An image detection method and system for gastrointestinal bleeding

Abstract

The present invention discloses an image detection method and system for gastrointestinal bleeding, which relates to the technical field of image detection. Based on the original data of the user to be detected and combined with the usage data of the capsule detection terminal in the data warehouse, the overall matching value between the user to be detected and each user is evaluated from two aspects of physical fitness and gastrointestinal peristalsis, so as to determine the set of detection body position and posture transformations of the user to be detected, with relatively high accuracy, ensuring the correctness of the posture of the user to be detected when using the capsule detection terminal for gastrointestinal bleeding examination, ensuring the smooth movement of the capsule detection terminal in the digestive tract, and improving the detection effect of the capsule detection terminal; based on the operation data of the capsule detection terminal of the user to be detected, it is judged whether there are problems such as image defects or too fast speed in the current capsule detection terminal, avoiding the influence on the recognition and determination of gastrointestinal bleeding of subsequent users to be detected, and ensuring the detection effect of the user to be detected.

Description

Technical Field

[0001] The present invention relates to the technical field of image detection, and particularly relates to an image detection method and system for gastrointestinal bleeding. Background Art

[0002] Gastrointestinal bleeding is a common and potentially life-threatening disease, and its diagnosis and treatment require promptness and precision. Image detection methods play a crucial role in the diagnosis of gastrointestinal bleeding. They can not only provide intuitive visual information but also help doctors accurately determine the location and cause of bleeding, thereby formulating effective treatment plans. Capsule detection terminals are increasingly widely used in the detection of gastrointestinal bleeding due to their small size and non-invasive characteristics. However, improper use of capsule detection terminals not only affects the detection efficiency of patients but also causes waste of medical resources. Therefore, it is extremely necessary to analyze the use of capsule detection terminals.

[0003] The prior art, such as an invention patent application with publication number CN106373137B, discloses a method for detecting gastrointestinal bleeding images for a capsule endoscope, which includes the following steps: Step 1: After a patient swallows a capsule endoscope, the capsule endoscope collects images in the gastrointestinal tract and sends the images to a computer through wireless communication, and the computer performs data preprocessing on the gastrointestinal tract images. Step 2: Perform dot bleeding detection and surface bleeding detection on the preprocessed gastrointestinal tract images respectively to determine whether there is bleeding in the gastrointestinal tract, and mark the dot bleeding area and the surface bleeding area. This invention can effectively identify image bleeding and improve the working efficiency of doctors.

[0004] The prior art, such as an invention patent application with publication number CN115251806B, discloses a medical device monitoring, analysis and regulation system based on magnetic field principle technology, including: a module for analyzing the rationality of wearing a target person's examination gown, a module for the autonomous operation of a capsule detection terminal, a module for analyzing the rationality of the capsule residence time, a module for analyzing the electrical safety of the capsule detection terminal, a database, and a warning terminal. This invention classifies the collected stomach pictures according to their respective structures, and then analyzes the set of stomach pictures of each component structure of the stomach, effectively ensuring the accuracy of the analysis of the stomach damage of medical personnel. This invention performs secondary image acquisition on the damaged part of the stomach, reducing the deviation rate of the judgment result, thereby ensuring the accuracy of the judgment of the stomach condition of medical personnel. This invention not only analyzes the residence time of the capsule detection terminal in the body of medical personnel but also analyzes the electrical safety of the capsule detection terminal, and the analysis results are relatively comprehensive, which is conducive to ensuring the safety of medical personnel.

[0005] Based on the above technical solutions, it is found that there are still certain defects in the use and analysis of capsule detection terminals in the prior art, which are specifically reflected in the following aspects: on the one hand, in the prior art, the attention to the change of the detection body position and posture of the user to be detected is not high. Due to the differences in physical qualities and intestines of the users to be detected, there are certain differences in the acceptance of the detection body position and posture. The prior art lacks the analysis of this aspect, making it difficult to ensure the correctness of the posture of the user to be detected when using the capsule detection terminal for gastrointestinal bleeding examination, which in turn affects the movement of the capsule detection terminal in the digestive tract and reduces the detection effect of the capsule detection terminal. On the other hand, there is rarely an early warning for abnormalities during the use of the capsule detection terminal. When the capsule detection terminal is in use, problems such as defective transmitted images or excessive speed are likely to occur. If no timely early warning is given and the responsible doctor intervenes, it will affect the identification and determination of gastrointestinal bleeding of subsequent users to be detected, and it is difficult to ensure the detection effect of the users to be detected. Summary of the Invention

[0006] The purpose of the present invention is to provide an image detection method and system for gastrointestinal bleeding, which solves the problems existing in the background technology.

[0007] To solve the above technical problems, the present invention adopts the following technical solutions: In the first aspect of the present invention, an image detection method for gastrointestinal bleeding is provided, including ST1. Obtain the physical sign data of the user to be detected from the hospital operation platform, and determine whether the user to be detected needs to perform target gastrointestinal bleeding detection. If so, execute ST2.

[0008] ST2. Obtain the original data of the user to be detected from the hospital operation platform, and combine the capsule detection terminal usage data in the data warehouse to determine the set of detection body position and posture changes and the set of key detection time points of the user to be detected.

[0009] ST3. After the user to be detected takes in the capsule detection terminal, conduct voice broadcast for adjusting the body position and posture of the user to be detected, and obtain the operation data of the capsule detection terminal of the user to be detected.

[0010] ST4. Based on the operation data of the capsule detection terminal of the user to be detected, give an early warning to the responsible doctor of the user to be detected.

[0011] ST5. Extract the image set of the capsule detection terminal of the user to be detected, evaluate each bleeding area and the risk value of each bleeding area in the area to be detected, and display them.

[0012] The second aspect of the present invention provides a system for executing the image detection method as described in the present invention, including: a target gastrointestinal bleeding detection and judgment module, configured to obtain the physical sign data of the user to be detected from the hospital operation platform, determine whether the user to be detected needs to undergo target gastrointestinal bleeding detection, and if so, execute the detection parameter determination module for the user to be detected.

[0013] The detection parameter determination module for the user to be detected is configured to obtain the original data of the user to be detected from the hospital operation platform, and combine it with the usage data of the capsule detection terminal in the data warehouse to determine the set of detection body position and posture transformations and the set of key detection time points for the user to be detected.

[0014] The detection module is configured to, after the user to be detected swallows the capsule detection terminal, adjust the voice broadcast of the body position and posture of the user to be detected, and obtain the operation data of the capsule detection terminal of the user to be detected.

[0015] The early warning terminal is configured to give an early warning to the attending doctor of the user to be detected based on the operation data of the capsule detection terminal of the user to be detected.

[0016] The display terminal is configured to extract the image set of the capsule detection terminal of the user to be detected, evaluate the bleeding areas in the area to be detected and the risk values of each bleeding area, and display them.

[0017] The beneficial effects of the present invention are as follows: In ST1, based on the physical sign data of the user to be detected, the present invention determines whether the user to be detected needs to use the capsule detection terminal for gastrointestinal bleeding detection, laying a foundation for the subsequent analysis of the detection body position and posture of the user to be detected.

[0018] In ST2, based on the original data of the user to be detected and combined with the usage data of the capsule detection terminal in the data warehouse, the present invention evaluates the overall matching value between the user to be detected and each user from two levels of physical quality and gastrointestinal peristalsis, so as to determine the set of detection body position and posture transformations of the user to be detected, with relatively high accuracy, making up for the defect of neglecting this level in the prior art, ensuring the correctness of the posture of the user to be detected when using the capsule detection terminal for gastrointestinal bleeding examination, ensuring the smooth movement of the capsule detection terminal in the digestive tract, and improving the detection effect of the capsule detection terminal.

[0019] In ST5, based on the operation data of the capsule detection terminal of the user to be detected, the present invention determines whether there are problems with image defects or excessive speed in the current capsule detection terminal, and gives an early warning to the attending doctor of the user to be detected, avoiding the impact on the recognition and determination of gastrointestinal bleeding of the subsequent user to be detected, and ensuring the detection effect of the user to be detected. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] 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 drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0021] Figure 1 It is a schematic flow chart of the implementation steps of the method of the present invention.

[0022] Figure 2 It is a schematic connection diagram of the system structure of the present invention. Detailed implementation manners

[0023] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0024] Refer to Figure 1 As shown, a method for image detection of gastrointestinal bleeding provided by the first aspect of the present invention includes: ST1. Obtain the physical sign data of the user to be detected from the hospital operation platform, determine whether the user to be detected needs to perform target gastrointestinal bleeding detection. If so, execute ST2.

[0025] It should be noted that the target gastrointestinal bleeding detection is to perform gastrointestinal bleeding detection using a capsule detection terminal.

[0026] As a preferred solution, the physical sign data includes a set of symptom description texts, blood pressure, heart rate, respiratory rate, a set of facial images, and characteristic values of each inspection index.

[0027] It should be noted that the characteristic values of each inspection index include hemoglobin concentration, red blood cell count, and hematocrit in blood routine examination, including prothrombin time, activated partial thromboplastin time, fibrinogen, etc. in coagulation function examination, and also include blood urea nitrogen concentration in blood urea nitrogen examination.

[0028] It should also be noted that the facial image set is specifically extracted from the monitoring of the hospital operation platform.

[0029] The original data includes characteristic values of each basic data, the activity base frequency of the gastrointestinal tract, esophageal distribution pressure map, anorectal distribution pressure map, excrement sample, small intestine transit rate, activity level characteristic value, and historical pathological data.

[0030] It should be noted that the characteristic values of the basic data include height, age, body fat percentage, etc. The specific fundamental frequency of gastrointestinal tract activity refers to the frequency of auscultating bowel sounds by palpating the abdomen. The esophageal distribution pressure map and anorectal distribution pressure map are specifically obtained by measuring the pressure of each segment of the digestive tract through esophageal manometry and anorectal manometry, and then drawing the esophageal distribution pressure map and anorectal distribution pressure map. The small intestine transit rate is specifically inferred by monitoring the change trend of hydrogen level in the user's breath over time. The activity level characteristic value specifically refers to the user's daily exercise amount. The activity level characteristic value is a value between 0 and 1, and the larger the activity level characteristic value, the greater the user's daily exercise amount, which is determined and uploaded by the user's attending physician.

[0031] It should also be noted that the small intestine transit rate is inferred based on the change trend of hydrogen level in the user's breath over time. According to the time point and peak value of the hydrogen concentration increase, the transmission rate of the small intestine can be inferred. Generally, the time point when the hydrogen concentration significantly increases reflects the time when the marker reaches the colon, thus indirectly reflecting the transmission rate of the small intestine. This method is a commonly used means in the prior art and will not be elaborated here.

[0032] In the ST1 of the present invention, based on the physical sign data of the user to be detected, it is determined whether the user to be detected needs to use a capsule detection terminal for gastrointestinal bleeding detection, laying a foundation for the subsequent analysis of the detection body position and posture of the user to be detected.

[0033] ST2. Obtain the original data of the user to be detected from the hospital operation platform, and combine the data of the capsule detection terminal usage in the data warehouse to determine the set of detection body position and posture transformations and the set of key detection time points of the user to be detected.

[0034] As a preferred solution, the specific determination method for determining whether the user to be detected needs to perform target gastrointestinal bleeding detection is as follows: Input the symptom description text set A, blood pressure X, heart rate P, and respiratory rate H in the physical sign data of the user to be detected into the first abnormal evaluation model to output the first abnormal value of the user to be detected, where A’ is the set of gastrointestinal bleeding symptom description texts preset in the data warehouse, and β m is the m-th vital sign data, β m =(X, P, H), β′ m is the requirement interval of the m-th vital sign data, β′ m =(XI′, PI′, HI′), and XI’, PI’, and HI’ respectively represent the required blood pressure interval, required heart rate interval, and required respiratory rate interval corresponding to the target gastrointestinal bleeding detection preset in the data warehouse.

[0035] Based on the facial image set of the user to be detected, evaluate the second abnormal value α of the user to be detected_2 。

[0036] Input the eigenvalue δ of each inspection index of the physical sign data of the user to be detected p into the third anomaly assessment model to output the third anomaly value of the user to be detected, where δ p ′ is the required eigenvalue interval corresponding to the p-th inspection index for the preset target gastrointestinal bleeding detection in the data warehouse, p is the number of each inspection index, p = 1, 2,..., q, and q is any integer greater than 2.

[0037] Import the first anomaly value, the second anomaly value, and the third anomaly value of the user to be detected into the target gastrointestinal bleeding detection and judgment model where α′ _1 、α′ _2 、α′ _3 respectively represent the preset first anomaly convergence value, the second anomaly convergence value, and the third anomaly convergence value, ∨ and ∧ respectively represent the logical symbols or and and, and output the target gastrointestinal bleeding detection and judgment value of the user to be detected. The target gastrointestinal bleeding detection and judgment value includes numerical values of 1 and -1. When the target gastrointestinal bleeding detection and judgment value is 1, it indicates that the user to be detected needs to undergo the target gastrointestinal bleeding detection. When the target gastrointestinal bleeding detection and judgment value is -1, it indicates that the user to be detected does not need to undergo the target gastrointestinal bleeding detection.

[0038] It should be noted that the present invention determines whether the user to be detected is suitable for detecting gastrointestinal bleeding in the form of a capsule detection terminal. On the one hand, it avoids detecting with a capsule detection terminal when the risk of gastrointestinal bleeding is relatively high, which may affect the timely treatment of the user to be detected. On the other hand, it alleviates the pressure of insufficient medical resources in the hospital.

[0039] As a preferred solution, for the evaluation of the second anomaly value of the user to be detected, the specific evaluation method is as follows: Based on the set of facial images of the user to be detected, identify the chromaticity values ST of each key part of several facial images through image recognition technology ij , and identify the temperature values SI of several facial images i , i is the number of each facial image, i = 1, 2,..., n, n is any integer greater than 2, and j is the number of each key part, j = 1, 2,..., k, and k is any integer greater than 2.

[0040] It should be noted that the specific method for identifying the temperature values of several facial images is as follows: By using image recognition technology to identify the chromaticity values of several facial images, obtaining the average chromaticity value of several facial images, and comparing it with the chromaticity value intervals corresponding to each preset temperature value in the data warehouse. If the average chromaticity value of a certain facial image is within the chromaticity value interval corresponding to a certain temperature value, then take this temperature value as the temperature value of this facial image, and obtain the temperature values of several facial images.

[0041] Evaluate the second outlier of the user to be detected Where SI′ and ST j ′ respectively represent the preset facial temperature value of gastrointestinal bleeding and the reference chromaticity value of the jth key part, k is the number of key parts, is the facial outlier of the ith facial image of the user to be detected.

[0042] It should be noted that when there is gastrointestinal bleeding, the facial skin temperature of the patient usually drops, and the skin of the key parts usually becomes pale. Therefore, it is necessary to analyze the facial skin temperature and the chromaticity value of the key parts of the patient, and in order to reduce the misjudgment rate, analyze the risk change rate of the patient.

[0043] As a preferred solution, the specific method for determining the detection body position posture transformation set and the key detection time point set of the user to be detected is as follows: Extract the characteristic values of each basic data, activity level characteristic values, and historical pathological data from the original data of the user to be detected, and obtain the capsule detection terminal usage data from the data warehouse. The capsule detection terminal usage data includes the original data of each user, the detection body position posture transformation set, and the key detection time point set, and extract the characteristic values of each basic data, activity level characteristic values, and historical pathological data of each user, and evaluate the physical fitness matching value between the user to be detected and each user.

[0044] Extract the activity fundamental frequency of the gastrointestinal tract, esophageal distribution pressure map, anorectal distribution pressure map, excrement sample, and small intestine transit rate from the original data of the user to be detected, and extract the activity fundamental frequency of the gastrointestinal tract, esophageal distribution pressure map, anorectal distribution pressure map, excrement sample, and small intestine transit rate from the original data of each user, and evaluate the gastrointestinal peristalsis matching value between the user to be detected and each user.

[0045] Import the physical fitness matching value and gastrointestinal peristalsis matching value between the user to be detected and each user into the overall matching value evaluation model where μ f is the physical fitness matching value between the user to be detected and the fth user, Let \(GI\) be the gastrointestinal motility matching value of the physical fitness matching value between the user to be detected and the \(f\)-th user, where \(f\) is the number of each user, \(f = 1, 2, \cdots, t\), and \(t\) is any integer greater than 2. \(\lambda_1\) and \(\lambda_2\) respectively represent the weight influence factors corresponding to the preset physical fitness matching and gastrointestinal motility matching. Output the overall matching value between the user to be detected and each user. If the overall matching value between the user to be detected and a certain user is the largest, then the detection posture transformation set and the key detection time point set of this user are respectively used as the detection posture transformation set and the key detection time point set of the user to be detected.

[0046] It should be noted that the preset weight influence factors corresponding to the physical fitness matching and gastrointestinal motility matching are specifically set by medical experts according to specific situations. For example, if clinical evidence shows that the gastrointestinal motility matching has a greater impact on the detection posture, then the weight influence factor corresponding to the gastrointestinal motility matching is set to be greater than the weight influence factor of the physical fitness matching.

[0047] As a preferred solution, the method for specifically evaluating the physical fitness matching value between the user to be detected and each user is as follows: Based on the historical pathological data of the user to be detected, extract each disease type and the emotion set \(W\) of each examination b and the characteristic values \(QI\) of each physiological data bx , construct the historical disease type set of the user to be detected, and evaluate the risk characteristic value of the examination emotion of the user to be detected where \(W'\) and \(QI\) x ' respectively represent the preset risk emotion set and the safety characteristic value interval of the \(x\)-th physiological data. \(b\) is the number of each examination, \(b = 1, 2, \cdots, d\), and \(d\) is any integer greater than 2. \(x\) is the number of each physiological data, \(x = 1, 2, \cdots, y\), and \(y\) is any integer greater than 2. Similarly, construct the historical disease type set of the user to be detected, and evaluate the risk characteristic value \(T\) of the examination of each user f '.

[0048] It should be noted that the risk emotion set includes panic, anxiety, depression, etc.

[0049] Specifically, the emotion set is specifically constructed by collecting the facial video of the user and then identifying various emotions of the user.

[0050] Let \(\eta\) be the characteristic value of each basic data of the user to be detected h , the activity level characteristic value \(T\), the historical disease type set \(E\), the risk characteristic value \(F\) of the examination emotion, and the characteristic value \(\eta'\) of each basic data of each user fh , the activity level characteristic value \(T\) f ', the historical disease type set \(E'\) f, check the emotional risk characteristic value F f ′ Import it into the physical literacy matching value evaluation model to output the physical literacy matching values of the user to be detected and each user. h is the number of each basic data, h = 1, 2,..., g, and g is any integer greater than 2.

[0051] As a preferred solution, the method for specifically calculating the gastrointestinal motility matching value of the user to be detected and each user is as follows: Based on the esophageal distribution pressure map, anorectal distribution pressure map, excrement sample of the user to be detected, and the esophageal distribution pressure map, anorectal distribution pressure map, excrement sample of each user, evaluate the internal distribution pressure matching degree N between the user to be detected and each user f , intestinal flora matching degree U f .

[0052] It should be noted that the specific evaluation method for the internal distribution pressure matching degree of the user to be detected and each user is as follows: Use image processing software to output the similarity between the esophageal distribution pressure map of the user to be detected and the esophageal distribution pressure maps of each user. Similarly, output the similarity between the anorectal distribution pressure map of the user to be detected and the anorectal distribution pressure maps of each user, and perform mean processing on the similarity between the esophageal distribution pressure map of the user to be detected and the anorectal distribution pressure maps of each user to obtain the internal distribution pressure matching degree between the user to be detected and each user.

[0053] It should also be noted that the specific evaluation method for the intestinal flora matching degree of the user to be detected and each user is as follows: Based on the excrement sample of the user to be detected, obtain each characteristic parameter of the excrement of the user to be detected, and based on the excrement samples of each user, obtain each characteristic parameter of the excrement of each user. The specific characteristic parameters of the excrement of each user were successfully recorded and uploaded to the data warehouse during historical use. By evaluating the similarity between the characteristic parameters of the excrement of the user to be detected and the characteristic parameters of the excrement of each user, and performing mean processing on them, the intestinal flora matching degree between the user to be detected and each user is obtained. The evaluation of the similarity between the characteristic parameters of the excrement of the user to be detected and the characteristic parameters of the excrement of each user can be obtained through numerical similarity evaluation in the prior art, which will not be elaborated here.

[0054] Once again, it should be noted that the characteristic parameters include chromaticity value, hardness, pH value, etc.

[0055] Evaluate the gastrointestinal motility matching value of the user to be detected and each user where HJ and V are respectively the activity fundamental frequency and small intestine transit rate of the gastrointestinal tract of the user to be detected, HJ′ f , V ff' are the fundamental frequency of gastrointestinal motility and the small intestine transit rate of the f-th user, respectively.

[0056] Based on the original data of the user to be detected in ST2, combined with the data of the capsule detection terminals used in the data warehouse, the present invention evaluates the overall matching value between the user to be detected and each user from two aspects of physical literacy and gastrointestinal motility, so as to determine the set of posture transformation for the user to be detected. The accuracy is relatively high, which makes up for the defect of neglecting this aspect in the prior art, ensures the correctness of the posture of the user to be detected when using the capsule detection terminal for gastrointestinal bleeding examination, ensures the smooth movement of the capsule detection terminal in the digestive tract, and improves the detection effect of the capsule detection terminal.

[0057] ST3. After the user to be detected takes in the capsule detection terminal, conduct voice broadcast for adjusting the body posture of the user to be detected, and obtain the operation data of the capsule detection terminal of the user to be detected.

[0058] The operation data includes the current output image and the current shooting duration.

[0059] ST4. Based on the operation data of the capsule detection terminal of the user to be detected, give an early warning to the responsible doctor of the user to be detected.

[0060] As a preferred solution, the method for giving an early warning to the responsible doctor of the user to be detected is as follows: obtain the current output image and the current shooting duration of the capsule detection terminal of the user to be detected from the operation data of the capsule detection terminal of the user to be detected, and identify the defect existence state of the current output image of the capsule detection terminal of the user to be detected. The defect existence state includes existence and non-existence.

[0061] It should be noted that the current shooting duration of the capsule detection terminal of the user to be detected is specifically the interval duration between the time point of the current output image and the time point of the previous output image.

[0062] It should also be noted that the method for identifying the defect existence state of the current output image of the capsule detection terminal of the user to be detected is to specifically identify the defects of the current output image of the capsule detection terminal of the user to be detected through image recognition technology. If there are defects, record the defect existence state as existence; otherwise, record the defect existence state as non-existence. The defects include blurring, noise, and low contrast, etc.

[0063] If the defect existence state of the current output image of the capsule detection terminal of the user to be detected is existence, give an early warning of abnormal image quality to the responsible doctor of the user to be detected. If the current shooting duration of the capsule detection terminal of the user to be detected is less than the preset shooting interval duration in the data warehouse, give an early warning of abnormal shooting speed to the responsible doctor of the user to be detected.

[0064] Based on the running data of the capsule detection terminal of the user to be detected in ST5, the present invention determines whether there are problems such as image defects or excessive speed in the current capsule detection terminal, and gives early warnings to the responsible doctors of the users to be detected, so as to avoid affecting the identification and determination of gastrointestinal bleeding of subsequent users to be detected, and ensure the detection effect of the users to be detected.

[0065] ST5. Extract the image set of the capsule detection terminal of the user to be detected, evaluate each bleeding area and the risk value of each bleeding area in the area to be detected, and display them.

[0066] As a preferred solution, the evaluation of each bleeding area and the risk value of each bleeding area of the user to be detected is specifically evaluated as follows: Extract several images from the image set of the capsule detection terminal of the user to be detected, and perform grid division on them to obtain each sub-image of each image, obtain the chromaticity value of each sub-image of each image, and then compare it with the preset bleeding chromaticity value interval in the data warehouse. If the chromaticity value of a certain sub-image is within the bleeding chromaticity value interval, then mark the sub-image as a risk sub-image, screen out each risk sub-image of each image, and count the number of risk sub-images and the number of sub-images of each image of the capsule detection terminal of the user to be detected.

[0067] If the number of risk sub-images of each image of the capsule detection terminal of the user to be detected is 0, then mark the image as a normal image; otherwise, mark the image as a bleeding image, mark the detection area corresponding to the image as a bleeding area, divide the number of risk sub-images of the image by the number of sub-images, and obtain the risk value of the bleeding area, so as to obtain each bleeding area and the risk value of each bleeding area of the user to be detected.

[0068] It should be noted that the detection area corresponding to the image is specifically: during the operation of the capsule detection terminal, the actual detection area of the output image can be located through the locator of the capsule detection terminal.

[0069] Refer to Figure 2 As shown, the second aspect of the present invention provides a system for executing the image detection method as described in the present invention, including:

[0070] The target gastrointestinal bleeding detection and judgment module is used to obtain the physical sign data of the user to be detected from the hospital operation platform, determine whether the user to be detected needs to perform target gastrointestinal bleeding detection, and if so, execute the detection parameter determination module for the user to be detected.

[0071] The detection parameter determination module for the user to be detected is used to obtain the original data of the user to be detected from the hospital operation platform, and combine the capsule detection terminal usage data in the data warehouse to determine the detection body position posture transformation set and the key detection time point set of the user to be detected.

[0072] A detection module, configured to, after a user to be detected puts on a capsule detection terminal, perform voice broadcast for adjusting the body posture of the user to be detected, and obtain the operation data of the capsule detection terminal of the user to be detected.

[0073] An early warning terminal, configured to give an early warning to the attending doctor of the user to be detected based on the operation data of the capsule detection terminal of the user to be detected.

[0074] A display terminal, configured to extract an image set of the capsule detection terminal of the user to be detected, evaluate each bleeding area in the area to be detected and the risk value of each bleeding area, and display them.

[0075] It should be noted that the present invention further includes a data warehouse, which is used to store a text set of descriptions of digestive tract bleeding symptoms, the required blood pressure range, required heart rate range, and required respiratory rate range corresponding to the target digestive tract bleeding detection, the required characteristic value range of each inspection index corresponding to the target digestive tract bleeding detection, the usage data of the capsule detection terminal, the shooting interval duration, and the bleeding chromaticity value range.

[0076] It should also be noted that the target digestive tract bleeding detection and judgment module is connected to the detection parameter determination module of the user to be detected, the detection parameter determination module of the user to be detected is connected to the detection module, the detection module is respectively connected to the early warning terminal and the display terminal, and the data warehouse is respectively connected to the target digestive tract bleeding detection and judgment module, the detection parameter determination module of the user to be detected, and the early warning terminal.

[0077] The above content is only an example and illustration of the concept of the present invention. Those skilled in the art of this technology can make various modifications, supplements, or use similar methods to replace the specific embodiments described, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. An image detection method for gastrointestinal bleeding, characterized in that: include: ST1. Obtain the vital sign data of the user to be tested from the hospital operation platform, determine whether the user to be tested needs to undergo target gastrointestinal bleeding testing, and if necessary, execute ST2; The physical sign data includes a symptom description text set, blood pressure, heart rate, respiratory rate, facial image set and characteristic values ​​of each examination index; The specific method for determining whether the user to be detected needs to undergo target digestive tract bleeding detection is as follows: The symptom description text set A, blood pressure X, heart rate P and respiratory rate H in the physical sign data of the user to be detected are input into the first abnormality assessment model , output the first abnormal value of the user to be detected, where A' is the text set of gastrointestinal bleeding symptom description preset in the data warehouse, β m is the mth vital sign data, β m =(X,P,H),β′ m is the required interval of the mth vital sign data, β′ m =(XI′,PI′,HI′), XI′, PI′, HI′ respectively represent the required blood pressure interval, required heart rate interval and required respiratory rate interval corresponding to the target gastrointestinal bleeding detection preset in the data bin; Based on the facial image set of the user to be detected, the second outlier value α of the user to be detected is evaluated _2 ; The characteristic value δ of each inspection index of the vital sign data of the user to be detected p Input to the third anomaly assessment model , output the third abnormal value of the user to be detected, where δ p ′ is the required characteristic value interval of the pth examination index corresponding to the target gastrointestinal bleeding detection preset in the data warehouse, p is the number of each examination index, p=1,2,...,q, q is any integer greater than 2; Import the first abnormal value, the second abnormal value and the third abnormal value of the user to be detected into the target gastrointestinal bleeding detection and judgment model where α′ _1 , α′ _2 , α′ _3 They are respectively represented as the preset first abnormal convergence value, the second abnormal convergence value and the third abnormal convergence value, ∨ and ∧ are respectively represented as the logical symbols or and, and the target gastrointestinal bleeding detection judgment value of the user to be detected is output, and the target gastrointestinal bleeding detection judgment value includes the values ​​of 1 and -1. When the target gastrointestinal bleeding detection judgment value is 1, it indicates that the user to be detected needs to perform the target gastrointestinal bleeding detection, and when the target gastrointestinal bleeding detection judgment value is -1, it indicates that the user to be detected does not need to perform the target gastrointestinal bleeding detection; The specific evaluation method of evaluating the second abnormal value of the user to be detected is: Based on the facial image set of the user to be detected, the chromaticity values ​​of each key part of several facial images are identified by image recognition technology. ij , and identify the temperature values ​​SI of several facial images i , i is the number of each facial image, i=1,2,...,n, n is any integer greater than 2, j is the number of each key part, j=1,2,...,k, k is any integer greater than 2; Evaluate the second outlier value of the user to be detected Among them, SI′, ST j ′ represents the preset facial temperature value of gastrointestinal bleeding and the reference chromaticity value of the jth key part, k is the number of key parts, is the facial outlier value of the i-th facial image of the user to be detected; ST2. Obtain the original data of the user to be tested from the hospital operation platform, combine the capsule detection terminal usage data in the data warehouse, and determine the detection posture transformation set and key detection time point set of the user to be tested; The source data include characteristic values ​​of basic data, gastrointestinal activity fundamental frequency, esophageal distribution pressure map, anorectal distribution pressure map, fecal samples, small intestinal transit rate, activity level characteristic value, historical pathological data; ST3. After the user to be tested is served into the capsule detection terminal, the posture of the user to be tested is adjusted by voice broadcast, and the operation data of the capsule detection terminal of the user to be tested is obtained; ST4. Based on the operation data of the capsule detection terminal of the user to be detected, the responsible doctor of the user to be detected is warned; ST5. Extract the image set of the capsule detection terminal of the user to be detected, evaluate and obtain the bleeding areas and risk values ​​of the bleeding areas in the area to be detected, and display them.

2. The image detection method for gastrointestinal bleeding according to claim 1, characterized in that: The specific determination method of determining the detection posture transformation set and key detection time point set of the user to be detected is: Extract the characteristic values ​​of each basic data, the characteristic values ​​of the activity level and the historical pathological data from the original data of the user to be detected, obtain the capsule detection terminal usage data from the data warehouse, wherein the capsule detection terminal usage data includes the original data of each user, the detection posture change set and the key detection time point set, and extract the characteristic values ​​of each basic data, the characteristic values ​​of the activity level and the historical pathological data of each user, and evaluate the physical fitness matching value between the user to be detected and each user; Extracting the gastrointestinal activity fundamental frequency, esophageal distribution pressure map, anorectal distribution pressure map, fecal sample and small intestinal transit rate from the original data of the user to be tested, and extracting the gastrointestinal activity fundamental frequency, esophageal distribution pressure map, anorectal distribution pressure map, fecal sample and small intestinal transit rate from the original data of each user, and evaluating the gastrointestinal motility matching value of the user to be tested and each user; Import the physical fitness matching value and gastrointestinal motility matching value of the user to be tested and each user into the overall matching value evaluation model μ f is the physical fitness matching value between the user to be detected and the fth user, is the physical fitness matching value and gastrointestinal motility matching value of the user to be detected and the f-th user, f is the number of each user, f=1,2,...,t, t is any integer greater than 2, λ1 and λ2 represent the weight influence factors corresponding to the preset physical fitness matching and gastrointestinal motility matching respectively, output the overall matching value of the user to be detected and each user, and screen out the detection posture transformation set and key detection time point set of the user to be detected.

3. The image detection method for gastrointestinal bleeding according to claim 2, characterized in that: The specific evaluation method of evaluating the physical fitness matching value of the user to be detected and each user is as follows: Extract the emotion set W for each disease type and each examination based on the historical pathological data of the user to be detected b , the characteristic value QI of each physiological data bx , construct a historical disease type set of the user to be tested, and evaluate the risk feature value of the user's examination emotion Where W′, QI x ′ represents the preset risk emotion set and the safety characteristic value interval of the xth physiological data, b is the number of each examination, b=1,2,...,d, d is any integer greater than 2, x is the number of each physiological data, x=1,2,...,y, y is any integer greater than 2, and similarly, the historical disease type set of the user to be tested is constructed, and the examination risk characteristic value T of each user is evaluated. f ′; The characteristic value η of each basic data of the user to be detected h , activity level characteristic value T, historical disease type set E, check emotional risk characteristic value F and characteristic value η′ of each user's basic data fh , activity level characteristic value T f ′ , historical disease type set E′ f , check the emotional risk characteristic value F f 'Imported into the physical literacy matching value assessment model In the output, the physical fitness matching value of the user to be detected and each user is output, h is the number of each basic data, h = 1, 2, ..., g, g is any integer greater than 2.

4. The image detection method for gastrointestinal bleeding according to claim 2, characterized in that: The specific calculation method of evaluating the gastrointestinal motility matching value between the user to be detected and each user is as follows: Based on the esophageal distribution pressure map, anorectal distribution pressure map, excrement sample of the user to be tested and the esophageal distribution pressure map, anorectal distribution pressure map, excrement sample of each user, evaluate the internal distribution pressure matching degree N of the user to be tested and each user f , intestinal flora matching degree U f ; Evaluate the gastrointestinal motility matching value between the user to be tested and each user Where HJ and V are the activity fundamental frequency and small intestinal transit rate of the gastrointestinal tract of the user to be tested, respectively, and HJ′ f 、V f ′ are the gastrointestinal activity fundamental frequency and small intestinal transit rate of the fth user respectively.

5. The image detection method for gastrointestinal bleeding according to claim 1, characterized in that: The specific method of giving an early warning to the responsible doctor of the user to be tested is as follows: Acquire a current output image and a current shooting duration of the capsule detection terminal of the user to be detected from the operation data of the capsule detection terminal of the user to be detected, and identify a defect existence state of the current output image of the capsule detection terminal of the user to be detected, wherein the defect existence state includes existence and non-existence; If the defect status of the current output image of the capsule detection terminal of the user to be detected is present, the responsible doctor of the user to be detected will be warned of abnormal image quality; if the current shooting time of the capsule detection terminal of the user to be detected is less than the shooting interval preset in the data warehouse, the responsible doctor of the user to be detected will be warned of abnormal shooting speed.

6. The image detection method for digestive tract bleeding according to claim 1, characterized in that: The evaluation obtains each bleeding area of ​​the user to be detected and the risk value of each bleeding area, and the specific evaluation method is: Extract several images from the image set of the capsule detection terminal of the user to be detected, divide them into grids, obtain sub-images of each image, obtain the chromaticity value of each sub-image of each image, and then compare it with the bleeding chromaticity value interval preset in the data warehouse. If the chromaticity value of a sub-image is within the bleeding chromaticity value interval, the sub-image is recorded as a risk sub-image, and the risk sub-images of each image are obtained by screening, and the number of risk sub-images and the number of sub-images of each image of the capsule detection terminal of the user to be detected are counted; If the number of risk sub-images of each image of the capsule detection terminal of the user to be detected is 0, the image is recorded as a normal image. Otherwise, the image is recorded as a bleeding image, and the detection area corresponding to the image is recorded as a bleeding area. The number of risk sub-images of the image is divided by the number of sub-images to obtain the risk value of the bleeding area, thereby obtaining each bleeding area of ​​the user to be detected and the risk value of each bleeding area.

7. A system for executing the image detection method according to any one of claims 1 to 6, characterized in that: include: The target gastrointestinal bleeding detection judgment module is used to obtain the vital sign data of the user to be detected from the hospital operation platform, determine whether the user to be detected needs to undergo the target gastrointestinal bleeding detection, and if necessary, execute the detection parameter determination module of the user to be detected; The module for determining the detection parameters of the user to be detected is used to obtain the original data of the user to be detected from the hospital operation platform, and determine the detection posture transformation set and key detection time point set of the user to be detected in combination with the capsule detection terminal usage data in the data warehouse; The detection module is used to adjust the posture of the user to be detected after the user to be detected takes the capsule detection terminal, and to obtain the operation data of the capsule detection terminal of the user to be detected; An early warning terminal is used to warn the doctor responsible for the user to be tested based on the operation data of the capsule testing terminal of the user to be tested; The display terminal is used to extract an image set of a capsule detection terminal of a user to be detected, evaluate and obtain each bleeding area and the risk value of each bleeding area in the area to be detected, and display them.

Images