Digestive endoscopy postoperative intelligent nursing system based on deep learning

Abstract

The invention relates to the technical field of medical care, in particular to a digestive endoscopy postoperative intelligent nursing system based on deep learning. The system comprises an endoscopic postoperative data acquisition and processing module, an endoscopic postoperative feature analysis module, an endoscopic postoperative recovery risk assessment module and a postoperative recovery risk intelligent nursing module, and can acquire physiological data of a patient after a digestive endoscopic surgery and digestive tract state image data after the endoscopic surgery; the long and short time memory network is used for obtaining the long and short time sequence characteristics of the physiological indexes after the endoscopic surgery, the convolutional neural network is used for identifying and obtaining the healing abnormal lesion characteristics after the endoscopic surgery, the characteristics are input into the preset deep neural network model for characteristic fusion, and postoperative recovery risk assessment calculation and postoperative intelligent nursing analysis are carried out. And generating digestive endoscopy postoperative patient recovery nursing suggestions so as to execute corresponding digestive endoscopy postoperative risk recovery nursing operation. According to the invention, postoperative data of the patient can be accurately analyzed, so that nursing decision-making work can be intelligently guided.

Description

Technical Field

[0001] The present invention relates to the field of medical care technology, and in particular to an intelligent nursing system for post-operative digestive endoscopy based on deep learning. Background Art

[0002] As a minimally invasive method, digestive endoscopy has been widely used in the analysis and care of gastrointestinal diseases. Although this technology has the advantages of less trauma and faster recovery, the quality of postoperative care directly affects the patient's recovery process and the occurrence of complications. In recent years, technologies based on artificial intelligence and deep learning have been widely used in the medical field, especially in image recognition and patient monitoring. Deep learning models can extract potential rules and patterns from a large amount of patient data and provide accurate prediction and decision support. In postoperative care after digestive endoscopy, the introduction of deep learning technology can provide more intelligent and personalized nursing services for postoperative patients. However, the current deep learning applications are mainly concentrated in the fields of image analysis and risk prediction. Intelligent methods for postoperative care have not been fully studied. The patient's vital signs, laboratory test results, postoperative symptoms and other information have not been fully utilized for comprehensive dynamic monitoring and risk prediction, making it difficult to achieve precision and personalization in the postoperative care process, and it is difficult to respond to the complex changes of patients in the postoperative recovery process in real time, thereby reducing the corresponding nursing efficiency of postoperative patients after digestive endoscopy. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide an intelligent postoperative care system for digestive endoscopy based on deep learning to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a deep learning-based intelligent nursing system for postoperative digestive endoscopy includes the following modules:

[0005] The post-endoscopic data acquisition and processing module is used to obtain the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy through the API interface, and to perform post-endoscopic data preprocessing and standardization on the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy, so as to obtain the physiological standardized data after digestive endoscopy and the standardized imaging data of the digestive tract after endoscopy;

[0006] The post-endoscopic feature analysis module is used to extract long-short time series features from the standardized physiological data after digestive endoscopy using the long-short time memory network to obtain the long-short time series features of post-endoscopic physiological indicators; the convolutional neural network is used to identify the post-endoscopic healing lesion features of the standard imaging data of the digestive tract after endoscopy to obtain the abnormal healing lesion features after endoscopy, including the degree of lesions corresponding to post-endoscopic digestive tract healing bleeding, ulcers and infections;

[0007] The post-endoscopic recovery risk assessment module is used to input the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal post-endoscopic healing lesions into the preset deep neural network model for feature fusion to generate a post-digestive endoscopy fusion feature set; perform post-digestive endoscopy recovery risk assessment calculation on the post-digestive endoscopy fusion feature set to obtain the degree of post-digestive endoscopy recovery risk;

[0008] The postoperative recovery risk intelligent nursing module is used to determine the corresponding recovery risk level of patients after digestive endoscopy according to the degree of recovery risk after digestive endoscopy, and to perform postoperative intelligent nursing analysis based on the corresponding recovery risk level of patients after digestive endoscopy, and to generate recovery nursing recommendations for patients after digestive endoscopy, so as to perform corresponding postoperative risk recovery nursing tasks after digestive endoscopy.

[0009] Furthermore, the post-endoscopic data acquisition and processing module includes the following functions:

[0010] A data collection security channel is built between the hospital's internal network and the external data requester using blockchain-based encryption technology and digital certificate authentication, so that the corresponding post-digestive endoscopy patient data interaction log is recorded through the blockchain's distributed ledger, and the external data requester is authenticated using a digital certificate to generate a post-digestive endoscopy data collection security channel.

[0011] Obtain the data characteristics of patients after digestive endoscopy, including the field storage accuracy and timestamp format of the patient's physiological data, and the resolution and encoding method of the endoscopic image data. Combined with the data structure of the hospital information system and the digestive endoscopy image storage system, finely adapt the API interface parameters corresponding to the external data request end to generate a data request adaptation API interface parameter set, which includes data screening conditions and data format conversion rules.

[0012] Based on the digestive endoscopy post-operative data collection security channel and the data request adaptation API interface parameter set and using the authorized external data request end to send an acquisition request to the corresponding hospital information system and digestive endoscopy image storage system in the hospital internal network, and obtain the patient's physiological data after digestive endoscopy and the digestive tract status image data after endoscopy;

[0013] Postoperative data preprocessing and standardization were performed on the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy to obtain physiological standardized data after digestive endoscopy and standard imaging data of the digestive tract after endoscopy.

[0014] Furthermore, the postoperative data preprocessing and standardization of the physiological data of the patient after digestive endoscopy and the postoperative digestive tract status imaging data after endoscopy includes:

[0015] The physiological data of patients after digestive endoscopy were abnormally removed and standardized to obtain physiological standardized data after digestive endoscopy;

[0016] Performing image data enhancement on the digestive tract status image data after endoscopy, including rotation, translation and flipping, to obtain enhanced image data of the digestive tract after endoscopy;

[0017] Perform grayscale histogram equalization on the enhanced image data of the digestive tract after endoscopy to obtain contrast-balanced image data of the digestive tract after endoscopy;

[0018] Perform pixel blurriness analysis on the contrast-balanced image data of the post-endoscopic digestive tract to obtain the pixel blurriness value of the post-endoscopic digestive tract image; perform image denoising on the corresponding contrast-balanced image data of the post-endoscopic digestive tract based on the pixel blurriness value of the post-endoscopic digestive tract image to obtain the denoised image data of the post-endoscopic digestive tract;

[0019] The denoised image data of the digestive tract after endoscopy were standardized to obtain the standard image data of the digestive tract after endoscopy.

[0020] Furthermore, the post-endoscopic feature analysis module includes the following functions:

[0021] Perform frequency domain conversion on physiological indexes of normalized data after digestive endoscopy to obtain the periodic rhythm spectrum of physiological indexes after digestive endoscopy.

[0022] The dynamic trend key node analysis of the cyclical rhythm spectrum of physiological indicators after digestive endoscopy is performed based on the slope change detection and extreme point tracking method, so as to analyze the slope changes corresponding to each physiological indicator curve along the time axis. When the slope mutation exceeds the preset threshold, it is marked as a potential key node, and the maximum and minimum points corresponding to each physiological indicator curve are tracked at the same time to obtain the dynamic trend key node set of physiological indicators after endoscopy.

[0023] According to the key node set of dynamic trends of post-endoscopic physiological indicators, long-scale time windows and short-scale time windows are determined on the time axis within the periodic rhythm spectrum of post-digestive endoscopy physiological indicators, and the corresponding post-digestive endoscopy physiological standardized data are divided into long- and short-window sequences based on the long-scale time window and the short-scale time window to obtain endoscopic surgery physiological long-time window sequence data and endoscopic surgery physiological short-time window sequence data;

[0024] Long short-term memory network is used to extract long and short time series features from the physiological long-term window sequence data and the physiological short-term window sequence data of endoscopic surgery, so as to obtain the long and short time series features of physiological indicators after endoscopy, including the low-frequency trend change features corresponding to each physiological indicator after endoscopy in the long-term window and the high-frequency fluctuation features corresponding to the short-term window;

[0025] Convolutional neural network was used to identify the postoperative healing lesion characteristics of standard post-endoscopic digestive tract imaging data, and the abnormal post-endoscopic healing lesion characteristics were obtained, including the post-endoscopic digestive tract healing bleeding points and the corresponding lesion degree of ulcer.

[0026] Furthermore, the simultaneous tracking of the maximum and minimum points corresponding to each physiological indicator curve includes:

[0027] Perform curve fluctuation frequency characteristic analysis on each physiological index curve in the periodic rhythm spectrum of post-digestive endoscopy physiological indexes to obtain the curve fluctuation frequency characteristics corresponding to each post-endoscopic physiological index;

[0028] Based on the curve fluctuation frequency characteristics corresponding to each post-endoscopic physiological index, each physiological index curve in the periodic rhythm spectrum of the post-endoscopic physiological index is divided into curve fluctuation pattern regions to generate curve fluctuation pattern sub-regions corresponding to each post-endoscopic physiological index;

[0029] The golden section search method is used to perform extreme value search processing on the curve fluctuation mode sub-regions corresponding to each post-endoscopic physiological index, and the curve fluctuation extreme value point search interval corresponding to each post-endoscopic physiological index is obtained;

[0030] The parabolic approximation method was used to perform curve local extreme point fitting calculation on the curve fluctuation extreme point search interval corresponding to each post-endoscopic physiological index, so as to obtain the maximum and minimum points corresponding to each physiological index curve.

[0031] Furthermore, the use of a convolutional neural network to identify postoperative healing lesion features of standard post-endoscopic digestive tract imaging data includes:

[0032] Obtain the absorption and scattering characteristics of the digestive tract tissue under different wavelengths of light through the standard image data of the digestive tract after endoscopic surgery, and perform image optical parameter analysis on the corresponding standard image data of the digestive tract after endoscopic surgery based on the absorption and scattering characteristics of the digestive tract tissue under different wavelengths of light combined with a preset light scattering model to obtain the absorption coefficient and scattering coefficient at each pixel point in the image of the digestive tract after endoscopic surgery;

[0033] Based on the absorption coefficient and scattering coefficient at each pixel point in the post-endoscopic digestive tract image, the corresponding post-endoscopic digestive tract standard image data is subjected to image texture fractal structure recognition to generate the post-endoscopic digestive tract image fractal feature structure, including the mucosal folds and vascular branch fractal feature structure; the corresponding post-endoscopic digestive tract image fractal dimension is obtained according to the post-endoscopic digestive tract image fractal feature structure;

[0034] By using the gray-level co-occurrence matrix to analyze the gray-scale texture characteristics of the standard image data of the post-endoscopic digestive tract, the distribution of gray-scale texture characteristics between pixel pairs in different directions and distances of the post-endoscopic digestive tract images was obtained.

[0035] Based on the fractal dimension of post-endoscopic digestive tract images and the grayscale texture feature distribution between pixel pairs in different directions and distances, the semantic segmentation network model is used to segment the corresponding post-endoscopic digestive tract standard image data into image anatomical regions, and obtain the images of the anatomical structure regions of each post-endoscopic digestive tract.

[0036] Convolutional neural networks were used to extract the features of postoperative healing lesions from the images of various post-endoscopic gastrointestinal anatomical structure regions, and the bleeding points of post-endoscopic gastrointestinal healing, the size of the ulcer surface of post-endoscopic gastrointestinal lesions, and the changes in the depth of post-endoscopic gastrointestinal lesions were obtained. The corresponding post-endoscopic gastrointestinal lesion ulcer expansion rate was obtained according to the changes in the depth of post-endoscopic gastrointestinal lesions. Based on the size of the ulcer surface and the expansion rate of post-endoscopic gastrointestinal lesions, the ulcer lesion degree calculation formula was used to perform lesion measurement calculation on the corresponding post-endoscopic gastrointestinal images, and the corresponding lesion degree of the post-endoscopic gastrointestinal ulcer was obtained.

[0037] Furthermore, the calculation formula for the degree of ulcer lesions is specifically:

[0038]

[0039] Where D is the lesion degree corresponding to the post-endoscopic digestive ulcer, Ω is the post-endoscopic digestive tract imaging area, and A xy is the size of the ulcer surface of the post-endoscopic digestive tract lesion at position (x, y), λ1 is the attenuation factor of the ulcer area, and R xy is the expansion rate of post-endoscopic digestive tract ulcer at position (x, y), λ2 is the attenuation factor of ulcer expansion rate, α1 is the weighting coefficient of post-endoscopic digestive tract ulcer, n is the number of image features corresponding to post-endoscopic digestive tract images, i is the item index corresponding to the image feature, and f i is the i-th image feature corresponding to the post-endoscopic digestive tract image, α2 is the image feature lesion weighting coefficient, and η is the correction coefficient of the lesion degree.

[0040] Furthermore, the post-endoscopic recovery risk assessment module includes the following functions:

[0041] The corresponding characteristic means and characteristic variances were obtained through the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal healing lesions after endoscopic surgery, and the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal healing lesions after endoscopic surgery were Z-score standardized based on the characteristic means and characteristic variances to obtain the standard characteristics of the long and short time series characteristics of post-endoscopic physiological indicators and the standard characteristics of abnormal healing lesions after endoscopic surgery;

[0042] Mutual information analysis is performed between each dimensional feature within the standard features of the length of physiological indicators after endoscopy and each characteristic component within the standard features of abnormal lesions after endoscopy, so as to obtain the mutual information between each physiological dimensional feature after endoscopy and each characteristic component of abnormal lesions; nonlinear correlation evaluation analysis is performed between each dimensional feature within the standard features of the length of physiological indicators after endoscopy and each characteristic component within the standard features of abnormal lesions after endoscopy according to the mutual information between each physiological dimensional feature after endoscopy and each characteristic component of abnormal lesions, so as to obtain the nonlinear correlation between each physiological dimensional feature after endoscopy and each characteristic component of abnormal lesions;

[0043] The linear correlation evaluation and analysis was performed on the various dimensional characteristics within the standard characteristics of the length of post-endoscopic physiological indicators and the various characteristic components within the standard characteristics of post-endoscopic abnormal lesions, and the linear correlation coefficients between the post-endoscopic physiological dimensional characteristics and the characteristic components of the abnormal lesions were obtained.

[0044] The long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal healing lesions after endoscopy are input into the preset deep neural network model, and the nonlinear correlation and linear correlation coefficient between the post-endoscopic physiological dimension characteristics and the characteristic components of each abnormal lesion are combined to perform feature fusion to generate a post-digestive endoscopy fusion feature set;

[0045] The postoperative recovery risk assessment calculation formula was used to evaluate the postoperative recovery risk of digestive endoscopy postoperative fusion feature set to obtain the degree of postoperative recovery risk of digestive endoscopy.

[0046] Furthermore, the postoperative recovery risk assessment calculation formula is specifically:

[0047]

[0048] Where R is the risk level of postoperative recovery after digestive endoscopy, T is the postoperative observation time range, t is the time variable parameter, P(t) is the time series characteristics of the postoperative physiological index corresponding to the patient at time point t after surgery, Q(t) is the time series characteristics of the abnormal healing lesions corresponding to the patient at time point t after surgery, β is the risk weight coefficient of recovery of abnormal healing lesions, H(t) is the long and short time series characteristics of the postoperative physiological index corresponding to the patient at time point t after surgery, γ is the risk weight coefficient of recovery of postoperative physiological indexes, ε is the risk influencing factor of postoperative recovery after digestive endoscopy, m is the total number of individual characteristics of patients after digestive endoscopy, X j is the specific value corresponding to the individual characteristics of the jth patient after digestive endoscopy, φ j is the weight coefficient corresponding to the individual characteristics of the jth patient after digestive endoscopy, and ξ is the correction coefficient of the risk of recovery after digestive endoscopy.

[0049] Furthermore, the postoperative recovery risk intelligent nursing module includes the following functions:

[0050] Obtain clinical experience after digestive endoscopy and determine the risk grading system for post-endoscopic recovery based on clinical experience after digestive endoscopy;

[0051] Based on the post-endoscopic recovery risk grading system, the recovery risk level of patients after digestive endoscopy is determined according to the degree of recovery risk after digestive endoscopy, including low recovery risk level, medium recovery risk level and high recovery risk level;

[0052] Postoperative intelligent nursing analysis is performed based on the corresponding recovery risk level of patients after gastrointestinal endoscopy, and recovery nursing recommendations for patients after gastrointestinal endoscopy are generated to execute corresponding risk recovery nursing tasks after gastrointestinal endoscopy.

[0053] Beneficial effects of the present invention:

[0054] The deep learning-based postoperative intelligent nursing system for digestive endoscopy proposed in the present invention is generally composed of a postoperative data acquisition and processing module, a postoperative feature analysis module, a postoperative recovery risk assessment module and a postoperative recovery risk intelligent nursing module. Compared with the prior art, the beneficial effect of the present application lies in that by using the API interface to automatically acquire the physiological data of patients after digestive endoscopy and the imaging data of the postoperative digestive tract status, real-time monitoring and evaluation of the patient's postoperative status can be achieved. Physiological data such as heart rate, blood pressure, body temperature, etc., and imaging data such as digestive tract images taken by endoscopy can help doctors fully understand the patient's postoperative condition. Data preprocessing and standardization are important links in data analysis, which can eliminate noise and deviation from different data sources, so that the data has a unified standard format, which is convenient for subsequent analysis and processing. After standardization, the data will be more comparable, which will help improve the accuracy and reliability of subsequent analysis results. For example, the standardization of physiological data can eliminate the impact of individual differences, allowing the model to focus on pathological changes rather than simple physiological fluctuations; the standardization of imaging data can improve the contrast and clarity of the image, making subsequent image recognition more accurate, thereby improving the accuracy of lesion recognition, providing a high-quality data foundation for subsequent deep learning models, and ensuring the accuracy and reliability of prediction and evaluation. Secondly, by using long short-term memory networks to extract time series features from physiological standardized data, in postoperative monitoring, physiological data usually have time series characteristics, and single moment data often cannot fully reflect the patient's recovery status. Using long short-term memory networks (LSTM) to extract long and short time series features of physiological data can capture the long-term and short-term dependencies of physiological parameter changes, which helps to deeply understand the patient's recovery trend and potential risks. For example, LSTM can identify risk warning signals in the postoperative recovery process, such as the occurrence of acute complications, by learning the change patterns of multiple physiological indicators such as the patient's heart rate, blood pressure, and body temperature at different time points. At the same time, through the analysis of post-endoscopic gastrointestinal imaging data, the convolutional neural network (CNN) can be used to accurately identify lesion characteristics. CNN extracts detailed information from the image through multiple convolutional layers and can effectively identify abnormal healing lesions in the digestive tract, such as areas of bleeding, ulcers, or infection. More importantly, CNN can assess the severity of these lesions and provide more detailed guidance for postoperative recovery. Through in-depth analysis of two data forms (physiological data and imaging data), a comprehensive assessment of the postoperative recovery status can be achieved, effectively predicting potential complications or risks.Then, by inputting the long and short time series features and the healing abnormal lesion features into the deep neural network model for feature fusion, a more comprehensive and comprehensive recovery assessment feature set can be obtained. This fusion can not only bring together the advantages of physiological data and imaging data, but also perform intelligent analysis through the deep neural network model to mine the complex relationship between different data. The data set after feature fusion can provide richer contextual information for recovery risk assessment and help reveal the interaction of multiple factors in the patient's recovery process. For example, changes in physiological characteristics are caused by lesions inside the digestive tract, while imaging data can accurately reveal the presence and severity of lesions. Through the calculation of the deep learning model, an accurate assessment of recovery risk can be achieved, thereby providing a scientific basis for postoperative care. The deep fusion of this step helps to personalize the recovery process of patients after digestive endoscopy, improve the accuracy of postoperative risk prediction, and help doctors identify potential high-risk patients in advance and carry out targeted interventions on them. Finally, different levels of nursing intervention can be performed according to the patient's recovery risk level to ensure that nursing resources are reasonably allocated and to avoid excessive or insufficient intervention. For example, low-risk patients can receive routine care, while high-risk patients require enhanced monitoring and intervention. Intelligent nursing analysis integrates the patient's personalized data to provide targeted nursing advice, such as diet adjustment, medication use, exercise advice, etc., thereby promoting the patient's rapid recovery after surgery and reducing the occurrence of complications. In addition, intelligent nursing analysis based on recovery risk can also help hospitals improve nursing efficiency, optimize nursing processes, reduce unnecessary manual intervention, reduce medical costs, and enhance patients' nursing experience. Through this intelligent nursing analysis, the patient's postoperative recovery process can be managed in a refined and personalized manner, ensuring that the patient receives the most suitable recovery care plan, effectively improving the quality and effectiveness of postoperative care, and thus significantly improving the quality and efficiency of postoperative care. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0056] Figure 1 This is a module schematic diagram of the intelligent nursing system for postoperative digestive endoscopy based on deep learning of the present invention;

[0057] Figure 2 for Figure 1 Schematic diagram of the functional flow of the post-endoscopic data acquisition and processing module;

[0058] Figure 3 for Figure 1 Schematic diagram of the functional flow of the post-endoscopic feature analysis module. DETAILED DESCRIPTION

[0059] The technical system of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0060] To achieve this, please refer to Figures 1 to 3 The present invention provides an intelligent nursing system for postoperative digestive endoscopy based on deep learning, and the system includes the following modules:

[0061] The post-endoscopic data acquisition and processing module is used to obtain the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy through the API interface, and to perform post-endoscopic data preprocessing and standardization on the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy, so as to obtain the physiological standardized data after digestive endoscopy and the standardized imaging data of the digestive tract after endoscopy;

[0062] The post-endoscopic feature analysis module is used to extract long-short time series features from the standardized physiological data after digestive endoscopy using the long-short time memory network to obtain the long-short time series features of post-endoscopic physiological indicators; the convolutional neural network is used to identify the post-endoscopic healing lesion features of the standard imaging data of the digestive tract after endoscopy to obtain the abnormal healing lesion features after endoscopy, including the degree of lesions corresponding to post-endoscopic digestive tract healing bleeding, ulcers and infections;

[0063] The post-endoscopic recovery risk assessment module is used to input the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal post-endoscopic healing lesions into the preset deep neural network model for feature fusion to generate a post-digestive endoscopy fusion feature set; perform post-digestive endoscopy recovery risk assessment calculation on the post-digestive endoscopy fusion feature set to obtain the degree of post-digestive endoscopy recovery risk;

[0064] The postoperative recovery risk intelligent nursing module is used to determine the corresponding recovery risk level of patients after digestive endoscopy according to the degree of recovery risk after digestive endoscopy, and to perform postoperative intelligent nursing analysis based on the corresponding recovery risk level of patients after digestive endoscopy, and to generate recovery nursing recommendations for patients after digestive endoscopy, so as to perform corresponding postoperative risk recovery nursing tasks after digestive endoscopy.

[0065] In the embodiment of the present invention, please refer to Figure 1 As shown, it is a module schematic diagram of the intelligent nursing system for postoperative digestive endoscopy based on deep learning of the present invention. In this example, the intelligent nursing system for postoperative digestive endoscopy based on deep learning includes the following modules:

[0066] S1: Post-endoscopic data acquisition and processing module, used to obtain the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy through the API interface, and perform post-endoscopic data preprocessing and standardization on the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy, so as to obtain the physiological standardized data after digestive endoscopy and the standardized imaging data of the digestive tract after endoscopy;

[0067] In an embodiment of the present invention, the physiological data of patients after digestive endoscopy and the imaging data of the post-endoscopic digestive tract status are obtained from the electronic health record system or device interface of the hospital by using an API interface. The physiological data generally include the patient's blood pressure, heart rate, body temperature, blood oxygen saturation, respiratory rate and other vital signs indicators. These data are presented in the form of time series and contain some missing values ​​or abnormal data. Therefore, data cleaning is essential. First, missing values ​​are filled by interpolation or mean filling. Outliers can be identified and corrected by box plots or standard deviation methods. After that, all physiological data are standardized, and the numerical range of each feature is scaled to between 0 and 1. The standardization method used is Z-score standardization. For imaging data, common image preprocessing methods used in deep learning are used, such as adjusting image size, normalization, grayscale, etc. The image is standardized by pixel mean and standard deviation so that the pixel value of the image falls between 0 and 1. The preprocessing of imaging data is performed by appropriate cropping and adjustment before input into the convolutional neural network (CNN) to make it meet the dimensional requirements of the model input. Through these preprocessing methods, standardized postoperative physiological data and imaging data after digestive endoscopy are obtained, and finally standardized physiological data after digestive endoscopy and standard imaging data of the digestive tract after endoscopy are obtained.

[0068] S2: Post-endoscopic feature analysis module, used to extract long-short time series features from standardized physiological data after digestive endoscopy using long-short time memory network, so as to obtain long-short time series features of post-endoscopic physiological indicators; use convolutional neural network to identify post-endoscopic healing lesion features from standard imaging data of the digestive tract after endoscopy, and obtain abnormal healing lesion features after endoscopy, including the degree of lesions corresponding to post-endoscopic digestive tract healing bleeding, ulcers and infections;

[0069] In an embodiment of the present invention, the standardized post-digestive endoscopy physiological data is processed by a long short-term memory network (LSTM) to extract time series features. The LSTM model can capture long-term dependencies and time series patterns in time series data, and is therefore suitable for analyzing physiological changes in patients after digestive endoscopy. In model construction, a multi-layer LSTM structure is selected, and the number of units in each layer of LSTM is 128. A bidirectional LSTM structure is used to enhance feature extraction capabilities. The input data is the patient's post-operative physiological data. The network effectively transfers past information to the current moment through a long short-term memory mechanism, thereby extracting time series features of physiological indicators. For post-operative gastrointestinal imaging data, A convolutional neural network (CNN) was used for feature recognition. CNN extracted local features from the image through several convolutional layers, and then downsampled through the pooling layer to obtain high-level features of the image. The input of the model was postoperative gastrointestinal imaging data. After processing by the convolutional layer, the features of endoscopic postoperative healing lesions were extracted, including but not limited to the manifestations of abnormal lesions such as bleeding, ulcers, and infection. The model was initialized with the pre-trained VGG16 model, and then fine-tuned on the postoperative imaging data to make the model more suitable for specific postoperative imaging features. Finally, the features of abnormal endoscopic postoperative healing lesions were obtained, including the degree of lesions corresponding to postoperative gastrointestinal healing bleeding, ulcers, and infections.

[0070] S3: Endoscopic postoperative recovery risk assessment module, used to input the long and short time series characteristics of postoperative physiological indicators and the characteristics of abnormal postoperative healing lesions into the preset deep neural network model for feature fusion to generate a postoperative fusion feature set for digestive endoscopy; perform postoperative recovery risk assessment calculation on the postoperative fusion feature set for digestive endoscopy to obtain the degree of postoperative recovery risk for digestive endoscopy;

[0071] In an embodiment of the present invention, a fusion feature set after digestive endoscopy is obtained by fusing physiological data features and image data features. First, the extracted long and short time series features of post-endoscopic physiological indicators are merged with the features of abnormal post-endoscopic healing lesions, and a fully connected layer (FC Layer) is used to connect them together to form a fusion feature vector. Subsequently, the fusion feature is input into a deep neural network model for deeper feature learning and combination. In the design of the deep neural network, a multi-layer perceptron (MLP) structure is adopted, which includes three hidden layers, each with 256 neurons, and ReLU (Rectified Linear Unit) is selected as the activation function. At this time, the output layer adopts a sigmoid activation function to map the fusion feature to the range of [0,1], indicating the probability of recovery risk. After this process, the result of the post-endoscopic recovery risk assessment is obtained, wherein the degree of recovery risk reflects the potential situation of the patient's postoperative recovery. A higher value indicates a greater recovery risk, and a lower value indicates a better recovery situation. Finally, the degree of recovery risk after digestive endoscopy is obtained.

[0072] S4: Postoperative recovery risk intelligent nursing module is used to determine the corresponding recovery risk level of patients after digestive endoscopy according to the degree of recovery risk after digestive endoscopy, and perform postoperative intelligent nursing analysis according to the corresponding recovery risk level of patients after digestive endoscopy, and generate recovery nursing suggestions for patients after digestive endoscopy to perform corresponding postoperative risk recovery nursing tasks after digestive endoscopy.

[0073] In an embodiment of the present invention, the recovery risk level of patients after digestive endoscopy is determined by setting a specific threshold value according to the recovery risk degree calculated in the aforementioned steps, and a standard for recovery risk scoring is set. For example, if the recovery risk probability is greater than 0.85, it is determined to be high risk; if the recovery risk probability is between 0.46 and 0.85, it is determined to be medium risk; if the recovery risk probability is less than 0.45, it is determined to be low risk. According to the patient's recovery risk level, a personalized postoperative intelligent nursing analysis is further generated. The intelligent nursing analysis will integrate the patient's physiological data, imaging data and risk level to generate specific nursing recommendations. The nursing recommendations for high-risk patients include more frequent monitoring, adjustment of drug treatment plans, and enhanced postoperative care; medium-risk patients require routine care and regular review; low-risk patients can receive routine care to avoid excessive intervention, and the patient's recovery care recommendations are provided to nursing staff and doctors through the hospital's information system to ensure that postoperative care can be effectively implemented, minimize the probability of postoperative complications, and ensure that the patient's postoperative recovery proceeds smoothly. Finally, recovery care recommendations for patients after digestive endoscopy are generated to perform corresponding postoperative risk recovery nursing tasks after digestive endoscopy.

[0074] Further, as an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 The functional flow diagram of the post-endoscopic data acquisition and processing module is shown in FIG. 1 . In this embodiment, the post-endoscopic data acquisition and processing module includes the following functions:

[0075] S11: Use blockchain-based encryption technology and digital certificate authentication to build a data collection security channel between the hospital's internal network and the external data request end, so as to record the corresponding digestive endoscopy patient data interaction log through the blockchain's distributed ledger, and use digital certificates to perform identity authentication on the external data request end to generate a digestive endoscopy postoperative data collection security channel;

[0076] In an embodiment of the present invention, by constructing a data collection security channel between the hospital's internal network and the external data request end, it is first necessary to introduce blockchain-based encryption technology and digital certificate authentication mechanism. The construction of this security channel relies on the distributed ledger function of the blockchain, which is used to record detailed logs of each data interaction. The specific steps include creating a private chain through a blockchain platform (such as Ethereum or Hyperledger, etc.), defining interaction behaviors and operation record nodes related to post-digestive endoscopy patient data, and verifying the identity of the external data request end through a smart contract each time a data request or data transmission occurs, and generating corresponding transaction records. These records will be permanently stored in the blockchain ledger to ensure that data access behavior cannot be tampered with. In order to enhance the security of data transmission, a digital certificate is used to authenticate the external data request end. The certificate generation is completed by a trusted certificate authority (CA), and digital signature verification is required for each data interaction. After passing the identity authentication, the data request end can interact with the hospital's internal information system or image storage system through an encrypted channel to ensure the security and privacy of patient information, and finally generate a post-digestive endoscopy data collection security channel.

[0077] S12: Obtain the data characteristics of patients after digestive endoscopy, including the field storage accuracy and timestamp format of the patient's physiological data and the resolution and encoding method of the endoscopic image data, and finely adapt the API interface parameters corresponding to the external data request end in combination with the data structure corresponding to the hospital information system and the digestive endoscopy image storage system to generate a data request adaptation API interface parameter set, which includes data screening conditions and data format conversion rules;

[0078] In the embodiment of the present invention, the process of obtaining the characteristics of post-endoscopic patient data first requires in-depth analysis of the different dimensions of patient data. For physiological data, the accuracy of data storage is analyzed in detail. For example, the storage field accuracy of heart rate, blood pressure and other data should be set to two decimal places, and the timestamp format should comply with ISO 8601 standard (for example, "2025-01-06T15:30:00Z"); for image data, the resolution and encoding method of the data need to be obtained. The common endoscopic image resolution is 1920x1080, and the encoding method can be H.264 or HEVC. Then, by combining the data structures in the hospital information system (HIS) and the digestive endoscopy image storage system (such as PACS), the interface parameters between the two are analyzed. Through detailed review and adaptation of these parameters, the API interface parameter set required by the external data request end is determined. This interface parameter set includes data screening conditions (such as patient data within a specified date range) and data format conversion rules (such as converting H.264-encoded images to JPEG format). In addition, to ensure interface compatibility, necessary data field verification and fault tolerance mechanisms need to be introduced in the API interface to cope with different data formats and possible missing values, and finally generate a data request adaptation API interface parameter set.

[0079] S13: Based on the digestive endoscopy postoperative data collection security channel and the data request adaptation API interface parameter set, and using the authorized external data request end, send an acquisition request to the corresponding hospital information system and digestive endoscopy image storage system in the hospital internal network, and obtain the postoperative digestive endoscopy patient physiological data and postoperative endoscopy digestive tract status image data;

[0080] In an embodiment of the present invention, after constructing a secure channel for post-digestive endoscopy data collection and completing the adaptation of the API interface parameter set, the next step is to initiate a data acquisition request based on an authorized external data request end. In this process, the external request end first needs to be authenticated and authorized. After the verification is passed, the data acquisition request is sent through a secure API interface. The data request will be transmitted to the hospital's internal network through an encrypted blockchain channel. The specific request content includes requesting the patient's physiological data (such as body temperature, blood pressure, heart rate, etc.) and endoscopic image data (such as digestive tract status images). The data request end is connected to the hospital information system (HIS) and the digestive endoscopy image storage system (PACS) and initiates a request according to a pre-set API interface parameter set. During the request process, the data is encrypted and transmitted through the blockchain network to ensure that the data is not leaked or tampered with during the transmission process. After receiving the request, the hospital system queries the corresponding patient data according to the API request content and returns the corresponding data set. The physiological data and image data received by the request end are decrypted to finally obtain the patient's physiological data after digestive endoscopy and the post-endoscopic digestive tract status image data.

[0081] S14: Perform postoperative data preprocessing and standardization on the patient's physiological data after digestive endoscopy and the imaging data of the digestive tract status after endoscopy to obtain physiological standardized data after digestive endoscopy and standardized imaging data of the digestive tract after endoscopy.

[0082] In an embodiment of the present invention, the patient's physiological data and endoscopic image data are converted and cleaned in a unified format to meet the needs of subsequent deep learning algorithm analysis. For physiological data, the data must first be time-aligned to ensure that data from different sources have consistent timestamps and format conversion is performed according to standardized rules. For example, the heart rate and blood pressure data are unified to be recorded once per minute, the accuracy is adjusted to two decimal places, and the units of each data field must be unified (such as blood pressure units are unified as millimeters of mercury); for image data, the original endoscopic images must first be quality-screened to remove unqualified images due to image blur or unclear acquisition, and then size standardization is performed. All images are adjusted to a fixed resolution (such as 1920x1080). In order to improve the utilization efficiency of image data, images can be compressed using compression algorithms (such as JPEG2000) and converted into encoding formats to ensure compatibility. In addition, image color, contrast and other parameters also need to be optimized through image processing algorithms to unify the presentation of image data. The processed and standardized physiological data and image data will be stored as standardized data sets, ready for subsequent deep learning model training and analysis, to ensure the integrity and non-tamperability of the data processing process, and finally obtain physiological standardized data after digestive endoscopy and standard image data of the digestive tract after endoscopy.

[0083] Furthermore, the postoperative data preprocessing and standardization of the physiological data of the patient after digestive endoscopy and the postoperative digestive tract status imaging data after endoscopy includes:

[0084] The physiological data of patients after digestive endoscopy were abnormally removed and standardized to obtain physiological standardized data after digestive endoscopy;

[0085] In an embodiment of the present invention, the collected physiological data of patients after digestive endoscopy are cleaned and processed, and abnormal values ​​exist in the data, such as extreme values ​​or missing values ​​of physiological data such as blood pressure, heart rate, and body temperature. For this reason, an abnormal value detection technology based on statistical methods, such as Z-score or IQR (interquartile range) method, is used to eliminate abnormal values ​​in the data that exceed the normal range. For missing data, an interpolation method (such as linear interpolation, spline interpolation) is used to fill the missing part to ensure data integrity. Next, the data is processed using a standardization method to unify the data dimensions of different physiological parameters. Specifically, the Z-score standardization formula is used: Z = σ / (X-μ), where X is the original data, μ is the mean of the data, and σ is the standard deviation. The standardized data can eliminate the dimensional differences between different physiological indicators, ensure that each physiological data is within the same magnitude range, and finally obtain physiological standardized data after digestive endoscopy.

[0086] Preferably, the image data of the digestive tract status after endoscopy is enhanced, including rotation, translation and flipping, to obtain enhanced image data of the digestive tract after endoscopy;

[0087] In an embodiment of the present invention, data enhancement processing is performed on the digestive tract image data obtained after digestive endoscopy to increase the diversity of the training set and improve the robustness and accuracy of the subsequent model. The specific operations include firstly rotating the original image, and the rotation angle range is set to ±30 degrees to simulate different viewing angles in endoscopic operation; secondly, translating the image, that is, moving a certain pixel value in the horizontal direction or vertical direction to ensure that the model can learn information at different positions; thirdly, performing a mirror flip operation, that is, flipping the image horizontally or vertically to increase the diversity of the data. All enhancement operations are performed in the image preprocessing stage to ensure that the basic structural information of the original image will not be changed while increasing the amount of data. The enhanced image data provides more samples for subsequent analysis and training, avoids the occurrence of overfitting problems, and finally obtains enhanced image data of the digestive tract after endoscopy.

[0088] Preferably, grayscale histogram equalization is performed on the enhanced image data of the digestive tract after endoscopy to obtain contrast-balanced image data of the digestive tract after endoscopy;

[0089] In an embodiment of the present invention, in order to enhance the contrast and clarity of the post-endoscopic digestive tract image, a grayscale histogram equalization process is performed. The purpose of the grayscale histogram equalization is to improve the contrast of the image by adjusting the grayscale value distribution of the image, thereby making the details in the digestive tract clearer and facilitating subsequent analysis. In the specific implementation process, the grayscale histogram of the image is first calculated, that is, the frequency of occurrence of each grayscale level pixel in the image is counted, and based on the histogram, the cumulative distribution function (CDF) of each grayscale value is calculated. Through the mapping relationship of the CDF, each pixel value in the image is mapped to a new grayscale value, thereby adjusting the brightness and contrast of the image, making the grayscale value distribution of the image more uniform, and the details of the digestive tract are better displayed. The image data after the processing has a higher contrast in visual effect, and finally the contrast-balanced image data of the digestive tract after the endoscopy is obtained.

[0090] Preferably, pixel blurriness analysis is performed on the contrast-balanced image data of the post-endoscopic digestive tract to obtain pixel blurriness values ​​of the post-endoscopic digestive tract images; image denoising is performed on the corresponding contrast-balanced image data of the post-endoscopic digestive tract based on the pixel blurriness values ​​of the post-endoscopic digestive tract images to obtain denoised image data of the post-endoscopic digestive tract;

[0091] In an embodiment of the present invention, pixel blurriness analysis is performed on the contrast-balanced image of the digestive tract after endoscopic surgery to evaluate the clarity of the image, and further denoising is performed. The pixel blurriness analysis adopts a common image clarity evaluation method, such as the Laplace operator or the gradient operator (Sobel operator). Specifically, the second-order derivative of the image is first calculated by the Laplace operator to obtain the edge information of the image. If the edge information of the image is relatively blurred, it means that the clarity of the image is poor; if the edge information is clear, it means that the clarity of the image is high. The quality of the image is evaluated according to the size of the image blurriness value. For images with high blurriness values, denoising algorithms are used for processing, such as median filtering, Gaussian filtering, etc. These denoising algorithms can effectively remove random noise in the image, maintain the details of the image, and obtain denoised image data, so that the image of the digestive tract is clearer and more accurate, and finally denoised image data of the digestive tract after endoscopic surgery is obtained.

[0092] Preferably, the denoised image data of the digestive tract after endoscopy is standardized to obtain standard image data of the digestive tract after endoscopy.

[0093] In an embodiment of the present invention, the denoised post-endoscopic digestive tract image data is standardized to ensure the consistency and comparability of the data. The standardization method adopts a zero mean unit variance standardization method. The specific steps are: first, the mean and standard deviation of the denoised image pixel values ​​are calculated, and then each pixel value is standardized by the following formula: Q = (xa) / b, where x is the pixel value in the denoised image data, a is the image pixel mean, and b is the image pixel standard deviation. The standardized image data has a uniform scale in the pixel value range, which can eliminate the differences caused by the different brightness and contrast of the image itself. This step of standardization can not only ensure the consistency of the data, but also avoid the brightness differences in the image data from having an adverse effect on subsequent analysis and deep learning model training, and finally obtain standard post-endoscopic digestive tract image data.

[0094] Further, as an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 The functional flow diagram of the post-endoscopic feature analysis module is shown in FIG. 1 . In this embodiment, the post-endoscopic feature analysis module includes the following functions:

[0095] S21: Performing frequency domain conversion on physiological indexes of post-digestive endoscopy physiological standardized data to obtain the periodic rhythm spectrum of post-digestive endoscopy physiological indexes;

[0096] In an embodiment of the present invention, frequency domain analysis is performed on physiological standardized data after digestive endoscopy, where the standardized data includes physiological indicators such as the patient's heart rate, respiratory rate, and blood oxygen saturation. The data after equalization processing should eliminate noise and outliers to ensure the accuracy of the data. Then, the fast Fourier transform (FFT) algorithm is used to convert these time domain data to the frequency domain. FFT can efficiently convert signals from the time domain to the frequency domain to generate a periodic information spectrum. For each physiological indicator, based on its frequency response characteristics, a periodic rhythm spectrum is obtained by analyzing its frequency domain characteristics. This can reveal the periodic law of changes in physiological indicators over time, and ultimately obtain the periodic rhythm spectrum of physiological indicators after digestive endoscopy.

[0097] S22: A dynamic trend key node analysis is performed on the cyclical rhythm spectrum of physiological indicators after digestive endoscopy using a slope change detection and extreme point tracking method, so as to analyze the slope changes corresponding to each physiological indicator curve along the time axis. When the slope mutation exceeds the preset threshold, it is marked as a potential key node, and the maximum and minimum points corresponding to each physiological indicator curve are tracked at the same time to obtain a dynamic trend key node set of physiological indicators after endoscopy;

[0098] In an embodiment of the present invention, a key node analysis is performed on the dynamic trend of the cyclical rhythm spectrum of physiological indicators after digestive endoscopy. First, a slope change detection method is used to capture the mutation points of each physiological indicator. Specifically, the derivative of the spectrum data is calculated and its change trend is analyzed. If the slope suddenly changes and the amplitude of the change exceeds a preset threshold, it is considered to be a potential key node. In this way, the time point when the physiological indicator changes more drastically can be accurately captured. In addition, the extreme point tracking algorithm is used to locate the maximum and minimum points in the cyclical rhythm spectrum. The extreme points usually represent significant fluctuations or stable states of physiological indicators. Combined with these slope mutation points and extreme points, a set of key nodes can be fully identified. This step effectively identifies potential physiological abnormalities and ensures that the patient's postoperative physiological state can be monitored in time, and finally a set of key nodes for the dynamic trend of post-endoscopic physiological indicators is obtained.

[0099] S23: Determine the long-scale time window and the short-scale time window on the time axis within the periodic rhythm spectrum of the post-digestive endoscopy physiological indicators between every two adjacent key nodes in the key node set of the dynamic trend of the post-endoscopic physiological indicators, and divide the corresponding post-digestive endoscopy physiological standardized data into long- and short-window sequences based on the long-scale time window and the short-scale time window, so as to obtain the endoscopic surgery physiological long-time window sequence data and the endoscopic surgery physiological short-time window sequence data;

[0100] In an embodiment of the present invention, time window division is performed based on the dynamic trend key node set obtained in the previous step. For every two adjacent key nodes, the time interval between them is first calculated on the time axis of the periodic rhythm spectrum. The length of this time interval determines the defined long-scale time window and short-scale time window. The long-scale time window usually covers a longer time period and is mainly used to analyze the overall trend and chronic fluctuations of physiological indicators. The short-scale time window is used to capture short-term, rapidly changing physiological fluctuations. In specific implementation, the two-scale time periods are extracted by a sliding window method, and data is divided to ensure that long-term and short-term physiological dynamic changes can be comprehensively analyzed. This process provides effective input sequence data for subsequent deep learning analysis, and finally obtains endoscopic surgery physiological long-term window sequence data and endoscopic surgery physiological short-term window sequence data.

[0101] S24: Using the long short-term memory network, extract the long and short time series features of the endoscopic surgery physiological long-term window sequence data and the endoscopic surgery physiological short-term window sequence data to obtain the long and short time series features of the post-endoscopic physiological indicators, including the low-frequency trend change features corresponding to the long-term window and the high-frequency fluctuation features corresponding to the short-term window of each post-endoscopic physiological indicator;

[0102] In an embodiment of the present invention, a long short-term memory network (LSTM) is used to extract features from the previously divided long-term window sequence data and short-term window sequence data. LSTM is a recurrent neural network suitable for time series data, which can effectively learn the long-term dependence and short-term dependence in the time series data. The long-term window sequence data mainly reflects the long-term change trend of the patient's physiological indicators, while the short-term window sequence data contains drastic fluctuations in the short term. The LSTM network processes these two types of data separately and extracts the time series features therein. Under the long-term window, LSTM extracts the low-frequency trend change features of the physiological indicators, which are usually related to chronic changes and stable states; under the short-term window, LSTM extracts the high-frequency fluctuation features, which are usually related to acute changes and rapid fluctuations. In this way, the macro-trends and micro-fluctuations of the patient's postoperative physiological indicators can be captured at the same time, and finally the long-short time series features of the post-endoscopic physiological indicators are obtained, including the low-frequency trend change features corresponding to each post-endoscopic physiological indicator under the long-term window and the high-frequency fluctuation features corresponding to the short-term window.

[0103] S25: Convolutional neural network is used to identify the postoperative healing lesion characteristics of the standard imaging data of the digestive tract after endoscopic surgery, and the abnormal lesion characteristics after endoscopic surgery are obtained, including the bleeding points of the digestive tract healing after endoscopic surgery and the corresponding lesion degree of the ulcer.

[0104] In an embodiment of the present invention, a convolutional neural network (CNN) is used to identify healing lesion features of standard image data of the digestive tract after endoscopic surgery. The convolutional neural network has significant advantages in the field of image processing and can automatically learn and extract important features in the image. In this step, the digestive tract image data of patients after endoscopic surgery is first collected, and after preprocessing, it is input into the CNN model. Through multi-layer convolution operations, CNN can automatically extract lesion features in the image, such as healing bleeding points, ulcer areas and other potential abnormal changes. By training the CNN model, it can accurately distinguish between normal healing and lesions, and then identify the abnormal healing lesion features of the postoperative digestive tract. In the feature recognition process, CNN can also classify according to different lesion degrees, and further judge the progress and abnormality of healing. Through the output of the model, the healing lesion features of the digestive tract after endoscopic surgery can be obtained, and finally the abnormal healing lesion features after endoscopic surgery are obtained.

[0105] Furthermore, the simultaneous tracking of the maximum and minimum points corresponding to each physiological indicator curve includes:

[0106] Perform curve fluctuation frequency characteristic analysis on each physiological index curve in the periodic rhythm spectrum of post-digestive endoscopy physiological indexes to obtain the curve fluctuation frequency characteristics corresponding to each post-endoscopic physiological index;

[0107] In an embodiment of the present invention, time series data of various physiological indicators, such as heart rate, blood pressure, body temperature, respiratory rate, etc., are obtained from a physiological monitoring system after digestive endoscopy. The data needs to be preprocessed, including denoising, smoothing and other operations to ensure the accuracy of the original signal. Then, Fourier transform (Fast Fourier Transform, FFT) is used to perform frequency domain analysis on these time series data to calculate the spectrum of each physiological indicator curve. Spectral analysis can reveal the intensity of each frequency component in the signal, so that the main frequency component of the periodic fluctuation can be identified, that is, the fluctuation frequency characteristics of each physiological indicator. These frequency characteristics reflect the periodic change law of the physiological system. For example, the fluctuation frequency of the heart rate may correspond to the heart's beating rhythm, while the body temperature fluctuation is related to the physiological rhythm. Through spectrum analysis, the periodic components of each physiological indicator can be extracted, and finally the curve fluctuation frequency characteristics corresponding to each post-endoscopic physiological indicator are obtained.

[0108] Preferably, the curve fluctuation pattern regions of each physiological indicator curve in the periodic rhythm spectrum of the post-endoscopic physiological indicator are divided based on the curve fluctuation frequency characteristics corresponding to each post-endoscopic physiological indicator, so as to generate the curve fluctuation pattern sub-regions corresponding to each post-endoscopic physiological indicator;

[0109] In an embodiment of the present invention, the fluctuation frequency characteristics of each physiological indicator are divided into different fluctuation pattern areas according to the previously obtained spectrum data. Specifically, it is necessary to first set a standard for the fluctuation pattern division. For example, the spectrum data is classified by a statistical analysis method (such as cluster analysis or a threshold-based segmentation method), and the fluctuations in different frequency bands are divided into different pattern areas. For each physiological indicator's spectrum diagram, it can be divided into high-frequency, medium-frequency and low-frequency areas according to the distribution of the main frequency components. For example, for heart rate fluctuations, the low-frequency part can be corresponded to the basic heart rhythm, and the high-frequency part can be corresponded to the fluctuation reflecting the autonomic nervous activity. A similar division is performed on each physiological indicator to obtain the corresponding fluctuation pattern sub-area. This process can simplify and structure the complex fluctuation pattern of each physiological indicator, and finally divide and generate the curve fluctuation pattern sub-area corresponding to each post-endoscopic physiological indicator.

[0110] Preferably, the golden section search method is used to perform extreme value search processing on the curve fluctuation pattern sub-regions corresponding to each post-endoscopic physiological index, so as to obtain the curve fluctuation extreme value point search interval corresponding to each post-endoscopic physiological index;

[0111] In an embodiment of the present invention, the golden section search method is used to locate the extreme points for the fluctuation pattern sub-region corresponding to each physiological indicator previously obtained. The golden section search method is an efficient optimization method, which is mainly used to search for the extreme points of a function within a known interval. For the fluctuation pattern sub-region of each physiological indicator, the search interval range is first determined. This interval is generally determined by the fluctuation pattern region divided in the previous step. Then, the interval is continuously divided into two parts by the golden section method, and which part contains the extreme points is determined according to the fluctuation characteristics of the spectrum. The golden section method can accurately locate the maximum and minimum points of the curve by continuously narrowing the search range. Through this search process, the precise extreme point search interval within each physiological indicator fluctuation pattern sub-region can be obtained, and finally the curve fluctuation extreme point search interval corresponding to each post-endoscopic physiological indicator is obtained.

[0112] Preferably, the parabolic approximation method is used to perform curve local extreme point fitting calculation on the curve fluctuation extreme point search interval corresponding to each post-endoscopic physiological index, so as to obtain the maximum and minimum points corresponding to each physiological index curve.

[0113] In an embodiment of the present invention, a parabola approximation method is used to fit the local extreme points of the curve according to the extreme point search interval obtained previously. Specifically, the parabola approximation method is a numerical optimization method, which is mainly used to fit the local extreme points on the curve. According to this method, within the extreme point search interval of the fluctuation pattern sub-region of each physiological indicator, a number of points within the interval are first taken, and a quadratic parabola is fitted according to the numerical values ​​of these points. By analyzing the vertices of the parabola, the local maximum point or minimum point within the interval can be accurately obtained. Since the parabola approximation method can efficiently calculate the extreme value of the curve, the maximum point and minimum point of each physiological indicator curve can be quickly and accurately determined. These extreme points represent the key fluctuation characteristics of the physiological indicator curve within a certain period, and can reflect the trend of changes in the physiological state of the patient after surgery, and finally the maximum point and minimum point corresponding to each physiological indicator curve are obtained.

[0114] Furthermore, the use of a convolutional neural network to identify postoperative healing lesion features of standard post-endoscopic digestive tract imaging data includes:

[0115] Obtain the absorption and scattering characteristics of the digestive tract tissue under different wavelengths of light through the standard image data of the digestive tract after endoscopic surgery, and perform image optical parameter analysis on the corresponding standard image data of the digestive tract after endoscopic surgery based on the absorption and scattering characteristics of the digestive tract tissue under different wavelengths of light combined with a preset light scattering model to obtain the absorption coefficient and scattering coefficient at each pixel point in the image of the digestive tract after endoscopic surgery;

[0116] In an embodiment of the present invention, the absorption coefficient and scattering coefficient of the digestive tract tissue under different wavelengths of light are obtained through standard image data of the digestive tract after endoscopic surgery. To this end, it is necessary to collect post-endoscopic images and use specific optical parameter models to analyze these images. Specifically, the absorption and scattering characteristics of the digestive tract tissue are affected by tissue type, blood flow and other physiological characteristics, and based on the optical characteristics of these tissues, a set of preset light scattering models, such as the Mie scattering model or the Rayleigh scattering model, are selected to simulate the interaction between light and tissue. Then, by combining the spectral data with the endoscopic image, each pixel of the image is analyzed using computer vision technology to extract the corresponding absorption coefficient and scattering coefficient. The value of each pixel will be mapped to a specific optical parameter, so that the final image data contains the absorption and scattering characteristics of each pixel at different wavelengths, and finally the absorption coefficient and scattering coefficient at each pixel in the post-endoscopic digestive tract image are obtained.

[0117] Preferably, based on the absorption coefficient and scattering coefficient at each pixel point in the post-endoscopic digestive tract image, the corresponding post-endoscopic digestive tract standard image data is subjected to image texture fractal structure recognition to generate a post-endoscopic digestive tract image fractal feature structure, including a mucosal fold and a vascular branch fractal feature structure; and the corresponding post-endoscopic digestive tract image fractal dimension is obtained according to the post-endoscopic digestive tract image fractal feature structure;

[0118] In an embodiment of the present invention, the texture fractal structure of the endoscopic image data is identified by using the absorption coefficient and the scattering coefficient. First, the optical parameter data of the post-endoscopic image is input into the texture analysis algorithm to extract the fractal features. Fractal analysis methods such as the box dimension (Box-Counting Dimension) can be used here to identify the fractal feature structure in the image. Through these methods, the mucosal folds, vascular branches and other structures in the image can be identified, and their corresponding fractal dimensions can be calculated. This fractal dimension reflects the complexity and self-similarity of the tissue morphology, which is helpful for the subsequent quantitative analysis of the structural characteristics of the digestive tract. By analyzing the fractal dimension, the morphological changes of the mucosa in the post-operative digestive tract, the distribution of the vascular network and other information can be further obtained, thereby providing detailed tissue information, and finally obtaining the fractal dimension of the post-endoscopic digestive tract image.

[0119] Preferably, grayscale texture feature analysis is performed on standard post-endoscopic digestive tract image data using a grayscale co-occurrence matrix to obtain grayscale texture feature distribution of post-endoscopic digestive tract images between pixel pairs in different directions and distances;

[0120] In an embodiment of the present invention, grayscale texture feature analysis is performed on post-endoscopic image data by using a grayscale co-occurrence matrix (GLCM). The specific operation is to convert the standard post-endoscopic image data into a grayscale image, and then analyze the grayscale relationship between each pixel and its surrounding pixels. By calculating the grayscale co-occurrence matrix, statistical information on grayscale changes in the image can be obtained, covering texture characteristics such as directionality, contrast, and homogeneity. In this process, by selecting different directions (such as 0 degrees, 45 degrees, 90 degrees, and 135 degrees) and different distances (such as 1 pixel, 2 pixels, etc.), grayscale texture features in different directions and distances are extracted. The obtained texture features can be used to further analyze the spatial distribution characteristics of the image and provide useful clues to distinguish the different manifestations of healthy tissue and lesion areas. Through the extraction and analysis of these features, data support can be better provided for subsequent anatomical region segmentation and lesion detection, and finally the grayscale texture feature distribution of post-endoscopic digestive tract images between pixel pairs in different directions and distances is obtained.

[0121] Preferably, based on the fractal dimension of the post-endoscopic digestive tract image and the grayscale texture feature distribution of the post-endoscopic digestive tract image between pixel pairs in different directions and distances, the corresponding post-endoscopic digestive tract standard image data is segmented into image anatomical regions using a semantic segmentation network model to obtain images of the anatomical structure regions of each post-endoscopic digestive tract;

[0122] In an embodiment of the present invention, based on the fractal dimension and grayscale texture features obtained in the first two steps, a semantic segmentation network model is used to perform anatomical region segmentation on the endoscopic post-operative gastrointestinal image. In this process, the previously extracted absorption coefficient, scattering coefficient and grayscale texture features are first input into the deep learning semantic segmentation network. The network can realize accurate digestive tract anatomical region recognition and segmentation through a convolutional neural network (CNN) structure, combining the spatial features and texture features of the image. Specifically, the network will separate different anatomical regions, such as gastric wall, mucosa, blood vessels, etc., according to the texture and fractal features in the image, and assign a label to each region. The segmented result will obtain multiple anatomical structure region images, which provide clear boundary and regional information for subsequent lesion detection, healing status analysis, etc., and can effectively identify various parts of the post-operative digestive tract, and finally obtain various endoscopic post-operative digestive tract anatomical structure region images.

[0123] Preferably, a convolutional neural network is used to extract the features of postoperative healing lesions from the images of various post-endoscopic gastrointestinal anatomical structure regions, and the bleeding points of post-endoscopic gastrointestinal healing, the size of the ulcer surface of post-endoscopic gastrointestinal lesions, and the changes in the depth of post-endoscopic gastrointestinal lesions ulcers are obtained; the corresponding post-endoscopic gastrointestinal lesion ulcer expansion rate is obtained according to the changes in the depth of post-endoscopic gastrointestinal lesions ulcers, and the ulcer lesion degree calculation formula is used to perform lesion measurement calculation on the corresponding post-endoscopic gastrointestinal images based on the size of the ulcer surface of post-endoscopic gastrointestinal lesions and the ulcer expansion rate of post-endoscopic gastrointestinal lesions, and the lesion degree corresponding to the post-endoscopic gastrointestinal ulcer is obtained.

[0124] In an embodiment of the present invention, a convolutional neural network (CNN) is used to further analyze the previously segmented images of various anatomical structural regions of the digestive tract after endoscopic surgery to extract the characteristics of postoperative healing lesions. The CNN model extracts features from the images of each anatomical region, paying special attention to changes in the lesion area, such as the identification of healing bleeding points, the measurement of the size of the ulcer surface, and the calculation of the ulcer depth. Through model training, the healing process in the digestive tract, the lesion area, and the specific manifestations of the ulcer can be accurately identified. Next, based on the change in the depth of post-endoscopic gastrointestinal ulcers, the expansion rate of the lesion was calculated. For the calculation of the lesion expansion rate, a time series-based analysis method was used to associate the change in the ulcer area in the endoscopic image with the time axis, and then the lesion expansion rate was calculated. A formula for calculating the degree of ulcer lesions was constructed by combining the post-endoscopic gastrointestinal image area, the size of the ulcer surface after endoscopic gastrointestinal lesions, the attenuation factor of the ulcer area, the ulcer expansion rate after endoscopic gastrointestinal lesions, the ulcer expansion rate attenuation factor, the weighted coefficient of post-endoscopic gastrointestinal ulcers, the number of image features, the image feature lesion weighted coefficient and related parameters to obtain the final lesion measurement value. This lesion measurement value can quantify the severity of postoperative gastrointestinal lesions and finally obtain the lesion degree corresponding to the post-endoscopic gastrointestinal ulcer.

[0125] Furthermore, the calculation formula for the degree of ulcer lesions is specifically:

[0126]

[0127] Where D is the lesion degree corresponding to the post-endoscopic digestive ulcer, Ω is the post-endoscopic digestive tract imaging area, and A xy is the size of the ulcer surface of the post-endoscopic digestive tract lesion at the position (x, y), λ1 is the attenuation factor of the ulcer area, R xy is the expansion rate of post-endoscopic digestive tract ulcer at position (x, y), λ2 is the attenuation factor of ulcer expansion rate, α1 is the weighting coefficient of post-endoscopic digestive tract ulcer, n is the number of image features corresponding to post-endoscopic digestive tract images, i is the item index corresponding to the image feature, and fi is the i-th image feature corresponding to the post-endoscopic digestive tract image, α2 is the image feature lesion weighting coefficient, and η is the correction coefficient of the lesion degree.

[0128] The present invention obtains a calculation formula for the degree of ulcer lesions by using a specific mathematical model and verifying it, which is used to perform lesion measurement calculation on the corresponding post-endoscopic digestive tract images. The calculation formula for the degree of ulcer lesions involves multiple influencing factors, such as the area of ​​the ulcer, the expansion rate, image characteristics, etc., and can comprehensively consider the influence of different lesions, so as to obtain a more accurate degree of lesions. The specific influencing factors include: the size of the lesion ulcer surface, which directly reflects the physical area of ​​the ulcer. Generally, the larger the area of ​​the ulcer, the more serious the lesion; the ulcer expansion rate, which measures the growth rate of the ulcer, and ulcers with faster growth rates usually represent more serious lesions; the texture features of the image, which reflect the local structure or changes of the image, can reflect the heterogeneity or irregularity of the tissue, and help to identify the morphological changes of the ulcer. Through the comprehensive analysis of the above variables, various aspects of digestive tract ulcers can be better captured, and the expansion and change process of the ulcer are particularly important. The formula uses weighting coefficients and attenuation factors, which can be adjusted according to the importance of different image regions or features. The weighting coefficient can adjust the importance of different features in the calculation of the total lesion degree. If the area of ​​the lesion is more important, the corresponding weighting coefficient can be increased to increase the influence of the area on the lesion degree. Similarly, if the texture features of the image are more important, the corresponding weighting coefficient can be adjusted. The attenuation factor takes into account the nonlinear effect of the ulcer area or expansion rate on the lesion degree. As the ulcer area or expansion rate increases, the introduction of the attenuation factor can effectively simulate the complex behavior of the ulcer. For example, the increase in area does not necessarily linearly affect the lesion degree. A larger area brings more complex pathological changes. These parameters can make the formula more flexible and adaptable to different conditions and image data. Through integral operations, the formula takes into account the information of spatial position, that is, the size and expansion rate of the lesion ulcer are not just the characteristics of a single point, but a regional and global information integration, which enables the model to consider the overall changes in the region rather than local characteristics, thereby better simulating the real clinical lesion situation. By incorporating the texture features of the image (such as the features extracted by the gray-level co-occurrence matrix) into the calculation of the lesion extent, it is possible to establish a connection between subtle changes in post-endoscopic images and actual tissue pathological changes. This method helps to discover some early lesions or tiny lesions and provide more accurate pathological information. Since the formula includes the lesion expansion rate, the formula can not only evaluate the current lesion extent of the ulcer, but can also be used to dynamically track the progression of the lesion, especially during multiple observations after endoscopy, to evaluate in real time whether the lesion is improving or worsening. In addition, the correction coefficient in the formula provides further flexibility and can be used to correct the deviation of the model under different cases or imaging data to ensure the accuracy of lesion measurement. In summary, the formula fully considers the lesion extent D corresponding to post-endoscopic gastrointestinal ulcers, the post-endoscopic gastrointestinal imaging area Ω, and the size of the post-endoscopic gastrointestinal lesion ulcer surface A at position (x, y). xy, ulcer area attenuation factor λ1, post-endoscopic digestive tract ulcer expansion rate R at position (x, y) xy , ulcer expansion rate attenuation factor λ2, weighting coefficient of post-endoscopic digestive tract ulcer α1, number of image features corresponding to post-endoscopic digestive tract images n, item index i corresponding to image features, i-th image feature f corresponding to post-endoscopic digestive tract images i , image feature lesion weighting coefficient α2, lesion degree correction coefficient η, according to the lesion degree D corresponding to post-endoscopic gastrointestinal ulcer and the mutual correlation between the above parameters constitute a functional relationship The formula can realize the lesion measurement calculation process of the corresponding post-endoscopic digestive tract images. At the same time, by introducing the correction coefficient η of the lesion degree, it can be adjusted according to the errors occurring in the calculation process, thereby improving the accuracy and applicability of the ulcer lesion degree calculation formula.

[0129] Furthermore, the post-endoscopic recovery risk assessment module includes the following functions:

[0130] The corresponding characteristic means and characteristic variances were obtained through the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal healing lesions after endoscopic surgery, and the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal healing lesions after endoscopic surgery were Z-score standardized based on the characteristic means and characteristic variances to obtain the standard characteristics of the long and short time series characteristics of post-endoscopic physiological indicators and the standard characteristics of abnormal healing lesions after endoscopic surgery;

[0131] In an embodiment of the present invention, data preprocessing is performed on the long and short time series characteristics of physiological indicators and the characteristics of abnormal healing lesions of patients after endoscopic surgery. The long and short time series characteristics of physiological indicators refer to the data of heart rate, respiratory rate, body temperature, blood pressure, etc. of patients after surgery that change over time. The characteristics of abnormal healing lesions include imaging data of healing progress, bleeding, scar tissue, etc. of the gastrointestinal tract of patients after endoscopic surgery. For these characteristics, the mean and variance of each feature are first calculated. These two statistics are used to reflect the central tendency and discreteness of the data. Then, for each feature, the Z-score standardization method is used to standardize the data. Specifically, for each data point, the Z-score standardization formula is standardized characteristic value = (original data point - characteristic mean) / characteristic standard deviation, and finally the standard characteristics of the length of physiological indicators after endoscopic surgery and the standard characteristics of abnormal lesions after endoscopic surgery are obtained.

[0132] Preferably, mutual information analysis is performed between each dimensional feature within the standard feature of the length of physiological indicators after endoscopy and each characteristic component within the standard feature of abnormal lesions after endoscopy, so as to obtain the mutual information between each physiological dimensional feature after endoscopy and each abnormal lesion characteristic component; nonlinear correlation evaluation analysis is performed between each dimensional feature within the standard feature of the length of physiological indicators after endoscopy and each characteristic component within the standard feature of abnormal lesions after endoscopy according to the mutual information between each physiological dimensional feature after endoscopy and each abnormal lesion characteristic component, so as to obtain the nonlinear correlation between each physiological dimensional feature after endoscopy and each abnormal lesion characteristic component;

[0133] In an embodiment of the present invention, mutual information analysis is performed on the standard features of the length of physiological indicators and the standard features of healing abnormal lesions that have been standardized before. Mutual information is a measure of the correlation between two variables, and can evaluate the mutual dependence between different dimensional features and lesion features. In the specific operation process, a suitable algorithm (such as an entropy-based mutual information calculation method) is first selected to calculate the mutual information value between each pair of physiological features and abnormal lesion features. The calculation process can use the sklearn.metrics.mutual_info_score function in Python, or use a custom mutual information calculation formula. After calculating the mutual information, the mutual information value between each pair of physiological indicators and abnormal lesion features is obtained. These values ​​reflect the dependency between different features. Based on these mutual information values, nonlinear correlation analysis is further performed to evaluate the complex nonlinear dependency between physiological features and lesion features, and finally the nonlinear correlation between each physiological dimensional feature after endoscopic surgery and each abnormal lesion feature component is obtained.

[0134] Preferably, a linear correlation evaluation analysis is performed between each dimension feature within the standard feature of the length of the physiological index after endoscopy and each characteristic component within the standard feature of the abnormal lesion after endoscopy, to obtain the linear correlation coefficient between each physiological dimension feature after endoscopy and each abnormal lesion characteristic component;

[0135] In the embodiment of the present invention, the linear correlation between the physiological indexes after endoscopy and the characteristics of abnormal healing lesions is further analyzed and calculated by using the Pearson Correlation Coefficient, which is widely used to evaluate the linear relationship between two variables. In the specific implementation, for each pair of physiological dimension characteristics and abnormal lesion characteristic components obtained previously, their Pearson correlation coefficients are calculated, and the calculation formula is: Where X u and Y u are the values ​​of physiological characteristics and abnormal pathological characteristics, respectively. and is their mean, and the obtained linear correlation coefficient reflects the strength of the linear relationship between physiological characteristics and lesion characteristics. According to the size of the correlation coefficient, the influence of different characteristics on postoperative recovery can be further evaluated, and finally the linear correlation coefficient between each physiological dimension characteristic after endoscopic surgery and each abnormal lesion characteristic component can be obtained.

[0136] Preferably, the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal healing lesions after endoscopy are input into a preset deep neural network model, and the nonlinear correlation and linear correlation coefficient between the post-endoscopic physiological dimension characteristics and the characteristic components of each abnormal lesion are combined to perform feature fusion, so as to generate a post-digestive endoscopy fusion feature set;

[0137] In an embodiment of the present invention, a deep learning model is used to fuse the long and short time series features of physiological indicators and the features of abnormal healing lesions. First, a deep neural network architecture suitable for this task (such as a multi-layer perceptron, a convolutional neural network, etc.) is selected, and the previously obtained standardized features are input into the model. The design of the deep neural network takes into account the nonlinear correlation between features. The model can automatically capture the complex relationship between different features through multiple layers of nonlinear transformations. When inputting data, the input features are weightedly fused in combination with the previously obtained mutual information value and the calculated linear correlation coefficient. The goal of this weighted fusion is to assign different weights to different features based on the correlation between the features, so that the model can more effectively learn the feature information that has an important impact on the prediction of postoperative recovery. In this way, the generated fused feature set can comprehensively reflect the potential risk factors for the patient's postoperative recovery, and finally generate a postoperative fusion feature set for digestive endoscopy.

[0138] Preferably, a postoperative recovery risk assessment calculation formula is used to perform a postoperative recovery risk assessment calculation on the digestive endoscopy postoperative fusion feature set to obtain the degree of postoperative recovery risk of the digestive endoscopy.

[0139] In an embodiment of the present invention, a suitable postoperative recovery risk assessment calculation formula is constructed by combining the postoperative observation time range, time variable parameters, postoperative physiological indicator time series characteristics, abnormal healing lesion time series characteristics, abnormal healing lesion recovery risk weight coefficient, postoperative physiological indicator long-short time series characteristics, postoperative physiological indicator recovery risk weight coefficient, digestive endoscopy postoperative recovery risk influencing factors, specific values ​​corresponding to individual characteristics of postoperative digestive endoscopy patients, weight coefficients and related parameters to perform postoperative recovery risk assessment calculation on the digestive endoscopy postoperative fusion feature set to quantitatively predict the patient's postoperative recovery risk level, and finally obtain the digestive endoscopy postoperative recovery risk level.

[0140] Furthermore, the postoperative recovery risk assessment calculation formula is specifically:

[0141]

[0142] Where R is the risk level of postoperative recovery after digestive endoscopy, T is the postoperative observation time range, t is the time variable parameter, P(t) is the time series characteristics of the postoperative physiological index corresponding to the patient at time point t after surgery, Q(t) is the time series characteristics of the abnormal healing lesions corresponding to the patient at time point t after surgery, β is the risk weight coefficient of recovery of abnormal healing lesions, H(t) is the long and short time series characteristics of the postoperative physiological index corresponding to the patient at time point t after surgery, γ is the risk weight coefficient of recovery of postoperative physiological indexes, ε is the risk influencing factor of postoperative recovery after digestive endoscopy, m is the total number of individual characteristics of patients after digestive endoscopy, X j is the specific value corresponding to the individual characteristics of the jth patient after digestive endoscopy, φ j is the weight coefficient corresponding to the individual characteristics of the jth patient after digestive endoscopy, and ξ is the correction coefficient of the risk of recovery after digestive endoscopy.

[0143] The present invention obtains a postoperative recovery risk assessment calculation formula by using a specific mathematical model and after verification, which is used to perform postoperative recovery risk assessment calculation on the postoperative fusion feature set of digestive endoscopy. P(t) and Q(t) in the postoperative recovery risk assessment calculation formula represent the characteristics of the patient's postoperative physiological indicators and abnormal healing lesions in the time series. Using these time series data, the dynamic changes in the postoperative recovery process can be captured. By performing time series analysis on the characteristics at different time points after surgery, the patient's recovery trend can be dynamically evaluated, and the postoperative risk changes can be identified in real time to ensure early intervention. Individual characteristics and corresponding weight coefficients, which can be the patient's clinical background, health status, etc., work together with other physiological and lesion characteristics on postoperative recovery. By weighted integration of individual characteristics, the formula can provide a quantitative personalized risk assessment, which can more accurately tailor recovery expectations for each patient and ensure that individual differences are fully considered. The β in the formula is the recovery risk weight coefficient of the abnormal healing lesion characteristics, which indicates the degree of influence of the risk of abnormal healing lesions on postoperative recovery. Different lesion types or recovery processes have different weight values. This weighting mechanism allows abnormal lesions with different behaviors in the healing process (such as infection, bleeding, scars, etc.) to receive more accurate attention, ensuring effective monitoring and adjustment of high-risk factors in postoperative recovery. In addition, the correction coefficient in the formula performs the final calibration of the recovery risk level, which is an adjustment for certain special circumstances or external influencing factors (such as patient comorbidities, emergencies, etc.). The introduction of the correction coefficient can flexibly adjust the risk assessment results according to external factors or unexpected changes, making the assessment more accurate and operational. In summary, this formula fully considers the risk level R of postoperative recovery after digestive endoscopy, the postoperative observation time range T, the time variable parameter t, the time series characteristics P(t) of the postoperative physiological indicators corresponding to the time point t after surgery, the time series characteristics Q(t) of the abnormal healing lesions corresponding to the time point t after surgery, the risk weight coefficient β of recovery of abnormal healing lesions, the long and short time series characteristics H(t) of the postoperative physiological indicators corresponding to the time point t after surgery, the risk weight coefficient γ of postoperative physiological indicators recovery, the risk influencing factor ε of postoperative recovery after digestive endoscopy, the total number m corresponding to the individual characteristics of patients after digestive endoscopy, and the specific value X corresponding to the individual characteristics of the jth postoperative patient after digestive endoscopy j , the weight coefficient φ corresponding to the individual characteristics of the jth post-endoscopic patient j , the correction coefficient ξ of the risk of recovery after digestive endoscopy, forms a functional relationship based on the correlation between the risk of recovery after digestive endoscopy R and the above parameters:

[0144]

[0145] The formula can realize the calculation process of postoperative recovery risk assessment for the postoperative fusion feature set of digestive endoscopy. At the same time, by introducing the correction coefficient ξ of the postoperative recovery risk degree of digestive endoscopy, it can be adjusted according to the errors occurring in the calculation process, thereby improving the accuracy and applicability of the calculation formula for postoperative recovery risk assessment.

[0146] Furthermore, the postoperative recovery risk intelligent nursing module includes the following functions:

[0147] Obtain clinical experience after digestive endoscopy and determine the risk grading system for post-endoscopic recovery based on clinical experience after digestive endoscopy;

[0148] In an embodiment of the present invention, by analyzing the clinical data of a large number of patients after digestive endoscopy, multi-dimensional data such as surgery type, basic patient information (such as age, gender, medical history), and postoperative recovery are collected. These data can be sourced from the hospital's electronic health record system (EHR) or clinical database. After data cleaning and preprocessing, key factors related to postoperative recovery are extracted, such as bleeding, infection, pain control, gastrointestinal function recovery, and other indicators. These data are trained using machine learning algorithms, especially deep learning models (such as convolutional neural networks CNN or recurrent neural networks RNN). The model can identify the main factors affecting postoperative recovery and construct a recovery risk prediction model based on these factors. In this process, the model will model the relationship between each variable and, based on the performance in historical cases, provide different recovery risk classification standards. The generated recovery risk grading system can subdivide the patient's recovery process into multiple stages and grade the patient according to the recovery status, for example, dividing them into three levels: low risk, medium risk, and high risk, and finally determine the endoscopic postoperative recovery risk grading system.

[0149] Preferably, the recovery risk level corresponding to the patient after digestive endoscopy is determined based on the post-endoscopic recovery risk grading system according to the degree of post-endoscopic recovery risk, including a low recovery risk level, a medium recovery risk level and a high recovery risk level;

[0150] In an embodiment of the present invention, the actual situation of each post-endoscopic patient is compared with the standards in the system based on the previously established recovery risk grading system. The specific operation is to use a trained deep learning model or other machine learning model, input the patient's basic post-operative information and clinical data, and automatically evaluate the patient's recovery risk based on the features extracted by the model and the established risk level classification standards. For example, the model can evaluate whether the patient has a low, medium or high recovery risk level by analyzing factors such as whether there are complications after surgery, post-operative body temperature changes, pain scores, bowel movements, etc. If the patient recovers well after surgery without obvious complications, and functions such as defecation and appetite are gradually restored, that is, the risk level is 0%-45%, it is assessed as a low risk level; if the patient has mild complications, recovers slowly or requires auxiliary drug control, that is, the risk level is 46%-85%, it is assessed as a medium risk level; if the patient has serious complications such as severe bleeding, infection or functional impairment after surgery, that is, the risk level is 86%-100%, it is assessed as a high risk level. This step automatically completes the determination of the recovery risk level through an algorithm, and finally obtains the corresponding recovery risk level for patients after gastrointestinal endoscopy.

[0151] Preferably, a postoperative intelligent nursing analysis is performed based on the corresponding recovery risk level of the patient after gastrointestinal endoscopy, and recovery nursing suggestions for the patient after gastrointestinal endoscopy are generated to execute the corresponding risk recovery nursing tasks after gastrointestinal endoscopy.

[0152] In an embodiment of the present invention, an intelligent analysis is performed on the postoperative care of each patient based on the previously determined recovery risk level. The specific method is to generate personalized nursing suggestions for the patient according to the patient's recovery risk level (low, medium, high) using the established nursing intervention model. For example, for low-risk patients, nursing suggestions may include routine monitoring, appropriate activity guidance, and dietary suggestions; for medium-risk patients, nursing suggestions involve more detailed monitoring, drug treatment, and timely intervention measures; for high-risk patients, nursing suggestions need to include close observation of vital signs, early detection of complications, preventive measures, timely postoperative review, etc., and will be based on the patient's recovery risk level (low, medium, high). Specific nursing plans are generated based on specific situations (such as postoperative bleeding risk, infection risk, gastrointestinal function recovery, etc.) combined with clinical experience rules. This process uses the reasoning ability of deep learning models and the clinical experience rules of medical experts to provide customized and precise nursing plans in a data-driven manner. This intelligent nursing system can also make real-time adjustments based on dynamic changes in the patient's recovery process. For example, if the patient's recovery condition changes, it can provide real-time feedback and adjust the nursing plan to ensure that the nursing measures always match the patient's recovery condition, and ultimately generate nursing recommendations for patient recovery after gastrointestinal endoscopy to perform corresponding risk recovery nursing tasks after gastrointestinal endoscopy.

[0153] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A deep learning-based intelligent nursing system for postoperative digestive endoscopy, characterized in that: Includes the following modules: The post-endoscopic data acquisition and processing module is used to obtain the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy through the API interface, and to perform post-endoscopic data preprocessing and standardization on the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy, so as to obtain the physiological standardized data after digestive endoscopy and the standardized imaging data of the digestive tract after endoscopy; The post-endoscopic feature analysis module is used to extract long-short time series features from the standardized physiological data after digestive endoscopy using the long-short time memory network to obtain the long-short time series features of post-endoscopic physiological indicators; the convolutional neural network is used to identify the post-endoscopic healing lesion features of the standard imaging data of the digestive tract after endoscopy to obtain the abnormal healing lesion features after endoscopy, including the degree of lesions corresponding to post-endoscopic digestive tract healing bleeding, ulcers and infections; The post-endoscopic recovery risk assessment module is used to input the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal post-endoscopic healing lesions into the preset deep neural network model for feature fusion to generate a post-digestive endoscopy fusion feature set; perform post-digestive endoscopy recovery risk assessment calculation on the post-digestive endoscopy fusion feature set to obtain the degree of post-digestive endoscopy recovery risk; The postoperative recovery risk intelligent nursing module is used to determine the corresponding recovery risk level of patients after digestive endoscopy according to the degree of recovery risk after digestive endoscopy, and to perform postoperative intelligent nursing analysis based on the corresponding recovery risk level of patients after digestive endoscopy, and to generate recovery nursing recommendations for patients after digestive endoscopy, so as to perform corresponding postoperative risk recovery nursing tasks after digestive endoscopy.

2. The deep learning-based postoperative intelligent nursing system for digestive endoscopy according to claim 1 is characterized in that: The post-endoscopic data acquisition and processing module includes the following functions: A data collection security channel is built between the hospital's internal network and the external data requester using blockchain-based encryption technology and digital certificate authentication, so that the corresponding post-digestive endoscopy patient data interaction log is recorded through the blockchain's distributed ledger, and the external data requester is authenticated using a digital certificate to generate a post-digestive endoscopy data collection security channel. Obtain the data characteristics of patients after digestive endoscopy, including the field storage accuracy and timestamp format of the patient's physiological data, and the resolution and encoding method of the endoscopic image data. Combined with the data structure of the hospital information system and the digestive endoscopy image storage system, finely adapt the API interface parameters corresponding to the external data request end to generate a data request adaptation API interface parameter set, which includes data screening conditions and data format conversion rules. Based on the digestive endoscopy post-operative data collection security channel and the data request adaptation API interface parameter set and using the authorized external data request end to send an acquisition request to the corresponding hospital information system and digestive endoscopy image storage system in the hospital internal network, and obtain the patient's physiological data after digestive endoscopy and the digestive tract status image data after endoscopy; Postoperative data preprocessing and standardization were performed on the physiological data of patients after digestive endoscopy and the imaging data of the digestive tract status after endoscopy to obtain physiological standardized data after digestive endoscopy and standard imaging data of the digestive tract after endoscopy.

3. The deep learning-based postoperative intelligent nursing system for digestive endoscopy according to claim 2 is characterized in that: The postoperative data preprocessing and standardization of the physiological data of patients after digestive endoscopy and the postoperative digestive tract status imaging data after endoscopy includes: The physiological data of patients after digestive endoscopy were abnormally removed and standardized to obtain physiological standardized data after digestive endoscopy; Performing image data enhancement on the digestive tract status image data after endoscopy, including rotation, translation and flipping, to obtain enhanced image data of the digestive tract after endoscopy; Perform grayscale histogram equalization on the enhanced image data of the digestive tract after endoscopy to obtain contrast-balanced image data of the digestive tract after endoscopy; Perform pixel blurriness analysis on the contrast-balanced image data of the post-endoscopic digestive tract to obtain the pixel blurriness value of the post-endoscopic digestive tract image; perform image denoising on the corresponding contrast-balanced image data of the post-endoscopic digestive tract based on the pixel blurriness value of the post-endoscopic digestive tract image to obtain the denoised image data of the post-endoscopic digestive tract; The denoised image data of the digestive tract after endoscopy were standardized to obtain the standard image data of the digestive tract after endoscopy.

4. The deep learning-based postoperative intelligent nursing system for digestive endoscopy according to claim 1 is characterized in that: The post-endoscopic feature analysis module includes the following functions: Perform frequency domain conversion on physiological indexes of normalized data after digestive endoscopy to obtain the periodic rhythm spectrum of physiological indexes after digestive endoscopy. The dynamic trend key node analysis of the cyclical rhythm spectrum of physiological indicators after digestive endoscopy is performed based on the slope change detection and extreme point tracking method, so as to analyze the slope changes corresponding to each physiological indicator curve along the time axis. When the slope mutation exceeds the preset threshold, it is marked as a potential key node, and the maximum and minimum points corresponding to each physiological indicator curve are tracked at the same time to obtain the dynamic trend key node set of physiological indicators after endoscopy. According to the key node set of dynamic trends of post-endoscopic physiological indicators, long-scale time windows and short-scale time windows are determined on the time axis within the periodic rhythm spectrum of post-digestive endoscopy physiological indicators, and the corresponding post-digestive endoscopy physiological standardized data are divided into long- and short-window sequences based on the long-scale time window and the short-scale time window to obtain endoscopic surgery physiological long-time window sequence data and endoscopic surgery physiological short-time window sequence data; Long short-term memory network is used to extract long and short time series features from the physiological long-term window sequence data and the physiological short-term window sequence data of endoscopic surgery, so as to obtain the long and short time series features of physiological indicators after endoscopy, including the low-frequency trend change features corresponding to each physiological indicator after endoscopy in the long-term window and the high-frequency fluctuation features corresponding to the short-term window; Convolutional neural network was used to identify the postoperative healing lesion characteristics of standard post-endoscopic digestive tract imaging data, and the abnormal post-endoscopic healing lesion characteristics were obtained, including the post-endoscopic digestive tract healing bleeding points and the corresponding lesion degree of ulcer.

5. The deep learning-based postoperative intelligent nursing system for digestive endoscopy according to claim 4 is characterized in that: The simultaneous tracking of the maximum and minimum points corresponding to each physiological indicator curve includes: Perform curve fluctuation frequency characteristic analysis on each physiological index curve in the periodic rhythm spectrum of post-digestive endoscopy physiological indexes to obtain the curve fluctuation frequency characteristics corresponding to each post-endoscopic physiological index; Based on the curve fluctuation frequency characteristics corresponding to each post-endoscopic physiological index, each physiological index curve in the periodic rhythm spectrum of the post-endoscopic physiological index is divided into curve fluctuation pattern regions to generate curve fluctuation pattern sub-regions corresponding to each post-endoscopic physiological index; The golden section search method is used to perform extreme value search processing on the curve fluctuation mode sub-regions corresponding to each post-endoscopic physiological index, and the curve fluctuation extreme value point search interval corresponding to each post-endoscopic physiological index is obtained; The parabolic approximation method was used to perform curve local extreme point fitting calculation on the curve fluctuation extreme point search interval corresponding to each post-endoscopic physiological index, so as to obtain the maximum and minimum points corresponding to each physiological index curve.

6. The deep learning-based postoperative intelligent nursing system for digestive endoscopy according to claim 4 is characterized in that: The method of using a convolutional neural network to identify postoperative healing lesion features of standard postoperative digestive tract imaging data after endoscopic surgery includes: Obtain the absorption and scattering characteristics of the digestive tract tissue under different wavelengths of light through the standard image data of the digestive tract after endoscopic surgery, and perform image optical parameter analysis on the corresponding standard image data of the digestive tract after endoscopic surgery based on the absorption and scattering characteristics of the digestive tract tissue under different wavelengths of light combined with a preset light scattering model to obtain the absorption coefficient and scattering coefficient at each pixel point in the image of the digestive tract after endoscopic surgery; Based on the absorption coefficient and scattering coefficient at each pixel point in the post-endoscopic digestive tract image, the corresponding post-endoscopic digestive tract standard image data is subjected to image texture fractal structure recognition to generate the post-endoscopic digestive tract image fractal feature structure, including the mucosal folds and vascular branch fractal feature structure; the corresponding post-endoscopic digestive tract image fractal dimension is obtained according to the post-endoscopic digestive tract image fractal feature structure; By using the gray-level co-occurrence matrix to analyze the gray-scale texture characteristics of the standard image data of the post-endoscopic digestive tract, the distribution of gray-scale texture characteristics between pixel pairs in different directions and distances of the post-endoscopic digestive tract images was obtained. Based on the fractal dimension of post-endoscopic digestive tract images and the grayscale texture feature distribution between pixel pairs in different directions and distances, the semantic segmentation network model is used to segment the corresponding post-endoscopic digestive tract standard image data into image anatomical regions, and obtain the images of the anatomical structure regions of each post-endoscopic digestive tract. Convolutional neural networks were used to extract the features of postoperative healing lesions from the images of various post-endoscopic gastrointestinal anatomical structure regions, and the bleeding points of post-endoscopic gastrointestinal healing, the size of the ulcer surface of post-endoscopic gastrointestinal lesions, and the changes in the depth of post-endoscopic gastrointestinal lesions were obtained. The corresponding post-endoscopic gastrointestinal lesion ulcer expansion rate was obtained according to the changes in the depth of post-endoscopic gastrointestinal lesions. Based on the size of the ulcer surface and the expansion rate of post-endoscopic gastrointestinal lesions, the ulcer lesion degree calculation formula was used to perform lesion measurement calculation on the corresponding post-endoscopic gastrointestinal images, and the corresponding lesion degree of the post-endoscopic gastrointestinal ulcer was obtained.

7. The deep learning-based intelligent nursing system for postoperative digestive endoscopy according to claim 6 is characterized in that: The calculation formula for the degree of ulcer lesions is specifically: Where D is the lesion degree corresponding to the post-endoscopic digestive ulcer, Ω is the post-endoscopic digestive tract imaging area, and A xy is the size of the ulcer surface of the post-endoscopic digestive tract lesion at position (x, y), λ1 is the attenuation factor of the ulcer area, and R xy is the expansion rate of post-endoscopic digestive tract ulcer at position (x, y), λ2 is the attenuation factor of ulcer expansion rate, α1 is the weighting coefficient of post-endoscopic digestive tract ulcer, n is the number of image features corresponding to post-endoscopic digestive tract images, i is the item index corresponding to the image feature, and f i is the i-th image feature corresponding to the post-endoscopic digestive tract image, α2 is the image feature lesion weighting coefficient, and η is the correction coefficient of the lesion degree.

8. The deep learning-based postoperative intelligent nursing system for digestive endoscopy according to claim 1 is characterized in that: The post-endoscopic recovery risk assessment module includes the following functions: The corresponding characteristic means and characteristic variances were obtained through the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal healing lesions after endoscopic surgery, and the long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal healing lesions after endoscopic surgery were Z-score standardized based on the characteristic means and characteristic variances to obtain the standard characteristics of the long and short time series characteristics of post-endoscopic physiological indicators and the standard characteristics of abnormal healing lesions after endoscopic surgery; Mutual information analysis is performed between each dimensional feature within the standard features of the length of physiological indicators after endoscopy and each characteristic component within the standard features of abnormal lesions after endoscopy, so as to obtain the mutual information between each physiological dimensional feature after endoscopy and each characteristic component of abnormal lesions; nonlinear correlation evaluation analysis is performed between each dimensional feature within the standard features of the length of physiological indicators after endoscopy and each characteristic component within the standard features of abnormal lesions after endoscopy according to the mutual information between each physiological dimensional feature after endoscopy and each characteristic component of abnormal lesions, so as to obtain the nonlinear correlation between each physiological dimensional feature after endoscopy and each characteristic component of abnormal lesions; The linear correlation evaluation and analysis was performed on the various dimensional characteristics within the standard characteristics of the length of post-endoscopic physiological indicators and the various characteristic components within the standard characteristics of post-endoscopic abnormal lesions, and the linear correlation coefficients between the post-endoscopic physiological dimensional characteristics and the characteristic components of the abnormal lesions were obtained. The long and short time series characteristics of post-endoscopic physiological indicators and the characteristics of abnormal healing lesions after endoscopy are input into the preset deep neural network model, and the nonlinear correlation and linear correlation coefficient between the post-endoscopic physiological dimension characteristics and the characteristic components of each abnormal lesion are combined to perform feature fusion to generate a post-digestive endoscopy fusion feature set; The postoperative recovery risk assessment calculation formula was used to evaluate the postoperative recovery risk of digestive endoscopy postoperative fusion feature set to obtain the degree of postoperative recovery risk of digestive endoscopy.

9. The deep learning-based postoperative intelligent nursing system for digestive endoscopy according to claim 8 is characterized in that: The postoperative recovery risk assessment calculation formula is specifically: Where R is the risk level of postoperative recovery after digestive endoscopy, T is the postoperative observation time range, t is the time variable parameter, P(t) is the time series characteristics of the postoperative physiological index corresponding to the patient at time point t after surgery, Q(t) is the time series characteristics of the abnormal healing lesions corresponding to the patient at time point t after surgery, β is the risk weight coefficient of recovery of abnormal healing lesions, H(t) is the long and short time series characteristics of the postoperative physiological index corresponding to the patient at time point t after surgery, γ is the risk weight coefficient of recovery of postoperative physiological indexes, ε is the risk influencing factor of postoperative recovery after digestive endoscopy, m is the total number of individual characteristics of patients after digestive endoscopy, X j is the specific value corresponding to the individual characteristics of the jth patient after digestive endoscopy, φ j is the weight coefficient corresponding to the individual characteristics of the jth patient after gastrointestinal endoscopy, and ξ is the correction coefficient of the risk of recovery after gastrointestinal endoscopy.

10. The deep learning-based postoperative intelligent nursing system for digestive endoscopy according to claim 1 is characterized in that: The postoperative recovery risk intelligent nursing module includes the following functions: Obtain clinical experience after digestive endoscopy and determine the risk grading system for post-endoscopic recovery based on clinical experience after digestive endoscopy; Based on the post-endoscopic recovery risk grading system, the recovery risk level of patients after digestive endoscopy is determined according to the degree of recovery risk after digestive endoscopy, including low recovery risk level, medium recovery risk level and high recovery risk level; Postoperative intelligent nursing analysis is performed based on the corresponding recovery risk level of patients after gastrointestinal endoscopy, and recovery nursing recommendations for patients after gastrointestinal endoscopy are generated to execute corresponding risk recovery nursing tasks after gastrointestinal endoscopy.

Images