Sepsis Prognosis Method Based on Intelligent Auxiliary Analysis of Sublingual Vascular Images

Abstract

The present invention relates to the field of medical image technology, and particularly to a method for predicting the prognosis of sepsis based on intelligent auxiliary analysis of sublingual vascular images, including: obtaining grayscale images of sublingual microcirculation at different observation times; analyzing the law of gray-scale rise and fall of similar vascular dilatation based on the architecture of sublingual blood vessels in the grayscale images of sublingual microcirculation to obtain the saliva sedimentation degree in the vascular structure area; analyzing the flow and evaporation of sublingual saliva between adjacent observation times, and correcting the saliva sedimentation degree to obtain the corrected saliva sedimentation degree of the grayscale images of sublingual microcirculation at different observation times; and performing prognostic auxiliary analysis on the grayscale images of sublingual microcirculation according to the corrected saliva sedimentation degree. The present invention reduces the imaging interference of saliva on the sublingual blood vessels of sepsis patients, making the analysis results of sepsis prognosis more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of medical images, and particularly to a method for predicting the prognosis of sepsis based on intelligent auxiliary analysis of sublingual vascular images. Background Art

[0002] Sepsis is one of the major diseases that seriously threaten human health. The body's inflammatory response to infection is dysregulated, and it is secondary to various severe infections, such as severe burns, surgical operations, pneumonia, meningitis, etc. It is necessary to accurately evaluate the development of the disease and deal with it in a timely manner to avoid serious complications such as respiratory failure and heart failure. The microcirculation process under the patient's tongue can significantly reflect the abnormal changes in the sublingual microcirculation, which reflects the dysfunction of the microcirculation function in patients with sepsis. Therefore, non-invasive evaluation of microcirculation blood flow can be achieved through sublingual microcirculation technology. By observing the changes in microcirculation perfusion through a sublingual microscope, if there is microcirculation dysfunction, tissue hypoxia or organ failure may occur. Therefore, the sublingual mucosa is usually used as a perfusion window for visceral microcirculation to evaluate the prognosis level of sepsis patients, that is, the control effect of the condition.

[0003] During the monitoring of sublingual microcirculation images, the sublingual surface inevitably contains saliva, which will flow as a whole on the sublingual surface. Affected by the different blood flow rates inside the blood vessel ducts under different dilation conditions of the patient's sublingual surface, the reflection of saliva on different blood vessel surfaces under the tongue to light is different, and the blood vessel area under the tongue cannot be well displayed. Summary of the Invention

[0004] The present invention provides a method for predicting the prognosis of sepsis based on intelligent auxiliary analysis of sublingual vascular images to solve the existing problems: during the monitoring of sublingual microcirculation images, the sublingual surface inevitably contains saliva, which will flow as a whole on the sublingual surface. Affected by the different blood flow rates inside the blood vessel ducts under different dilation conditions of the patient's sublingual surface, the reflection of saliva on different blood vessel surfaces under the tongue to light is different, and the blood vessel area under the tongue cannot be well displayed.

[0005] The method for predicting the prognosis of sepsis based on intelligent auxiliary analysis of sublingual vascular images of the present invention adopts the following technical solutions:

[0006] It includes the following steps:

[0007] Obtain sublingual microcirculation grayscale images at different observation times;

[0008] For the gray-scale image of sublingual microcirculation at any observation moment, analyze the architecture of sublingual blood vessels in the gray-scale image of sublingual microcirculation, divide a blood vessel structure area from the gray-scale image of sublingual microcirculation, and the blood vessel structure area contains multiple blood vessel channel areas; based on the blood vessel channel areas, analyze the law of gray-scale rise and fall of similar blood vessel dilation, and obtain the saliva sedimentation degree of the blood vessel structure area;

[0009] Analyze the flow and evaporation of sublingual saliva between adjacent observation moments, correct the saliva sedimentation degree, and obtain the corrected saliva sedimentation degree of the gray-scale image of sublingual microcirculation at different observation moments;

[0010] According to the corrected saliva sedimentation degree, perform prognostic auxiliary analysis on the gray-scale image of sublingual microcirculation.

[0011] Further, the method for obtaining the blood vessel structure area is as follows:

[0012] The gray-scale image of sublingual microcirculation obtains several blood vessel skeleton points through a trained neural network; perform window sliding expansion on the blood vessel skeleton points to construct the blood vessel structure area.

[0013] Further, the method for obtaining the saliva sedimentation degree is as follows:

[0014] For any blood vessel structure area, analyze the similarity of the diameter expansion of blood vessel channels in the blood vessel structure area, divide an initial blood vessel swelling similar area from the blood vessel structure area; analyze the extension comparison of the initial blood vessel swelling similar area in the horizontal and vertical directions to obtain the blood vessel inclination degree of the initial blood vessel swelling similar area; analyze the gray-scale interval difference of blood vessel channels in the initial blood vessel swelling similar area to obtain the blood vessel saliva sedimentation degree of the initial blood vessel swelling similar area; comprehensively consider the blood vessel inclination degree and the blood vessel saliva sedimentation degree to obtain the saliva sedimentation degree of the blood vessel structure area; the saliva sedimentation degree is negatively correlated with the blood vessel inclination degree.

[0015] Further, the method for obtaining the initial blood vessel swelling similar area is as follows:

[0016] Obtain the channel widths of different blood vessel channel areas in the blood vessel structure area; analyze the similarity of the channel widths between different blood vessel channel areas, and integrate different blood vessel channel areas into the initial blood vessel swelling similar area.

[0017] Further, the method for obtaining the blood vessel inclination degree is as follows:

[0018] Obtain the horizontal area length in the initial blood vessel swelling similar area; obtain the vertical area length in the initial blood vessel swelling similar area; according to the comparison difference between the horizontal area length and the vertical area length, obtain the blood vessel inclination degree of the initial blood vessel swelling similar area.

[0019] Further, the method for obtaining the vascular saliva sedimentation degree is as follows:

[0020] Analyze the gray-scale difference between longitudinal vascular channels in the area with similar initial vascular swelling to obtain the vascular saliva sedimentation degree of the area with similar initial vascular swelling.

[0021] Further, after calculating the saliva sedimentation degree, it further includes:

[0022] Normalize the saliva sedimentation degree.

[0023] Further, the method for obtaining the corrected saliva sedimentation degree is as follows:

[0024] Take any observation moment as the target observation moment, compare the change difference of the saliva sedimentation degree before and after the target observation moment to obtain the abnormal dryness degree of saliva at the target observation moment; according to the abnormal dryness degree of saliva, obtain the corrected saliva sedimentation degree of the sublingual microcirculation gray-scale image at the target observation moment.

[0025] Further, the method for obtaining the abnormal dryness degree of saliva is as follows:

[0026] Take the observation moment before the target observation moment as the historical observation moment, take the observation moment after the target observation moment as the later observation moment, compare the difference in saliva sedimentation degree between the historical observation moment and the target observation moment to obtain the historical saliva sedimentation difference value; compare the difference in saliva sedimentation degree between the target observation moment and the later observation moment to obtain the later saliva sedimentation difference value; according to the historical saliva sedimentation difference value and the later saliva sedimentation difference value, obtain the abnormal dryness degree of saliva at the target observation moment.

[0027] Further, after calculating the corrected saliva sedimentation degree, it further includes:

[0028] Normalize the corrected saliva sedimentation degree.

[0029] The beneficial effects of the technical solution of the present invention are as follows: Based on analyzing the architecture of the sublingual blood vessels in the gray-scale image of the sublingual microcirculation, the law of gray-scale rise and fall of similar blood vessel dilation is analyzed to obtain the saliva sedimentation degree in the blood vessel structure area; the saliva sedimentation degree is used to describe the influence of the sublingual blood vessel structure of sepsis patients on the movement of sublingual saliva, making the connection between the influence on saliva flow and the physical state of sepsis patients closer; then, the flow and evaporation of sublingual saliva between adjacent observation times are analyzed to correct the saliva sedimentation degree, and the corrected saliva sedimentation degree of the gray-scale image of the sublingual microcirculation at different observation times is obtained; the corrected saliva sedimentation degree is used to describe the influence of the sublingual blood vessel structure of sepsis patients on the movement of sublingual saliva under the comprehensive influence of time dynamics, reducing the interference of normal sublingual blood vessels and time flow on the judgment of saliva state; finally, based on the corrected saliva sedimentation degree, prognostic auxiliary analysis is performed on the gray-scale image of the sublingual microcirculation; by analyzing the change relationship between the sublingual blood vessel structure of sepsis patients and the movement state of sublingual saliva, the present invention measures the corrected saliva sedimentation degree for prognostic analysis of sepsis; it reduces the imaging interference of saliva on the sublingual blood vessels of sepsis patients and makes the analysis result of sepsis prognosis more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0031] Figure 1 It is a flowchart of the steps of the sepsis prognosis method for intelligent auxiliary analysis based on sublingual blood vessel images of the present invention;

[0032] Figure 2 It is a schematic diagram of the gray-scale image of the sublingual microcirculation of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following combines the drawings and preferred embodiments to detail the specific implementation manner, structure, features and effects of the sepsis prognosis method for intelligent auxiliary analysis based on sublingual blood vessel images proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0034] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0035] The following specifically describes the specific solution of the sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images provided by the present invention in conjunction with the accompanying drawings.

[0036] Please refer to Figure 1 , which shows a flowchart of the steps of the sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images provided by an embodiment of the present invention. The method includes the following steps:

[0037] Step S001: Obtain sublingual microcirculation grayscale images at different observation times.

[0038] It should be noted that when monitoring sublingual microcirculation images, saliva is inevitably present on the sublingual surface. This saliva will flow as a whole on the sublingual surface. Affected by the different blood flow rates inside the blood vessel channels under different dilation conditions of the sublingual surface of the patient, the reflection of light by the saliva on different blood vessel surfaces under the tongue is different, and it cannot well display the blood vessel area under the tongue.

[0039] In a specific implementation manner of the embodiment of the present invention, the method for obtaining sublingual microcirculation grayscale images is as follows: Use a handheld vital microscope (HVM) to take a number of sublingual microcirculation images, and perform grayscale processing on each sublingual microcirculation image to obtain a sublingual microcirculation grayscale image. Among them, grayscale processing is a well-known technology and will not be elaborated in this embodiment; in addition, each sublingual microcirculation image corresponds to an observation time. Please refer to Figure 2 , which is a schematic diagram of a sublingual microcirculation grayscale image.

[0040] Specifically, in this embodiment, taking the total shooting duration of 10 seconds at a frequency of 1 time per second as an example, sublingual microcirculation images at multiple observation times are collected; the shooting frequency and the total shooting duration in this embodiment are not specifically limited, and the shooting frequency and the total shooting duration can be determined according to specific implementation situations.

[0041] So far, sublingual microcirculation grayscale images at different observation times are obtained through the above method.

[0042] Step S002: For the sublingual microcirculation grayscale image at any observation time, analyze the architecture of the sublingual blood vessels in the sublingual microcirculation grayscale image, divide the blood vessel structure area from the sublingual microcirculation grayscale image, and the blood vessel structure area contains multiple blood vessel channel areas; based on the blood vessel channel area, analyze the gray level rise and fall law of similar blood vessel dilation, and obtain the saliva sedimentation degree of the blood vessel structure area.

[0043] It should be noted that there are many small blood vessels with obvious branches distributed on the sublingual surface of the patient, which form different concave and convex distributions on the sublingual surface. As a result, when the saliva produced in the patient's oral cavity accumulates under the patient's tongue, it will be obstructed differently. Therefore, the structure of the blood vessels in the sublingual microcirculation gray image can be analyzed, and the blood vessel structure area can be divided from the sublingual microcirculation gray image.

[0044] Furthermore, it should be noted that when collecting the sublingual microcirculation gray image, the tongue will be lifted as much as possible. Under the action of natural gravity, usually, there is less saliva on the upper surface of the tongue, and more saliva on the lower surface close to the tongue bed. Therefore, generally, the imaging effect is better above, and the imaging effect is worse below due to the influence of saliva. So, based on the blood vessel channel area, the gray-scale rise and fall law of similar blood vessel dilation can be analyzed to obtain the saliva sedimentation degree of the blood vessel structure area; among them, the greater the saliva sedimentation degree, the faster the flow of saliva under the tongue in the blood vessel structure area, indicating that the interference of saliva on the angiography in the blood vessel structure area is greater.

[0045] Preferably, in some implementation manners of the embodiment of the present invention, the method for obtaining the blood vessel structure area is as follows: the sublingual microcirculation gray image obtains a number of blood vessel skeleton points through a trained neural network; the blood vessel skeleton points are expanded by window sliding to construct the blood vessel structure area. The specific process is as follows:

[0046] Input the sublingual microcirculation gray image into the trained neural network to obtain the blood vessel skeleton points of the sublingual microcirculation gray image; the neural network used in this embodiment is Resnet50. The method for obtaining the dataset for training this neural network is: collect a large number of sublingual microcirculation gray images, and artificially mark the positions of the blood vessel skeleton points in each sublingual microcirculation gray image, that is, mark the positions of the blood vessel skeleton points in the sublingual microcirculation gray image as 1, and mark the positions of non-blood vessel skeleton points as 0. This marking result is recorded as the label of each sublingual microcirculation gray image; collect a large number of sublingual microcirculation gray images and their corresponding labels to form a dataset; use this dataset to train this neural network, and the loss function used in the training process is the cross-entropy loss function; the specific training process is well-known content of the neural network, and the specific training process will not be elaborated in this embodiment.

[0047] Furthermore, taking any blood vessel skeleton point as an example, with this blood vessel skeleton point as the center, a sized window is preset, and the image area occupied by this window is used as the local blood vessel skeleton area of this blood vessel skeleton point; obtain the local blood vessel skeleton areas of all blood vessel skeleton points; the blood vessel skeleton points with intersections between different local blood vessel skeleton areas are used as real blood vessel skeleton points; the local blood vessel skeleton area of each real blood vessel skeleton point is used as the blood vessel channel area; and the union of all blood vessel channel areas is used as the blood vessel structure area. In this embodiment, For the sake of description, no specific limitation is imposed in this embodiment, where it can be determined according to the specific implementation situation.

[0048] It should be noted that each sublingual microcirculation grayscale image contains multiple vascular structure regions, and these regions are all independent of each other, without any overlapping parts.

[0049] Preferably, in some implementation manners of the embodiments of the present invention, the method for obtaining the saliva sedimentation degree is as follows: for any vascular structure region, analyze the similarity of the path diameter expansion of the blood vessel channels within the vascular structure region, and divide an initial blood vessel swelling similar region from the vascular structure region; analyze the extension comparison of the initial blood vessel swelling similar region in the horizontal and vertical directions to obtain the blood vessel inclination of the initial blood vessel swelling similar region; analyze the gray-scale interval difference of the blood vessel channels within the initial blood vessel swelling similar region to obtain the blood vessel saliva sedimentation degree of the initial blood vessel swelling similar region; combine the blood vessel inclination and the blood vessel saliva sedimentation degree to obtain the saliva sedimentation degree of the vascular structure region; the saliva sedimentation degree is negatively correlated with the blood vessel inclination. The specific process is as follows:

[0050] It should be noted that a large amount of inflammatory mediators will be produced in the body of patients with sepsis. These inflammatory mediators will affect vascular endothelial cells, especially endothelial cells in the microcirculation, causing similar damage to the tissue structure of the fine blood vessels under the tongue and resulting in a blood vessel swelling phenomenon with similar widths. Therefore, the similarity of the path diameter expansion of the blood vessel channels within the vascular structure region can be analyzed, and an initial blood vessel swelling similar region can be divided from the vascular structure region.

[0051] Preferably, in some implementation manners of the embodiments of the present invention, the method for obtaining the initial blood vessel swelling similar region is as follows: obtain the channel widths of different blood vessel channel regions within the vascular structure region; analyze the similarity of the channel widths between different blood vessel channel regions, and integrate different blood vessel channel regions into an initial blood vessel swelling similar region. The specific process is as follows:

[0052] Taking any blood vessel channel region within this vascular structure region as an example, the average value of the number of pixel points in all rows within this blood vessel channel region is used as the channel width of this blood vessel channel region; obtain the channel widths of all blood vessel channel regions; taking any two blood vessel channel regions as an example, preset a difference threshold , if the absolute value of the difference in the channel widths between these two blood vessel channel regions is less than , then these two blood vessel channel regions are jointly used as the initial blood vessel swelling similar region; obtain several initial blood vessel swelling similar regions. Here, this embodiment takes For the sake of description, no specific limitation is imposed in this embodiment, where it can be determined according to the specific implementation situation.

[0053] It should be noted that when saliva sinks under the patient's tongue, the vascular structure under vertical distribution provides the best support for the flow of saliva, and the vascular structures with inclined distribution will generate certain resistance to hinder the flow of saliva. Therefore, the extension comparison of the initial vascular swelling similar areas in the horizontal and vertical directions can be analyzed to obtain the vascular inclination of the initial vascular swelling similar areas. The greater the vascular inclination, the greater the hindrance of the sublingual blood vessels to the settlement of saliva.

[0054] Preferably, in some implementation manners of the embodiments of the present invention, the method for obtaining the vascular inclination is as follows: obtaining the length of the horizontal area within the initial vascular swelling similar area; obtaining the length of the vertical area within the initial vascular swelling similar area; and obtaining the vascular inclination of the initial vascular swelling similar area according to the comparison difference between the length of the horizontal area and the length of the vertical area. The specific process is as follows:

[0055] Taking any initial vascular swelling similar area as an example, the average value of the channel widths of all vascular channel areas within the initial vascular swelling similar area is used as the length of the horizontal area; the average value of the number of all pixel points in the columns occupied by all vascular channel areas within the initial vascular swelling similar area is used as the length of the vertical area; and the absolute value of the difference between the length of the horizontal area and the length of the vertical area is used as the vascular inclination of the initial vascular swelling similar area.

[0056] It should be noted that the greater the vascular inclination, the greater the hindrance of the sublingual blood vessels to the settlement of saliva.

[0057] Preferably, in some implementation manners of the embodiments of the present invention, the method for obtaining the vascular saliva settlement degree is as follows: analyzing the gray-scale difference between the longitudinal vascular channels within the initial vascular swelling similar area to obtain the vascular saliva settlement degree of the initial vascular swelling similar area. The specific process is as follows:

[0058] The average value of the gray-scale values of all pixel points in each column occupied by all vascular channel areas within the initial vascular swelling similar area is used as the longitudinal vascular gray-scale of each column; and the accumulated sum of the absolute values of the differences between the longitudinal vascular gray-scales of different columns is used as the vascular saliva settlement degree of the initial vascular swelling similar area.

[0059] Furthermore, the ratio of the vascular saliva settlement degree of the initial vascular swelling similar area to the vascular inclination is used as the regional saliva settlement degree of the initial vascular swelling similar area; and the normalized value of the average value of the regional saliva settlement degrees of all initial vascular swelling similar areas within the vascular structure area is used as the saliva settlement degree of the vascular structure area.

[0060] It should be especially noted that in this embodiment The function is normalized, and the normalization function can be determined according to specific implementation situations, which will not be elaborated in this embodiment.

[0061] It should be noted that the greater the salivary sedimentation degree, the faster the flow of saliva under the tongue in the vascular structure area, indicating that the interference of saliva on the contrast in the vascular structure area is greater.

[0062] Thus, the salivary sedimentation degree of the vascular structure area is obtained through the above method.

[0063] Step S003: Analyze the flow and evaporation of saliva under the tongue between adjacent observation times, correct the salivary sedimentation degree, and obtain the corrected salivary sedimentation degree of the gray-scale image of sublingual microcirculation at different observation times.

[0064] It should be noted that in actual situations, the saliva under the patient's tongue is in a dynamic flowing state, continuously flowing and evaporating over time; and the salivary sedimentation degree obtained by analyzing the saliva flow situation at a single observation time will have different interference effects over time. Therefore, the flow and evaporation of saliva under the tongue between adjacent observation times can be analyzed to correct the salivary sedimentation degree and obtain the corrected salivary sedimentation degree of the gray-scale image of sublingual microcirculation at different observation times.

[0065] Preferably, in some implementation manners of the embodiment of the present invention, the method for obtaining the corrected salivary sedimentation degree is as follows: Take any observation time as the target observation time, compare the change difference of the salivary sedimentation degree before and after the target observation time to obtain the abnormal dryness degree of saliva at the target observation time; According to the abnormal dryness degree of saliva, obtain the corrected salivary sedimentation degree of the gray-scale image of sublingual microcirculation at the target observation time. The specific process is as follows:

[0066] Preferably, in some implementation manners of the embodiment of the present invention, the method for obtaining the abnormal dryness degree of saliva is as follows: Take the observation time before the target observation time as the historical observation time, and the observation time after the target observation time as the later observation time. Compare the difference in salivary sedimentation degree between the historical observation time and the target observation time to obtain the historical salivary sedimentation difference value; Compare the difference in salivary sedimentation degree between the target observation time and the later observation time to obtain the later salivary sedimentation difference value; According to the historical salivary sedimentation difference value and the later salivary sedimentation difference value, obtain the abnormal dryness degree of saliva at the target observation time.

[0067] It should be especially noted that if the target observation time is the first observation time, take the later salivary sedimentation difference value of the target observation time as the abnormal dryness degree of saliva; if the target observation time is the last observation time, take the historical salivary sedimentation difference value of the target observation time as the abnormal dryness degree of saliva.

[0068] Further, as an example, the corrected salivary sedimentation can be calculated by the following formula:

[0069] ;

[0070] In the formula, represents the corrected salivary sedimentation of the sublingual microcirculation grayscale image at the th observation time; represents the abnormal dryness of saliva at the th observation time; represents the abnormal dryness of saliva at the th observation time; represents the mean value of the salivary sedimentation of all vascular structure regions in the sublingual microcirculation grayscale image at the th observation time; represents taking the absolute value;

[0071] Specifically, if there is no observation time before the th observation time, then is not considered when calculating ; if there is no observation time after the th observation time, then is not considered when calculating .

[0072] Thus, the corrected salivary sedimentation of the sublingual microcirculation grayscale image at different observation times is obtained by the above method.

[0073] Step S004: Perform prognostic auxiliary analysis on the sublingual microcirculation grayscale image according to the corrected salivary sedimentation.

[0074] Taking any observation time as an example, the corrected salivary sedimentation of the sublingual microcirculation grayscale image at this observation time is used as the grayscale weight, and the grayscale value after multiplying the grayscale weight by the grayscale value of each pixel in the sublingual microcirculation grayscale image at this observation time is used as the enhanced grayscale value of each pixel; the image composed of the enhanced grayscale values of all pixels at this observation time is used as the enhanced sublingual microcirculation image at this observation time.

[0075] Further, analyze the length of the vascular network and the perfusion quality in the enhanced sublingual microcirculation image, calculate the proportion of perfused blood vessels, and perform sepsis prognosis analysis. Among them, the process of calculating the proportion of perfused blood vessels according to the length of the vascular network and the perfusion quality for sepsis prognosis analysis is a well-known technology, and will not be elaborated in this embodiment.

[0076] Thus, this embodiment is completed.

[0077] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images, characterized in that: The method comprises the following steps: Obtain grayscale images of sublingual microcirculation at different observation times; For a grayscale image of sublingual microcirculation at any observation time, the structure of the sublingual blood vessels in the grayscale image of sublingual microcirculation is analyzed, and a blood vessel structure area is divided from the grayscale image of sublingual microcirculation, wherein the blood vessel structure area includes a plurality of blood vessel channel areas; based on the blood vessel channel area, the grayscale rise and fall law of similar blood vessel dilation is analyzed to obtain the saliva sedimentation degree of the blood vessel structure area; The flow and evaporation of sublingual saliva between adjacent observation moments were analyzed, and the saliva sedimentation degree was corrected to obtain the corrected saliva sedimentation degree of the sublingual microcirculation grayscale image at different observation moments; The grayscale images of sublingual microcirculation were analyzed for prognosis according to the corrected salivary sedimentation.

2. The sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images according to claim 1, characterized in that: The method for obtaining the vascular structure region is: The grayscale image of sublingual microcirculation obtains several vascular skeleton points through the trained neural network; the vascular skeleton points are expanded by window sliding to construct the vascular structure area.

3. The sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images according to claim 1, characterized in that: The method for obtaining the saliva sedimentation degree in the vascular structure area is: For any vascular structure area, the similarity of the expansion of the path of the vascular channel in the vascular structure area is analyzed, and the initial vascular swelling-proximate area is divided from the vascular structure area; the extension comparison of the initial vascular swelling-proximate area in the horizontal and vertical directions is analyzed to obtain the vascular inclination of the initial vascular swelling-proximate area; the grayscale interval difference of the vascular channel in the initial vascular swelling-proximate area is analyzed to obtain the vascular saliva sedimentation degree of the initial vascular swelling-proximate area; the vascular inclination and the vascular saliva sedimentation degree of the initial vascular swelling-proximate area are comprehensively considered to obtain the saliva sedimentation degree of the vascular structure area; the saliva sedimentation degree is negatively correlated with the vascular inclination.

4. The sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images according to claim 3, characterized in that: The method for obtaining the initial vascular swelling proximal region is: The channel widths of different vascular channel regions within the vascular structure region are obtained; the similarities of the channel widths between different vascular channel regions are analyzed, and the different vascular channel regions are integrated into regions similar to the initial vascular swelling.

5. The sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images according to claim 4, characterized in that: The method for obtaining the blood vessel inclination is: The horizontal region length in the initial blood vessel swelling proximal region is obtained; the vertical region length in the initial blood vessel swelling proximal region is obtained; and the blood vessel inclination in the initial blood vessel swelling proximal region is obtained based on the comparison difference between the horizontal region length and the vertical region length.

6. The sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images according to claim 4, characterized in that: The method for obtaining the vascular saliva sedimentation degree is: The grayscale differences between the longitudinal vascular channels in the area with similar initial vascular swelling were analyzed to obtain the vascular saliva sedimentation degree in the area with similar initial vascular swelling.

7. According to claim 1, the method for prognosis of sepsis based on intelligent auxiliary analysis of sublingual vascular images, after calculating the saliva sedimentation degree, further comprising: The salivary sedimentation was normalized.

8. The sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images according to claim 1, characterized in that: The method for obtaining the modified saliva sedimentation degree is: Taking any observation moment as the target observation moment, the difference in saliva sedimentation before and after the target observation moment is compared to obtain the abnormal saliva dryness at the target observation moment; based on the abnormal saliva dryness, the corrected saliva sedimentation of the sublingual microcirculation grayscale image at the target observation moment is obtained.

9. The sepsis prognosis method based on intelligent auxiliary analysis of sublingual vascular images according to claim 8, characterized in that: The method for obtaining the abnormal dryness of saliva is: The observation moment before the target observation moment is taken as the historical observation moment, and the observation moment after the target observation moment is taken as the later observation moment. The difference in saliva sedimentation degree between the historical observation moment and the target observation moment is compared to obtain the historical saliva sedimentation difference value; the difference in saliva sedimentation degree between the target observation moment and the later observation moment is compared to obtain the later saliva sedimentation difference value; according to the historical saliva sedimentation difference value and the later saliva sedimentation difference value, the abnormal dryness of saliva at the target observation moment is obtained.

10. According to claim 1, the method for prognosis of sepsis based on intelligent auxiliary analysis of sublingual vascular images, after calculating the corrected salivary sedimentation degree, further comprising: Normalization was performed on the corrected salivary sedimentation.

Images