, characterized in that: The S3 further includes combining a deep learning model and a regression analysis method to evaluate the healing progress of the wound by analyzing the area, shape, and color in the wound image data, and comprehensively analyzing the wound image and the patient's physiological data through multimodal learning. The model identifies different healing stages of the wound and predicts the healing trend through a quantitative scoring method.

5. The image-based intelligent wound management method according to claim 1, characterized in that: The S4 further includes using a deep learning algorithm to detect abnormalities in wound images, extracting the color and texture features of the wound through color space analysis and texture analysis, assisting in determining whether there are abnormalities, combining the subjective data uploaded by the patient to perform multimodal analysis, setting intelligent thresholds to monitor the size and color changes of the wound, automatically issuing alarms and prompting potential infections or abnormalities, and determining whether abnormalities occur during the wound healing process through long-term dynamic monitoring and trend analysis.

6. The image-based intelligent wound management method according to claim 1, characterized in that: The S5 further includes patients regularly uploading wound images and subjective perception data via mobile devices, and automatically synchronizing the data to the cloud or local server, associating and storing the images and subjective data in the patient's electronic health record, and providing real-time analysis, progress assessment, infection detection and early warning functions.

7. The image-based intelligent wound management method according to claim 1, characterized in that: The S6 further includes a combination with a smartphone or tablet device to support patients in taking photos or recording videos to upload wound images, and automatically processes, evaluates quality and detects wound areas through a cloud platform. Based on deep learning and image analysis technology, it has a built-in automatic reminder mechanism to regularly send nursing reminders to patients and provide specific nursing suggestions based on the patient's living habits. Doctors and caregivers adjust nursing plans and optimize home care effects through real-time data monitoring, remote medical consultation and interactive communication.

8. An image-based intelligent wound management system, applied to the image-based intelligent wound management method according to claims 1-7, characterized in that: It includes image acquisition module, wound recognition module, healing progress module, abnormality detection module and home reminder module; The image acquisition module is responsible for taking wound images with a camera after surgery, and performing image preprocessing; The wound recognition module uses a deep learning algorithm to automatically identify the wound area, mark the edge of the wound, and use a convolutional neural network to automatically segment the wound area; The healing progress module is used to evaluate the healing progress of the wound and generate a healing progress report; The abnormality detection module intelligently detects whether the wound has signs of infection or other abnormalities through image analysis and combined with subjective data uploaded by the patient; The home reminder module automatically generates personalized home care suggestions based on the wound image and the patient's healing progress, and sends a reminder to the patient.

Images