Auxiliary evaluation system for deep venous thrombosis occurrence risk of critical patient in plateau region

Abstract

The invention belongs to the technical field of intelligent medical treatment, and discloses an auxiliary assessment system for deep venous thrombosis occurrence risks of critical patients in plateau regions. Comprising a data acquisition module used for acquiring basic data, numerical feature data and image feature data of a patient; the index evaluation module is used for performing group division on the numerical feature data and evaluating real-time indexes; the index detection module is used for acquiring historical indexes and identifying abnormal indexes in the real-time indexes based on the historical indexes and the real-time indexes; the image analysis module is used for analyzing the image feature data, extracting blood vessel feature data and quantifying a form anomaly index; the risk assessment module is used for fusing the abnormal index and the form abnormal index, assessing a thrombus risk coefficient and obtaining a risk degree; the method can effectively improve the assessment accuracy of the occurrence risk of deep venous thrombosis, timely identify high-risk patients, and reduce the occurrence rate of deep venous thrombosis and related complications.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical technology, and more specifically, to an auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas. Background Art

[0002] Due to its unique hypoxic and low-temperature environment, the plateau area has a significant impact on the physiological functions of critically ill patients, especially the risk of deep vein thrombosis (DVT) is significantly increased; deep vein thrombosis refers to the formation of thrombus in deep veins, commonly in the deep veins of the lower extremities, which may lead to blood flow obstruction, pain, and swelling, and may also cause fatal complications such as pulmonary embolism in severe cases; research shows that high-altitude exposure is an important risk factor for thromboembolic diseases, and the thrombotic tendency of people above 2000 meters above sea level is more than twice that of people at low altitudes, and the risk of people at high altitudes above 3000 meters can increase to 30 times; the pathological mechanism of this increased risk is closely related to the coagulation dysfunction induced by the plateau environment: hypoxia and low temperature activate hypoxia-inducible factor, up-regulate the expression of transferrin, and then enhance the activity of thrombin and coagulation factors, resulting in a hypercoagulable state of the blood; in addition, critically ill patients in plateau areas are often further aggravated by factors such as trauma, surgery, infection, or long-term bed rest, resulting in stasis of blood flow and vascular endothelial damage, forming a superimposed effect of Virchow's triad (stasis of blood flow, hypercoagulable state, vascular damage), thereby significantly increasing the risk of deep vein thrombosis; therefore, constructing an intelligent deep vein thrombosis risk assessment system suitable for critically ill patients in plateau areas has become an urgent problem to be solved.

[0003] The patent application with the publication number CN116264116A discloses a modeling method for a deep vein thrombosis risk prediction model for patients after lower limb fractures; including: S1, establishing a database and collecting data; S2, processing the collected clinical feature data so that the data can be used in a patterned manner; S3, establishing a decision model for the processed data so that the established model can estimate the incidence rate of DVT in patients after lower limb fractures; S4, optimizing the model and parameters; S5, selecting multiple optimal options of the above models and parameters; S6, integrating multiple above models and using deep neural network and recurrent neural network algorithms for learning to finally complete the model establishment; this invention can greatly improve the assessment accuracy and reduce the offset error.

[0004] However, although the above-mentioned technology can achieve the risk assessment of deep vein thrombosis, it mainly targets patients after lower limb fractures and does not consider the unique effects of environmental factors such as hypoxia and low air pressure in high-altitude areas on thrombosis formation, as well as the physiological characteristic differences between critically ill patients and ordinary patients in high-altitude areas. Therefore, the above-mentioned technology cannot comprehensively identify the special factors of thrombosis risk in high-altitude areas, resulting in a decrease in the accuracy of risk assessment results. At the same time, it is difficult to identify high-risk patients in a timely manner, resulting in patients missing the best prevention and treatment opportunities and increasing the incidence of deep vein thrombosis and related complications.

[0005] In view of this, the present invention proposes an auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in high-altitude areas to solve the above problems. Summary of the Invention

[0006] To overcome the above defects of the prior art and to achieve the above object, the present invention provides the following technical solution: An auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in high-altitude areas, comprising:

[0007] A data acquisition module for real-time acquisition of patient characteristic data, where the patient characteristic data includes patient basic data, numerical characteristic data, and image characteristic data;

[0008] An index evaluation module for grouping the numerical characteristic data by using a clustering algorithm, evaluating the corresponding physiological indexes according to the data within each group, and marking them as real-time indexes;

[0009] An index detection module for obtaining historical indexes according to the patient basic data, calculating the coefficient of variation corresponding to each real-time index based on the historical indexes and the real-time indexes, analyzing the coefficient of variation, and identifying abnormal indexes among the real-time indexes;

[0010] An image analysis module for analyzing the image characteristic data, extracting vascular characteristic data, and quantifying the morphological abnormality index based on the vascular characteristic data;

[0011] A risk assessment module for fusing the abnormal indexes and the morphological abnormality index, evaluating the thrombus risk coefficient, and obtaining the risk level corresponding to the thrombus risk coefficient according to a pre-constructed risk classification standard.

[0012] Further, the steps of grouping the numerical characteristic data include:

[0013] Step S101: Using a pre-trained word embedding model, converting each data in the numerical characteristic data into a corresponding vector and marking it as a physiological vector;

[0014] Step S102: Regarding each physiological vector as a point to be divided, and the points to be divided correspond to the physiological vectors one by one;

[0015] Step S103: Preset the number of groups \(a\), randomly select \(a\) points to be divided as the center points, and mark each center point as \(\alpha\). b , where \(b\in[1,a]\); mark the sample points that are not used as center points as division points, and mark each division point as \(\beta\). c , where \(c\in[1,A - a]\), and \(A\) is the number of physiological vectors.

[0016] Step S104: Establish \(a\) corresponding groups according to the \(a\) center points, and calculate the cosine similarity between each division point and each center point in turn, and mark it as similarity.

[0017] Step S105: Compare all the similarities corresponding to the division point \(\beta\). c Mark the center point corresponding to the maximum similarity as the best point, and divide the division point \(\beta\). c into the group corresponding to the best point.

[0018] Step S106: Let \(c = c + 1\).

[0019] Step S107: Loop steps S105 to S106 until \(c = A - a\), then the loop ends, and enter step S108.

[0020] Step S108: Recalculate the center point corresponding to each group.

[0021] Step S109: Loop steps S104 to S108 until the center points of each group recalculated in step S108 are the same as the corresponding center points calculated in the previous loop, then the loop ends, and obtain \(a\) groups and the corresponding division points, and the groups correspond to the physiological indicators one by one.

[0022] In step S108, the method for recalculating the center point corresponding to each group includes:

[0023] Count the number of points to be divided in each group, and mark it as the punctuation number; add the center points corresponding to each group in turn, and then divide by the corresponding punctuation number to obtain the new center point of each group.

[0024] Further, the method for evaluating physiological indicators includes:

[0025] Regard the data in each group as a set of data, and the data sets correspond to the groups one by one; input each set of data sets into the corresponding index evaluation model to evaluate the corresponding physiological indicators; there are \(a\) evaluation models in the index evaluation model, and the evaluation models correspond to the physiological indicators one by one; the \(a\) evaluation models are all deep neural network models, and the training processes of the \(a\) evaluation models are all the same.

[0026] The training process of the evaluation model includes:

[0027] Pre-collect r sets of data collections, where each of the r sets of data collections corresponds to a physiological index. Set the corresponding physiological index for each of the r sets of data collections. r is an integer greater than 1. Convert the data collection and the corresponding physiological index into a corresponding set of feature vectors. Use each set of feature vectors as the input of an evaluation model. The evaluation model outputs a set of predicted physiological indices corresponding to each data collection, and uses the actual physiological index corresponding to each data collection as the prediction target. The actual physiological index is the pre-set physiological index corresponding to the data collection. Use minimizing the sum of prediction errors of all data collections as the training target. Train the evaluation model until the sum of prediction errors converges and then stop training.

[0028] Further, the historical index is the physiological index under the normal physiological state evaluated for the same patient at historical moments. The same patient refers to a patient whose basic patient data is the same as the real-time collected basic patient data.

[0029] The steps of calculating the coefficient of variation corresponding to each real-time index include:

[0030] Step S201: Sequentially incrementally set digital labels for each physiological index and mark them as index labels. The range of the index label is [1, a].

[0031] Step S202: Select the physiological index with the index label d and mark it as an element, where d ∈ [1, a].

[0032] Step S203: Screen out the physiological index corresponding to the element from the historical indices and mark it as the detection element. Mark the real-time index corresponding to the element as the real-time element. Both the detection element and the real-time element are used as analysis elements.

[0033] Step S204: Calculate the element difference and the reference distance corresponding to each analysis element.

[0034] Step S205: Determine the reference element corresponding to each analysis element.

[0035] Step S206: Calculate the adjacent distance between each analysis element and its corresponding reference element according to the element difference and the reference distance.

[0036] Step S207: Calculate the local density corresponding to each analysis element according to the adjacent distance.

[0037] Step S208: Calculate the coefficient of variation corresponding to the real-time element according to the local density.

[0038] Step S209: Let d = d + 1.

[0039] Step S210: Loop through steps S202 to S209 until all physiological indicators are marked as elements, and then the loop ends;

[0040] The method for identifying abnormal indicators in real-time indicators includes:

[0041] Preset an identification coefficient h, where 0 < h < 1; subtract 1 from the dispersion coefficient of each real-time element, then take the absolute value to obtain the judgment coefficient of each real-time element; compare the judgment coefficient of each real-time element with the identification coefficient; if the judgment coefficient is greater than or equal to the identification coefficient, mark the real-time indicator corresponding to the corresponding real-time element as an abnormal indicator; if the judgment coefficient is less than the identification coefficient, do not mark the real-time indicator corresponding to the corresponding real-time element.

[0042] Further, in step S204, the method for calculating the element difference corresponding to each analysis element is: obtain the adjacent elements corresponding to each analysis element, and the adjacent elements are the remaining f analysis elements, where f is the number of detected elements; subtract each analysis element from the corresponding adjacent element to obtain the element difference corresponding to each analysis element;

[0043] The method for calculating the reference distance corresponding to each analysis element is:

[0044] Take the element differences corresponding to each analysis element as an element set, and the element set corresponds to the analysis element one by one; sort the element differences in each element set from small to large to generate an element sorting table corresponding to each element set; preset a screening quantity g, and take the element difference ranked at the g-th position in each element sorting table as the reference distance of the analysis element corresponding to the corresponding element set;

[0045] In step S205, the method for determining the reference element corresponding to each analysis element is: take the adjacent elements corresponding to the first g element differences in each element sorting table as the reference element of the analysis element corresponding to the corresponding element set;

[0046] In step S206, the expression for the adjacent distance is: xl(p,q) = max(jl(p), yc(p,q)); where xl(p,q) is the adjacent distance between the p-th analysis element and the corresponding q-th reference element, max is the maximum value function, jl(p) is the reference distance of the p-th analysis element, yc(p,q) is the element difference between the p-th analysis element and the corresponding q-th reference element, p ∈ [1, f + 1], q ∈ [1, g], and f > g;

[0047] In the step S207, the method for calculating the local density corresponding to each analysis element is as follows: successively add the adjacent distances corresponding to each analysis element to obtain the comprehensive distance corresponding to each analysis element; take the reciprocal of the comprehensive distance corresponding to each analysis element as the local density corresponding to each analysis element.

[0048] In the step S208, the method for calculating the dispersion coefficient corresponding to the real-time element is as follows: obtain the reference element corresponding to the real-time element and mark it as the evaluation element; divide the local density corresponding to each evaluation element by the local density corresponding to the real-time element respectively to obtain the relative density corresponding to each evaluation element; successively add the relative densities corresponding to each evaluation element, and then divide by g to obtain the dispersion coefficient corresponding to the real-time element.

[0049] Furthermore, the vascular feature data includes a curvature set and a diameter set.

[0050] The method for extracting the curvature set includes:

[0051] Preset a gray threshold, obtain the gray value corresponding to each pixel point in the image feature data, and compare the gray value of each pixel point with the gray threshold respectively; if the gray value is greater than the gray threshold, mark the corresponding pixel point as a vascular point; if the gray value is less than or equal to the gray threshold, do not mark the corresponding pixel point; based on all vascular points, use a skeletonization algorithm to extract the vascular centerline; randomly select k vascular points from all the vascular points corresponding to the vascular centerline and mark them as midline points, where 1 < k < l and l is the number of vascular points on the vascular centerline.

[0052] According to the image coordinate system built in the image feature data, obtain the coordinates corresponding to the k midline points and mark them as midline coordinates; according to the midline coordinates of the k midline points, use a curve fitting algorithm to perform curve fitting on the centerline to obtain the mathematical expression corresponding to the centerline and mark it as the curve expression; perform a first-order derivative on the curve expression to obtain the first-order expression; perform a second-order derivative on the curve expression to obtain the second-order expression; substitute each midline coordinate into the first-order expression to obtain the first-order value corresponding to each midline point; substitute each midline coordinate into the second-order expression to obtain the second-order value corresponding to each midline point; square each first-order value and then add one to obtain the first value; take the square root of the cube of each first value to obtain the second value corresponding to each midline point; divide the absolute value of the second-order value corresponding to each midline point by the corresponding second value to obtain the vascular curvature corresponding to each midline point; take the vascular curvature corresponding to each midline point as the curvature set.

[0053] Furthermore, the method for extracting the diameter set includes:

[0054] Take the negative reciprocal of the first-order value corresponding to each midline point as the slope of the normal line corresponding to each midline point; according to the midline coordinates of each midline point and the slope of the corresponding normal line, use the point-slope form to represent the normal line equation corresponding to each midline point; the expression of the normal line equation is: y = z(x - x0) + y0; where y is the ordinate of any point on the normal line, z is the slope, x is the abscissa of any point on the normal line, x0 is the abscissa of the midline point, and y0 is the ordinate of the midline point; analyze the adjacent points corresponding to each blood vessel point, and the adjacent points are the pixel points adjacent to the blood vessel point; if there are pixel points that are not marked as blood vessel points among the adjacent points, then mark the corresponding blood vessel point as an edge point; if all the adjacent points are blood vessel points, then do not mark the corresponding blood vessel point; substitute each edge point into each normal line equation, if the normal line equation holds, then take the corresponding edge point as the satisfied point of the corresponding normal line equation; obtain the coordinates corresponding to each satisfied point and mark them as satisfied coordinates; according to the satisfied coordinates, calculate the Euclidean distance between two satisfied points corresponding to each normal line equation and take it as the satisfied length; take the satisfied length corresponding to each normal line equation as the image diameter of the corresponding midline point; obtain the scaling factor, multiply each image diameter by the scaling factor to obtain the blood vessel diameter corresponding to each midline point; take the blood vessel diameter corresponding to each midline point as the diameter set.

[0055] Further, the method for quantifying the morphological abnormality index includes:

[0056] Add up each blood vessel curvature in turn and then divide by k to obtain the curvature mean value; subtract the curvature mean value from each blood vessel curvature respectively and square it to obtain the squared curvature difference; add up each squared curvature difference in turn and divide by k, and then take the square root to obtain the curvature standard deviation; divide the curvature standard deviation by the curvature mean value to obtain the curvature abnormality index; add up each blood vessel diameter in turn and then divide by k to obtain the diameter mean value; subtract the diameter mean value from each blood vessel diameter respectively and square it to obtain the squared diameter difference; add up each squared diameter difference in turn and divide by k, and then take the square root to obtain the diameter standard deviation; divide the diameter standard deviation by the diameter mean value to obtain the diameter abnormality coefficient;

[0057] Preset threshold coefficients, and the threshold coefficients include a curvature coefficient and a diameter coefficient; compare each blood vessel curvature with the curvature coefficient respectively, if the blood vessel curvature is greater than or equal to the curvature coefficient, then mark the corresponding blood vessel point as a curvature abnormal point, if the blood vessel curvature is less than the curvature coefficient, then do not mark the corresponding blood vessel point; compare each blood vessel diameter with the diameter coefficient respectively, if the blood vessel diameter is greater than or equal to the diameter coefficient, then mark the corresponding blood vessel point as a diameter abnormal point, if the blood vessel diameter is less than the diameter coefficient, then do not mark the corresponding blood vessel point; take the curvature abnormal points and the diameter abnormal points as local abnormal points, count the number of local abnormal points and mark it as the number of local abnormal points; divide the number of local abnormal points by 2k to obtain the local abnormality coefficient;

[0058] A preset weight set, which includes weight coefficients corresponding to the curvature anomaly index, the diameter anomaly index, and the local anomaly coefficient; multiply the curvature anomaly index, the diameter anomaly index, and the local anomaly coefficient by their corresponding weight coefficients respectively, and then add them in sequence to obtain the morphological anomaly index.

[0059] Further, the method for evaluating the thrombus risk coefficient includes:

[0060] Obtain environmental factors, which include environmental temperature, environmental air pressure, and oxygen concentration; use the environmental factors, the anomaly index, and the morphological anomaly index as evaluation data, and input the evaluation data into the trained risk assessment model to evaluate the thrombus risk coefficient of the patient; the training process of the risk assessment model is the same as that of the evaluation model, and both are deep neural network models.

[0061] Further, the method for obtaining the risk level corresponding to the thrombus risk coefficient includes:

[0062] According to the environmental factors, screen out the corresponding degree evaluation criteria from the pre-constructed risk classification criteria and mark them as the real-time evaluation criteria; where the risk classification criteria include u different degree evaluation criteria, u is an integer greater than 1, and each degree evaluation criterion includes a coefficient range corresponding to different risk levels; compare the thrombus risk coefficient with each coefficient range in the real-time evaluation criteria to obtain the risk level corresponding to the coefficient range where the thrombus risk coefficient is located.

[0063] The steps for screening the real-time evaluation criteria include:

[0064] Step S301: Construct a corresponding fuzzy set for each data in the environmental factors, and each fuzzy set includes multiple fuzzy categories;

[0065] Step S302: Map each data in the environmental factors to the membership degree of the corresponding each fuzzy category through the fuzzification technique respectively;

[0066] Step S303: Define fuzzy rules;

[0067] Step S304: Match the fuzzified environmental factors with the fuzzy rules, perform fuzzy inference to obtain the fuzzy inference result, and the fuzzy inference result is the membership degree corresponding to each degree evaluation criterion;

[0068] Step S305: Compare each membership degree in the fuzzy inference result, and select the degree evaluation criterion with the highest membership degree as the real-time evaluation criterion.

[0069] The technical effects and advantages of an auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas of the present invention:

[0070] By collecting various types of characteristic data of patients in real time, key information on the health status of patients can be comprehensively obtained; clustering algorithms and deep learning technologies are used to evaluate physiological indicators by group, and abnormal detection is performed on the evaluated physiological indicators to identify abnormal indicators, enabling timely detection of abnormal conditions in patients' physiological indicators; through image analysis, the curvature and diameter characteristics of blood vessels are quantitatively extracted, the morphological abnormality index is calculated, and multi-dimensional information is formed with abnormal indicators and environmental factors to accurately evaluate the thrombus risk coefficient of patients, and the specific risk level is obtained based on the pre-constructed risk classification criteria; fully considering the influence of environmental factors such as low oxygen and low air pressure in plateau areas, special factors of thrombus risk in plateau areas are comprehensively identified, the accuracy of the assessment of the risk of deep vein thrombosis in critically ill patients in plateau areas is improved, high-risk patients are identified in a timely manner, so as to provide effective early warning and intervention measures for clinical practice and reduce the incidence of deep vein thrombosis and related complications. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 FIG. is a schematic diagram of an auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to Embodiment 1 of the present invention;

[0072] Figure 2 FIG. is a flowchart of the functional relationship between modules according to Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0073] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0074] Embodiment 1

[0075] Please refer to Figure 1 As shown, the auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas described in this embodiment includes a data collection module, an index evaluation module, an index detection module, an image analysis module, and a risk assessment module; each module is connected by wired and / or wireless means to realize data transmission between modules; the functional relationship between modules is as Figure 2 shown.

[0076] The data collection module is used to collect patient characteristic data in real time, and the patient characteristic data includes patient basic data, numerical characteristic data, and image characteristic data.

[0077] Patient basic data includes personal information and medical history information; personal information includes gender, age, height, and weight; medical history information includes past medical history and family medical history; past medical history refers to the diseases that the patient had before seeking medical treatment, such as heart disease, hypertension, etc., which helps to judge the patient's health status and potential risks; family medical history refers to the diseases that the patient's family members (such as parents, siblings, etc.) have had or currently have, including genetic diseases, chronic diseases, or other health problems, which helps to identify genetic risks, understand the health problems that the patient may face, and the susceptibility to diseases; patient basic data will affect the normal range of physiological indicators, and personalized normal physiological indicators need to be obtained based on patient basic data, which helps to provide basic information for subsequent indicator detection and ensure the accuracy of abnormal indicator identification;

[0078] Numerical feature data is the patient's physiological data, which refers to objective physiological-related data that can be quantified numerically, such as blood oxygen saturation, red blood cell packing, blood viscosity, prothrombin time, fibrinogen level, etc.; numerical feature data helps to evaluate the patient's physiological indicators, detect the patient's physiological state, discover the patient's abnormalities in a timely manner, and achieve real-time physiological monitoring of the patient and early warning of thrombosis;

[0079] Image feature data is magnetic resonance imaging, which is used to evaluate the patient's vascular status and provide important visualization information for subsequent risk assessment of deep vein thrombosis;

[0080] Patient basic data is obtained through the electronic medical record system within the hospital system.

[0081] The index evaluation module is used to divide the numerical feature data into groups by using a clustering algorithm, evaluate the corresponding physiological indicators according to the data within each group, and mark them as real-time indicators.

[0082] The steps of dividing the numerical feature data into groups include:

[0083] Step S101: Use a pre-trained word embedding model (such as Word2Vec, GloVe, BERT, etc.) to convert each data in the numerical feature data into a corresponding vector and mark it as a physiological vector; it should be understood that when converting the numerical feature data into vectors, what is converted is the name of each data in the numerical feature data, rather than the specific values corresponding to these data; the word embedding model is an existing technology, and the specific training process will not be elaborated here;

[0084] Step S102: Take each physiological vector as a point to be divided, and the points to be divided correspond to the physiological vectors one by one;

[0085] Step S103: Preset the number of groups a, randomly select a points to be divided as the center points, and mark each center point as α b, b ∈ [1, a]; Mark the sample points that are not used as the center points as partition points, and mark each partition point as β c , c ∈ [1, A - a], where A is the number of physiological vectors; The number of groups a is preset by those skilled in the art according to the actual number of physiological indicators;

[0086] Step S104: Establish a corresponding a groups according to the a center points, and calculate the cosine similarity between each partition point and each center point in turn, and mark it as similarity. The calculation method of cosine similarity is prior art and will not be elaborated here;

[0087] Step S105: Compare all the similarities corresponding to the partition point β c , mark the center point corresponding to the maximum similarity as the best point, and divide the partition point β c into the group corresponding to the best point;

[0088] Step S106: Let c = c + 1;

[0089] Step S107: Loop steps S105 to S106 until c = A - a, then the loop ends and enter step S108;

[0090] Step S108: Recalculate the center point corresponding to each group;

[0091] Step S109: Loop steps S104 to S108 until the center points of each group recalculated in step S108 are the same as the corresponding center points calculated in the previous loop, then the loop ends, and obtain a groups and the corresponding partition points. The groups correspond to the physiological indicators one by one; Physiological indicators such as hypoxic tolerance, coagulation status, hemodynamic status, etc. Among them, hypoxic tolerance corresponds to blood oxygen saturation, red blood cell accumulation, etc. in the numerical characteristic data, coagulation status corresponds to prothrombin time, fibrinogen level, etc. in the numerical characteristic data, and hemodynamic status corresponds to blood viscosity, etc. in the numerical characteristic data.

[0092] In the above step S108, the method for recalculating the center point corresponding to each group includes:

[0093] Count the number of points to be divided in each group and mark it as the punctuation number; Add up the center points corresponding to each group in turn, and then divide by the corresponding punctuation number to obtain the new center point of each group.

[0094] The method for evaluating physiological indicators includes:

[0095] Take the data within each group as a set of data, and the data sets correspond to the groups one by one; input each set of data into the corresponding index evaluation model to evaluate the corresponding physiological index; there are a evaluation models in the index evaluation model, and the evaluation models correspond to the physiological indexes one by one, that is, one evaluation model is used to evaluate one physiological index; the a evaluation models are all deep neural network models, and the training processes of the a evaluation models are all the same.

[0096] The training process of the evaluation model includes:

[0097] Pre-collect r sets of data, and the r sets of data all correspond to a physiological index. Set the corresponding physiological index for the r sets of data. r is an integer greater than 1. Convert the data set and the corresponding physiological index into a corresponding set of feature vectors; the physiological index corresponding to the data set is collected by those skilled in the art during the process of historical evaluation of physiological indexes. r sets of data are collected, and each set of data is analyzed in turn in combination with the actual situation to evaluate the physiological index corresponding to each set of data, and the corresponding physiological index is set for the r sets of data in turn;

[0098] Take each set of feature vectors as the input of the evaluation model. The evaluation model outputs a set of predicted physiological indexes corresponding to each set of data, and takes the actual physiological index corresponding to each set of data as the prediction target. The actual physiological index is the physiological index preset corresponding to the data set; take minimizing the sum of the prediction errors of all data sets as the training target; among them, the calculation formula of the prediction error is η w =(θ w -ε w ) 2 , where η w is the prediction error, w is the group number of the feature vector corresponding to the data set, θ w is the predicted physiological index corresponding to the wth set of data, and ε w is the actual physiological index corresponding to the wth set of data; train the evaluation model until the sum of the prediction errors reaches convergence and then stop training.

[0099] The index detection module is used to obtain historical indexes according to the basic patient data, calculate the coefficient of variation corresponding to each real-time index based on the historical indexes and real-time indexes, analyze the coefficient of variation, and identify abnormal indexes in the real-time indexes.

[0100] The historical index is the physiological index in the normal physiological state evaluated at the historical moment for the same patient; the same patient is the patient with the same basic patient data as the real-time collected basic patient data; the historical index is obtained through the electronic medical record system in the hospital system.

[0101] The steps of calculating the coefficient of variation corresponding to each real-time index include:

[0102] Step S201: Sequentially incrementally set digital tags for each physiological index and mark them as index tags, where the range of the index tags is [1, a];

[0103] Step S202: Select the physiological index with the index tag d and mark it as an element, where d ∈ [1, a];

[0104] Step S203: Screen out the physiological indexes corresponding to the element from the historical indexes and mark them as detected elements, mark the real-time indexes corresponding to the element as real-time elements, and use both the detected elements and the real-time elements as analysis elements;

[0105] Step S204: Calculate the element difference and the reference distance corresponding to each analysis element;

[0106] Step S205: Determine the reference element corresponding to each analysis element;

[0107] Step S206: Calculate the adjacent distance between each analysis element and the corresponding reference element according to the element difference and the reference distance;

[0108] Step S207: Calculate the local density corresponding to each analysis element according to the adjacent distance;

[0109] Step S208: Calculate the coefficient of variation corresponding to the real-time element according to the local density;

[0110] Step S209: Let d = d + 1;

[0111] Step S210: Loop through steps S202 to S209 until all physiological indexes are marked as elements, and then the loop ends.

[0112] In the above step S204, the method for calculating the element difference corresponding to each analysis element is as follows: Obtain the adjacent elements corresponding to each analysis element, where the adjacent elements are the remaining f analysis elements, and f is the number of detected elements; Subtract each analysis element from the corresponding adjacent element to obtain the element difference corresponding to each analysis element;

[0113] The method for calculating the reference distance corresponding to each analysis element is as follows:

[0114] Take the element difference corresponding to each analysis element as an element set, and the element set corresponds to the analysis element one by one; Sort the element differences in each element set from small to large to generate an element sorting table corresponding to each element set; Preset a screening quantity g, and the screening quantity g is preset by those skilled in the art according to the actual situation; Take the element difference ranked at the g-th position in each element sorting table as the reference distance of the analysis element corresponding to the corresponding element set.

[0115] In the above step S205, the method for determining the reference element corresponding to each analysis element is as follows: The adjacent elements corresponding to the difference of the first g elements in each element sorting table are used as the reference elements of the analysis elements corresponding to the corresponding element set.

[0116] In the above step S206, the expression for the adjacent distance is: xl(p,q) = max(jl(p), yc(p,q)); where xl(p,q) is the adjacent distance between the p-th analysis element and the corresponding q-th reference element, max is the maximum value function, jl(p) is the reference distance of the p-th analysis element, yc(p,q) is the element difference between the p-th analysis element and the corresponding q-th reference element, p ∈ [1, f + 1], q ∈ [1, g], and f > g.

[0117] In the above step S207, the method for calculating the local density corresponding to each analysis element is as follows: The adjacent distances corresponding to each analysis element are added in sequence to obtain the comprehensive distance corresponding to each analysis element; The reciprocal of the comprehensive distance corresponding to each analysis element is used as the local density corresponding to each analysis element.

[0118] In the above step S208, the method for calculating the dispersion coefficient corresponding to the real-time element is as follows: Obtain the reference element corresponding to the real-time element and mark it as the evaluation element; Divide the local density corresponding to each evaluation element by the local density corresponding to the real-time element to obtain the relative density corresponding to each evaluation element; Add the relative densities corresponding to each evaluation element in sequence, and then divide by g to obtain the dispersion coefficient corresponding to the real-time element.

[0119] The method for identifying abnormal indicators in real-time indicators includes:

[0120] Preset an identification coefficient h, 0 < h < 1, and the identification coefficient h is preset by those skilled in the art according to the actual situation; Subtract 1 from the dispersion coefficient of each real-time element, and then take the absolute value to obtain the judgment coefficient of each real-time element; Compare the judgment coefficient of each real-time element with the identification coefficient; If the judgment coefficient is greater than or equal to the identification coefficient, mark the real-time indicator corresponding to the corresponding real-time element as an abnormal indicator; If the judgment coefficient is less than the identification coefficient, do not mark the real-time indicator corresponding to the corresponding real-time element.

[0121] An image analysis module for analyzing image feature data, extracting vascular feature data, and quantifying the morphological abnormality index based on the vascular feature data.

[0122] The vascular feature data includes a curvature set and a diameter set;

[0123] The method for extracting the curvature set includes:

[0124] A preset gray threshold value, which is preset by those skilled in the art according to the actual situation; obtain the gray value corresponding to each pixel point in the image feature data, and compare the gray value of each pixel point with the gray threshold value respectively; if the gray value is greater than the gray threshold value, mark the corresponding pixel point as a blood vessel point; if the gray value is less than or equal to the gray threshold value, do not mark the corresponding pixel point; based on all the blood vessel points, use a skeletonization algorithm (such as Zhang-Suen thinning algorithm, Guo-Hall thinning algorithm, etc.) to extract the blood vessel center line; randomly select k blood vessel points from all the blood vessel points corresponding to the blood vessel center line and mark them as midline points, where 1 < k < l and l is the number of blood vessel points on the blood vessel center line;

[0125] According to the image coordinate system built in the image feature data, obtain the coordinates corresponding to the k midline points and mark them as midline coordinates; according to the midline coordinates of the k midline points, use a curve fitting algorithm (such as the least squares method, B-spline curve algorithm, etc.) to perform curve fitting on the center line, obtain the mathematical expression corresponding to the center line and mark it as the curve expression; perform a first-order derivative on the curve expression to obtain the first-order expression; perform a second-order derivative on the curve expression to obtain the second-order expression; substitute each midline coordinate into the first-order expression to obtain the first-order value corresponding to each midline point; substitute each midline coordinate into the second-order expression to obtain the second-order value corresponding to each midline point; square each first-order value and then add one to obtain the first value; take the square root of the cube of each first value to obtain the second value corresponding to each midline point; divide the absolute value of the second-order value corresponding to each midline point by the corresponding second value to obtain the blood vessel curvature corresponding to each midline point; use the blood vessel curvature corresponding to each midline point as the curvature set.

[0126] The method for extracting the diameter set includes:

[0127] Take the negative reciprocal of the first-order value corresponding to each midline point as the slope of the normal line corresponding to each midline point; according to the midline coordinates of each midline point and the slope of the corresponding normal line, use the point-slope form to represent the normal line equation corresponding to each midline point; the expression of the normal line equation is: y = z(x - x0) + y0; where y is the ordinate of any point on the normal line, z is the slope, x is the abscissa of any point on the normal line, x0 is the abscissa of the midline point, and y0 is the ordinate of the midline point; analyze the adjacent points corresponding to each blood vessel point, and the adjacent points are the pixel points adjacent to the blood vessel point; if there are pixel points among the adjacent points that are not marked as blood vessel points, then mark the corresponding blood vessel point as an edge point; if all the adjacent points are blood vessel points, then do not mark the corresponding blood vessel point; substitute each edge point into each normal line equation, if the normal line equation holds, then take the corresponding edge point as the satisfaction point of the corresponding normal line equation; obtain the coordinates corresponding to each satisfaction point and mark them as satisfaction coordinates; according to the satisfaction coordinates, calculate the Euclidean distance between the two satisfaction points corresponding to each normal line equation and use it as the satisfaction length; the calculation method of the Euclidean distance is the prior art and will not be elaborated here; take the satisfaction length corresponding to each normal line equation as the image diameter corresponding to the midline point; obtain the scaling factor, multiply each image diameter by the scaling factor to obtain the blood vessel diameter corresponding to each midline point; take the blood vessel diameter corresponding to each midline point as the diameter set; the scaling factor is obtained according to the image resolution of the image feature data, and the image resolution is obtained through the metadata of the image feature data.

[0128] The method for quantifying the morphological abnormality index includes:

[0129] Add up each blood vessel curvature in turn and then divide by k to obtain the curvature mean; subtract the curvature mean from each blood vessel curvature respectively and square it to obtain the squared curvature difference; add up each squared curvature difference in turn, divide by k, and take the square root to obtain the curvature standard deviation; divide the curvature standard deviation by the curvature mean to obtain the curvature abnormality index; add up each blood vessel diameter in turn and then divide by k to obtain the diameter mean; subtract the diameter mean from each blood vessel diameter respectively and square it to obtain the squared diameter difference; add up each squared diameter difference in turn, divide by k, and take the square root to obtain the diameter standard deviation; divide the diameter standard deviation by the diameter mean to obtain the diameter abnormality coefficient;

[0130] A preset threshold coefficient, where the threshold coefficient includes a curvature coefficient and a diameter coefficient, and the threshold coefficient is preset by those skilled in the art according to the actual situation; compare each blood vessel curvature with the curvature coefficient respectively. If the blood vessel curvature is greater than or equal to the curvature coefficient, mark the corresponding blood vessel point as a curvature anomaly point. If the blood vessel curvature is less than the curvature coefficient, do not mark the corresponding blood vessel point; compare each blood vessel diameter with the diameter coefficient respectively. If the blood vessel diameter is greater than or equal to the diameter coefficient, mark the corresponding blood vessel point as a diameter anomaly point. If the blood vessel diameter is less than the diameter coefficient, do not mark the corresponding blood vessel point; regard both the curvature anomaly points and the diameter anomaly points as local anomaly points, count the number of local anomaly points, and mark it as the number of local anomaly points; divide the number of local anomaly points by 2k to obtain the local anomaly coefficient.

[0131] A preset weight set, where the weight set includes weight coefficients corresponding to a curvature anomaly index, a diameter anomaly index, and a local anomaly coefficient, and the weight set is preset by those skilled in the art according to the actual situation; multiply the curvature anomaly index, the diameter anomaly index, and the local anomaly coefficient by their corresponding weight coefficients respectively, and then add them up in sequence to obtain the morphological anomaly index.

[0132] A risk assessment module, which is used to fuse the anomaly indicators and the morphological anomaly index, evaluate the thrombus risk coefficient, and obtain the risk level corresponding to the thrombus risk coefficient according to a pre-constructed risk classification standard.

[0133] The method for evaluating the thrombus risk coefficient includes:

[0134] Obtain environmental factors, where the environmental factors include environmental temperature, environmental air pressure, and oxygen concentration; the environmental temperature is obtained through a temperature sensor installed in the intensive care unit, the environmental air pressure is obtained through a barometric pressure sensor installed in the intensive care unit, and the oxygen concentration is obtained through an oxygen concentration sensor installed in the intensive care unit; it should be understood that the reason for obtaining environmental factors is that the air pressure in the plateau area is relatively low, which will cause the air to be thin and the oxygen partial pressure to decrease. In order to adapt to the hypoxic environment, patients will have an increase in blood viscosity and obstruction of venous return, thus increasing the risk of deep vein thrombosis; moreover, due to the thin air in the plateau area, the oxygen concentration in the blood will decrease, and hypoxia will affect blood circulation, including vasoconstriction, increase in coagulation factors, etc., thus increasing the risk of deep vein thrombosis; in addition, the temperature in the plateau area is usually relatively low, and the cold environment will cause vasoconstriction, affecting blood circulation, especially the blood flow in the limbs slows down, making the risk of deep vein thrombosis more serious; therefore, by obtaining environmental factors, it is possible to comprehensively consider the unique effects of environmental factors such as hypoxia and low air pressure in the plateau area on thrombus formation, which helps to more accurately evaluate the risk of deep vein thrombosis in critically ill patients in the plateau area.

[0135] Taking environmental factors, abnormal indicators, and morphological abnormality indices as evaluation data, inputting the evaluation data into a trained risk assessment model to evaluate the patient's thrombus risk coefficient; the training process of the risk assessment model is the same as that of the evaluation model, and both are deep neural network models.

[0136] The methods for obtaining the risk level corresponding to the thrombus risk coefficient include:

[0137] According to the environmental factors, select the corresponding degree evaluation criteria from the pre-constructed risk classification criteria and mark them as real-time evaluation criteria; among them, the risk classification criteria include u different degree evaluation criteria, where u is an integer greater than 1, and each degree evaluation criterion includes a coefficient range corresponding to different risk levels, such as the coefficient range corresponding to high risk, the coefficient range corresponding to medium risk, the coefficient range corresponding to low risk, etc.; the risk classification criteria are pre-constructed by those skilled in the art according to the actual situation; compare the thrombus risk coefficient with each coefficient range in the real-time evaluation criteria to obtain the risk level corresponding to the coefficient range where the thrombus risk coefficient is located.

[0138] The steps for screening the real-time evaluation criteria include:

[0139] Step S301: Construct a corresponding fuzzy set for each data in the environmental factors, and each fuzzy set includes multiple fuzzy categories; for example: the fuzzy categories corresponding to environmental temperature are low temperature, medium temperature, high temperature, etc., the fuzzy categories corresponding to environmental air pressure are low air pressure, medium air pressure, high air pressure, etc., and the fuzzy categories corresponding to oxygen concentration are low concentration, medium concentration, high concentration, etc.;

[0140] Step S302: Map each data in the environmental factors to the membership degree of the corresponding each fuzzy category through a fuzzification technique respectively; Fuzzification is the process of mapping an exact value to the membership degree corresponding to a fuzzy category, and fuzzification techniques such as triangular membership function, trapezoidal membership function, etc.; for example, if the value of environmental air pressure is low, then it is inferred that the membership degree of low air pressure is 0.9, the membership degree of medium air pressure is 0.1, and the membership degree of high air pressure is 0;

[0141] Step S303: Define fuzzy rules, and the fuzzy rules are defined according to expert knowledge or relevant literature; for example, if the environmental temperature is low, the environmental air pressure is low, and the oxygen concentration is low, then it is inferred that the probability of degree evaluation criterion A as the real-time evaluation criterion is high; if the environmental temperature is medium, the environmental air pressure is medium, and the oxygen concentration is medium, then it is inferred that the probability of degree evaluation criterion B as the real-time evaluation criterion is high; among them, in degree evaluation criterion A, the coefficient range corresponding to low risk is relatively narrow, and the coefficient ranges corresponding to medium risk and high risk are relatively wide, which is applicable to plateau areas; in degree evaluation criterion B, the coefficient ranges corresponding to low risk, medium risk, and high risk are evenly spaced, which is applicable to conventional environments;

[0142] Step S304: Match the fuzzified environmental factors with the fuzzy rules, perform fuzzy inference, and obtain the fuzzy inference result. The fuzzy inference result is the membership degree corresponding to each degree evaluation criterion. The fuzzy inference method can be, for example, the Mamdani or Sugeno fuzzy inference method. The fuzzy inference result is, for example, the membership degree of degree evaluation criterion C is 0.1, the membership degree of degree evaluation criterion B is 0.3, and the membership degree of degree evaluation criterion A is 0.6.

[0143] Step S305: Compare each membership degree in the fuzzy inference result, and select the degree evaluation criterion with the highest membership degree as the real-time evaluation criterion.

[0144] In this embodiment, by collecting various types of characteristic data of patients in real time, the key information of the patients' health status can be comprehensively obtained; using the clustering algorithm and deep learning technology to evaluate the physiological indicators by group, and performing anomaly detection on the evaluated physiological indicators to identify abnormal indicators, the abnormal conditions in the patients' physiological indicators can be timely discovered; through image analysis, the curvature and diameter characteristics of blood vessels are quantitatively extracted, the morphological abnormality index is calculated, and multi-dimensional information is formed with the abnormal indicators and environmental factors to accurately evaluate the thrombus risk coefficient of the patients, and the specific risk level is obtained based on the pre-constructed risk division standard; fully considering the influence of environmental factors such as low oxygen and low air pressure in the plateau area, comprehensively identifying the special factors of thrombus risk in the plateau area, improving the evaluation accuracy of the risk of deep vein thrombosis in critically ill patients in the plateau area, timely identifying high-risk patients, so as to provide effective early warning and intervention measures for clinical practice, and reducing the incidence of deep vein thrombosis and related complications.

[0145] Embodiment 2

[0146] The present application also provides an electronic device. The electronic device may include one or more processors and one or more memories. Among them, computer-readable code is stored in the memory, and when the computer-readable code is run by one or more processors, it can execute a system for assisting in evaluating the risk of deep vein thrombosis in critically ill patients in the plateau area as described above.

[0147] The method or system according to the embodiment of the present application can also be implemented by means of the architecture of the electronic device shown in the present application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to the network, input / output, a hard disk, etc. The storage device in the electronic device, such as ROM or hard disk, can store a system for assisting in evaluating the risk of deep vein thrombosis in critically ill patients in the plateau area provided by the present application. Further, the electronic device may also include a user interface. Of course, the architecture shown in the present application is only exemplary, and when implementing different devices, one or more components shown in the electronic device of the present application can be omitted according to actual needs.

[0148] Embodiment 3

[0149] One embodiment of the present application discloses a computer-readable storage medium. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are run by a processor, a risk auxiliary assessment system for deep vein thrombosis in critically ill patients in plateau areas according to the embodiments of the present application described with reference to the above drawings can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory, etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0150] In addition, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium storing machine-readable instructions that can be run by a processor to execute instructions corresponding to the method steps provided by the present application, such as: a risk auxiliary assessment system for deep vein thrombosis in critically ill patients in plateau areas. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0151] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, and all should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

[0152] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas, characterized in that, Including: A data acquisition module, configured to acquire patient characteristic data in real time, where the patient characteristic data includes patient basic data, numerical characteristic data, and image characteristic data; An index evaluation module, configured to use a clustering algorithm to divide the numerical characteristic data into groups, evaluate the corresponding physiological indexes according to the data within each group, and mark them as real-time indexes; An index detection module, configured to obtain historical indexes according to the patient basic data, calculate the coefficient of variation corresponding to each real-time index based on the historical indexes and the real-time indexes, analyze the coefficient of variation, and identify abnormal indexes in the real-time indexes; An image analysis module, configured to analyze the image characteristic data, extract vascular characteristic data, and quantify the morphological abnormality index based on the vascular characteristic data; A risk assessment module, configured to fuse the abnormal indexes and the morphological abnormality index, evaluate the thrombus risk coefficient, and obtain the risk level corresponding to the thrombus risk coefficient according to a pre-constructed risk division standard.

2. The auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to claim 1, wherein The steps of dividing the numerical characteristic data into groups include: Step S101: Use a pre-trained word embedding model to convert each data in the numerical characteristic data into a corresponding vector, and mark it as a physiological vector; Step S102: Take each physiological vector as a point to be divided, and the points to be divided correspond to the physiological vectors one by one; Step S103: Preset the number of groups \(a\), randomly select \(a\) points to be divided as the center points, and mark each center point as \(\alpha_b\), where \(b\in[1,a]\); mark the sample points that are not used as center points as division points, and mark each division point as \(\beta\). c , where \(c\in[1,A - a]\) and \(A\) is the number of physiological vectors. Step S104: Establish corresponding a groups according to a center points, and calculate the cosine similarity between each division point and each center point in turn, and mark it as similarity; Step S105: Compare all the similarities corresponding to the division point β c Mark the center point corresponding to the maximum similarity as the best point, and divide the division point β c into the group corresponding to the best point; Step S106: Let c = c + 1; Step S107: Loop steps S105 to S106 until the loop ends when c = A - a, and enter step S108; Step S108: Recalculate the center point corresponding to each group; Step S109: Loop steps S104 to S108 until the center points of each group recalculated in step S108 are the same as the corresponding center points calculated in the previous loop, then the loop ends, and a groups and the corresponding division points are obtained, and the groups correspond to the physiological indexes one by one; In the step S108, the method for recalculating the center point corresponding to each group includes: Count the number of points to be divided in each group, and mark it as the punctuation number; add up the center points corresponding to each group in turn, and then divide by the corresponding punctuation number to obtain the new center point of each group.

3. The auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to claim 2, characterized in that, The method for evaluating physiological indexes includes: Take the data within each group as a data set, and the data sets correspond to the groups one by one; input each data set into the corresponding index evaluation model to evaluate the corresponding physiological index; the index evaluation model includes a evaluation models, and the evaluation models correspond to the physiological indexes one by one; the a evaluation models are all deep neural network models, and the training processes of the a evaluation models are all the same; The training process of the evaluation model includes: Pre-collect r sets of data collections. Each of the r sets of data collections corresponds to a physiological index. Set the corresponding physiological index for each of the r sets of data collections. r is an integer greater than 1. Convert the data collection and the corresponding physiological index into a corresponding set of feature vectors; use each set of feature vectors as the input of the evaluation model. The evaluation model outputs a set of predicted physiological indices corresponding to each data collection, and uses the actual physiological index corresponding to each data collection as the prediction target. The actual physiological index is the pre-set physiological index corresponding to the data collection; use minimizing the sum of prediction errors of all data collections as the training target; train the evaluation model until the sum of prediction errors reaches convergence and then stop training.

4. An auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to claim 3, characterized in that, The historical index is the physiological index evaluated in the normal physiological state at the historical moment for the same patient; The same patient refers to a patient whose basic patient data is the same as the real-time collected basic patient data; The step of calculating the coefficient of variation corresponding to each real-time index includes: Step S201: Sequentially incrementally set digital labels for each physiological index and mark them as index labels. The range of the index labels is [1, a]; Step S202: Select the physiological index with the index label d and mark it as an element, where d ∈ [1, a]; Step S203: Screen out the physiological index corresponding to the element from the historical indices and mark it as the detected element. Mark the real-time index corresponding to the element as the real-time element. Both the detected element and the real-time element are used as analysis elements; Step S204: Calculate the element difference and the reference distance corresponding to each analysis element; Step S205: Determine the reference element corresponding to each analysis element; Step S206: Calculate the adjacent distance between each analysis element and each corresponding reference element according to the element difference and the reference distance; Step S207: Calculate the local density corresponding to each analysis element according to the adjacent distance; Step S208: Calculate the coefficient of variation corresponding to the real-time element according to the local density; Step S209: Let d = d + 1; Step S210: Loop through steps S202 to S209 until all physiological indices are marked as elements, and then the loop ends; The method for identifying abnormal indices in the real-time indices includes: Preset an identification coefficient h, where 0 < h < 1; subtract 1 from the coefficient of variation of each real-time element and then take the absolute value to obtain the judgment coefficient of each real-time element; compare the judgment coefficient of each real-time element with the identification coefficient; if the judgment coefficient is greater than or equal to the identification coefficient, mark the real-time index corresponding to the corresponding real-time element as an abnormal index; if the judgment coefficient is less than the identification coefficient, do not mark the real-time index corresponding to the corresponding real-time element.

5. The auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to claim 4, characterized in that In step S204, the method for calculating the element difference corresponding to each analysis element is: obtain the adjacent elements corresponding to each analysis element. The adjacent elements are the remaining f analysis elements, where f is the number of detected elements; subtract each analysis element from the corresponding adjacent element to obtain the element difference corresponding to each analysis element; The method for calculating the reference distance corresponding to each analysis element is: Take the element difference corresponding to each analysis element as an element set, and the element set corresponds one-to-one with the analysis element; sort the element differences in each element set from small to large to generate an element sorting table corresponding to each element set; preset the screening quantity g, and take the element difference ranked at the g-th position in each element sorting table as the reference distance of the analysis element corresponding to the corresponding element set. In the step S205, the method for determining the reference element corresponding to each analysis element is: take the adjacent elements corresponding to the first g element differences in each element sorting table as the reference element of the analysis element corresponding to the corresponding element set. In the step S206, the expression of the adjacent distance is: xl(p,q) = max(jl(p), yc(p,q)); where xl(p,q) is the adjacent distance between the p-th analysis element and the corresponding q-th reference element, max is the maximum value function, jl(p) is the reference distance of the p-th analysis element, yc(p,q) is the element difference between the p-th analysis element and the corresponding q-th reference element, p ∈ [1, f + 1], q ∈ [1, g], f > g. In the step S207, the method for calculating the local density corresponding to each analysis element is: successively add the adjacent distances corresponding to each analysis element to obtain the comprehensive distance corresponding to each analysis element; take the reciprocal of the comprehensive distance corresponding to each analysis element as the local density corresponding to each analysis element. In the step S208, the method for calculating the discrete coefficient corresponding to the real-time element is: obtain the reference element corresponding to the real-time element and mark it as the evaluation element; divide the local density corresponding to each evaluation element by the local density corresponding to the real-time element respectively to obtain the relative density corresponding to each evaluation element; successively add the relative densities of each evaluation element, and then divide by g to obtain the discrete coefficient corresponding to the real-time element.

6. The auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to claim 5, wherein The vascular feature data includes a curvature set and a diameter set. The method for extracting the curvature set includes: Preset a gray threshold, obtain the gray value corresponding to each pixel point in the image feature data, and compare the gray value of each pixel point with the gray threshold respectively; if the gray value is greater than the gray threshold, mark the corresponding pixel point as a vascular point; if the gray value is less than or equal to the gray threshold, do not mark the corresponding pixel point; based on all vascular points, use a skeletonization algorithm to extract the vascular centerline; randomly select k vascular points from all vascular points corresponding to the vascular centerline and mark them as midline points, 1 < k < l, where l is the number of vascular points on the vascular centerline. According to the image coordinate system built in the image feature data, obtain the coordinates corresponding to k midline points and mark them as midline coordinates; according to the midline coordinates of the k midline points, use the curve fitting algorithm to perform curve fitting on the center line, obtain the mathematical expression corresponding to the center line, and mark it as the curve expression; perform the first-order derivative on the curve expression to obtain the first-order expression; perform the second-order derivative on the curve expression to obtain the second-order expression; substitute each midline coordinate into the first-order expression to obtain the first-order value corresponding to each midline point; substitute each midline coordinate into the second-order expression to obtain the second-order value corresponding to each midline point; square each first-order value and then add one to obtain the first value; cube each first value and then take the square root to obtain the second value corresponding to each midline point; divide the absolute value of the second-order value corresponding to each midline point by the corresponding second value to obtain the blood vessel curvature corresponding to each midline point; use the blood vessel curvature corresponding to each midline point as the curvature set.

7. The auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to claim 6, characterized in that, The method for extracting the diameter set includes: Take the negative reciprocal of the first-order value corresponding to each midline point as the slope of the normal line corresponding to each midline point; according to the midline coordinates of each midline point and the slope of the corresponding normal line, use the point-slope form to represent the normal line equation corresponding to each midline point; the expression of the normal line equation is: y = z(x - x0) + y0; where, y is the ordinate of any point on the normal line, z is the slope, x is the abscissa of any point on the normal line, x0 is the abscissa of the midline point, and y0 is the ordinate of the midline point; analyze the adjacent points corresponding to each blood vessel point, and the adjacent points are the pixel points adjacent to the blood vessel point; if there are pixel points among the adjacent points that are not marked as blood vessel points, then mark the corresponding blood vessel point as an edge point; if all the adjacent points are blood vessel points, then do not mark the corresponding blood vessel point; substitute each edge point into each normal line equation, if the normal line equation holds, then use the corresponding edge point as the satisfied point of the corresponding normal line equation; obtain the coordinates corresponding to each satisfied point and mark them as satisfied coordinates; according to the satisfied coordinates, calculate the Euclidean distance between two satisfied points corresponding to each normal line equation and use it as the satisfied length; use the satisfied length corresponding to each normal line equation as the image diameter corresponding to the corresponding midline point; obtain the scaling factor, multiply each image diameter by the scaling factor to obtain the blood vessel diameter corresponding to each midline point; use the blood vessel diameter corresponding to each midline point as the diameter set.

8. An auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to claim 7, characterized in that, The method for quantifying the morphological abnormality index includes: Add up each blood vessel curvature in turn and then divide by k to obtain the curvature mean value; subtract the curvature mean value from each blood vessel curvature and square it to obtain the squared curvature difference; add up each squared curvature difference in turn, divide by k, and take the square root to obtain the curvature standard deviation; divide the curvature standard deviation by the curvature mean value to obtain the curvature abnormality index; add up each blood vessel diameter in turn and then divide by k to obtain the diameter mean value; subtract the diameter mean value from each blood vessel diameter and square it to obtain the squared diameter difference; add up each squared diameter difference in turn, divide by k, and take the square root to obtain the diameter standard deviation; divide the diameter standard deviation by the diameter mean value to obtain the diameter abnormality coefficient; Preset threshold coefficients, where the threshold coefficients include curvature coefficients and diameter coefficients; compare each blood vessel curvature with the curvature coefficient respectively. If the blood vessel curvature is greater than or equal to the curvature coefficient, mark the corresponding blood vessel point as a curvature abnormal point. If the blood vessel curvature is less than the curvature coefficient, do not mark the corresponding blood vessel point; compare each blood vessel diameter with the diameter coefficient respectively. If the blood vessel diameter is greater than or equal to the diameter coefficient, mark the corresponding blood vessel point as a diameter abnormal point. If the blood vessel diameter is less than the diameter coefficient, do not mark the corresponding blood vessel point; regard both the curvature abnormal points and the diameter abnormal points as local abnormal points, count the number of local abnormal points, and mark it as the number of local abnormal points; divide the number of local abnormal points by 2k to obtain the local abnormal coefficient. Preset weight set, where the weight set includes weight coefficients corresponding to the curvature abnormal index, the diameter abnormal index, and the local abnormal coefficient; multiply the curvature abnormal index, the diameter abnormal index, and the local abnormal coefficient by their corresponding weight coefficients respectively, and then add them up in sequence to obtain the morphological abnormal index.

9. An auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to claim 8, characterized in that, The method for evaluating the thrombus risk coefficient includes: Obtain environmental factors, where the environmental factors include environmental temperature, environmental air pressure, and oxygen concentration; use the environmental factors, abnormal indicators, and morphological abnormal index as evaluation data, and input the evaluation data into the trained risk assessment model to evaluate the patient's thrombus risk coefficient; the training process of the risk assessment model is consistent with the training process of the evaluation model, and both are deep neural network models.

10. An auxiliary assessment system for the risk of deep vein thrombosis in critically ill patients in plateau areas according to claim 9, characterized in that, The method for obtaining the risk level corresponding to the thrombus risk coefficient includes: According to the environmental factors, select the corresponding degree evaluation criteria from the pre-constructed risk division criteria, and mark them as the real-time evaluation criteria; among them, the risk division criteria include u different degree evaluation criteria, where u is an integer greater than 1, and each degree evaluation criterion includes a coefficient range corresponding to different risk levels; compare the thrombus risk coefficient with each coefficient range in the real-time evaluation criteria to obtain the risk level corresponding to the coefficient range where the thrombus risk coefficient is located. The steps for screening the real-time evaluation criteria include: Step S301: Construct corresponding fuzzy sets for each data in the environmental factors, and each fuzzy set includes multiple fuzzy categories. Step S302: Map each data in the environmental factors to the membership degree of the corresponding each fuzzy category through the fuzzification technique respectively. Step S303: Define fuzzy rules. Step S304: Match the fuzzified environmental factors with the fuzzy rules, perform fuzzy inference, and obtain the fuzzy inference result. The fuzzy inference result is the membership degree corresponding to each degree evaluation criterion. Step S305: Compare each membership degree in the fuzzy inference result, and select the degree evaluation criterion with the highest membership degree as the real-time evaluation criterion.

Images