System and method for evaluating and detecting intestinal preparation of elderly patients before enteroscopy

Abstract

The invention relates to the field of medical health management, in particular to a system and a method for evaluating and detecting intestinal preparation of an elderly patient before enteroscopy, and the system comprises an image acquisition module, a feature extraction module, a scheme generation module and a quality evaluation module; the image acquisition module is used for acquiring a first image and a second image; the feature extraction module is used for obtaining a first feature and a second feature through the first image and the second image respectively, the first feature is an excrement shape feature before the monitoring target executes the intestinal preparation scheme, and the second feature is an excrement shape feature after the monitoring target executes the intestinal preparation scheme; a scheme generation module generates an intestinal preparation scheme based on the first feature; the quality evaluation module evaluates the intestinal preparation quality of the monitoring target according to the first feature and the second feature. Through intelligent evaluation and personalized intervention, the intestinal preparation quality of the elderly patient can be improved, and the success rate of colonoscopy can be improved.

Description

Technical Field

[0001] The present invention relates to the field of medical health management, and in particular to a system and method for evaluating and detecting intestinal preparation of elderly patients before colonoscopy. Background Art

[0002] As an important means of diagnosing and preventing colorectal cancer, colonoscopy is becoming more and more common among the elderly. The incidence of colorectal cancer increases significantly with age, and early screening is crucial to improving cure rates and reducing mortality. However, due to the special physiological conditions of the elderly, such as weakened intestinal function, decreased drug metabolism, and slower gastrointestinal motility, they face greater challenges in bowel preparation before undergoing colonoscopy.

[0003] Traditional bowel preparation programs usually rely on the use of strong laxatives, requiring patients to drink a lot of water and take medicine in a short period of time to empty the intestines before the examination. However, this type of method may cause some problems for the elderly: First, poor drug absorption leads to decreased excretion efficiency, which in turn affects the cleanliness of the intestines. Secondly, due to physical weakness or problems such as heart and kidney dysfunction, the elderly may not be able to withstand the side effects of large amounts of laxatives, such as dehydration and electrolyte imbalance. These side effects not only increase the risks of colonoscopy, but may also cause serious health problems.

[0004] In addition, the exercise capacity of elderly patients generally decreases, and exercise plays an important role in maintaining intestinal motility and digestive function. Insufficient exercise makes the intestinal preparation process of the elderly slower and less effective. In this case, even if the patient strictly follows the doctor's instructions for intestinal preparation, insufficient exercise and slow intestinal motility may still lead to intestinal residues not being discharged in time, thus affecting the clarity and accuracy of colonoscopy.

[0005] Due to the above reasons, the quality of intestinal preparation for elderly patients before colonoscopy is usually not ideal. This not only increases the risk of examination failure, but may also cause doctors to be unable to fully observe the condition of the colon wall, and may also miss early lesions due to incomplete intestinal cleaning, increasing the risk of misdiagnosis or missed diagnosis. Summary of the invention

[0006] The present invention provides a system and method for evaluating and detecting intestinal preparation of elderly patients before colonoscopy, aiming to provide elderly patients with a safe, effective and personalized intestinal preparation plan to improve the quality of colonoscopy.

[0007] In a first aspect, the intestinal preparation assessment and detection system for elderly patients before colonoscopy provided by the present invention includes: an image acquisition module, the image acquisition module is used to acquire a first image and a second image, the first image is a stool image of the monitoring target before the intestinal preparation plan is executed, and the second image is a stool image of the monitoring target after the intestinal preparation plan is executed; a feature extraction module, the feature extraction module is used to extract a first feature from the first image, and to extract a second feature from the second image, the first feature is a stool shape feature of the monitoring target before the intestinal preparation plan is executed, and the second feature is a stool shape feature of the monitoring target after the intestinal preparation plan is executed; a plan generation module, the plan generation module is used to generate an intestinal preparation plan based on the first feature, the intestinal preparation plan is a drug intervention measure, an exercise intervention measure and / or an abdominal vibration intervention measure when the monitoring target performs intestinal preparation; a quality assessment module, the quality assessment module is used to evaluate the intestinal preparation quality of the monitoring target according to the first feature and the second feature. The present invention can generate personalized bowel preparation plans from multiple dimensions by monitoring the stool shape characteristics of the target before executing the bowel preparation plan to solve problems such as weakened intestinal function and poor drug absorption in elderly patients; and can evaluate the bowel preparation effect through changes in stool shape characteristics before and after bowel preparation to ensure that elderly patients can timely adjust and optimize the intervention plan during the bowel preparation process, thereby reducing the risk of failure of subsequent colonoscopy.

[0008] In the second aspect, based on the intestinal preparation assessment and detection system for elderly patients before colonoscopy provided in the first aspect, the intestinal preparation assessment and detection method for elderly patients before colonoscopy provided in the present invention comprises the following steps: the image acquisition module acquires a first image; the feature extraction module obtains a first feature through the first image; the solution generation module generates an intestinal preparation solution based on the first feature; the image acquisition module acquires a second image; the feature extraction module obtains a second feature through the second image; the quality assessment module evaluates the intestinal preparation quality of the monitored target based on the first feature and the second feature. The intestinal preparation assessment and detection method for elderly patients before colonoscopy provided in the present invention is applicable to the intestinal preparation assessment and detection system for elderly patients before colonoscopy provided in the present invention.

[0009] Furthermore, the feature extraction module obtains a first feature through the first image, comprising the following steps: using an image feature recognition model to recognize a first feature of the first image, wherein the first feature is one of a feature of type A, a feature of type B, a feature of type C, a feature of type D, a feature of type E, a feature of type F, and a feature of type G. The present invention uses image recognition technology to accurately recognize the shape features of stool before intestinal preparation, and provides basic data support for the personalized generation of subsequent intestinal preparation plans.

[0010] Furthermore, the feature extraction module obtains the second feature through the second image, including the following steps: using an image feature recognition model to recognize the second feature of the second image, wherein the second feature is one of features of type A, type B, type C, type D, type E, type F, and type G. The present invention uses image recognition technology to accurately recognize the shape features of stool after intestinal preparation, providing basic data support for subsequent quality assessment of intestinal preparation.

[0011] Furthermore, the intestinal preparation assessment and detection method for elderly patients before colonoscopy provided by the present invention also includes the following steps: the feature extraction module uses a sample data set to train and verify the image feature recognition model, and any sample in the sample data set is a stool image with a feature label, and the feature label is a feature label of type A feature, type B feature, type C feature, type D feature, type E feature, type F feature, and type G feature. The present invention trains and verifies the image feature recognition model through a large number of sample data sets with a single feature label, thereby ensuring the accuracy of stool shape feature recognition, thereby providing a reliable data basis for the personalized generation of intestinal preparation plans and the quality assessment of intestinal preparation.

[0012] Furthermore, the solution generation module generates a bowel preparation solution based on the first feature, including the following steps: matching a bowel preparation solution with the first feature tag from a bowel preparation solution library based on the first feature. The present invention provides the most suitable drug, exercise or abdominal vibration intervention solution according to the specific stool shape characteristics of the patient, thereby improving the effect of bowel cleaning, improving the accuracy and efficiency of the preparation process, and is particularly suitable for the special needs of elderly patients.

[0013] Furthermore, the intestinal preparation assessment and detection method for elderly patients before colonoscopy provided by the present invention also includes the following steps: the solution generation module obtains a library of intestinal preparation solutions, and any intestinal preparation solution in the library has at least one characteristic tag. The intestinal preparation solution library obtained by the present invention can efficiently match intervention measures that match the patient's characteristics, ensuring the accuracy and personalization of the intestinal preparation solution.

[0014] Furthermore, the scheme generation module generates a bowel preparation scheme based on the first feature, and further includes the following steps: obtaining a third feature, the third feature being a state feature of the detection target; and matching a bowel preparation scheme having the third feature from bowel preparation schemes having the first feature label based on the third feature. The present invention further incorporates the overall health status characteristics of the patient to better adapt to the overall health status of the patient, further improve the effect of bowel preparation, and is particularly suitable for the complex health needs of elderly patients.

[0015] Furthermore, the quality assessment module assesses the intestinal preparation quality of the monitoring target according to the first feature and the second feature, including the following steps: based on the feature quantification standard, quantify the first feature and the second feature respectively to obtain the first quantification parameter and the second quantification parameter; evaluate the effectiveness of the intestinal preparation program executed by the monitoring target through the difference between the first quantification parameter and the second quantification parameter; and evaluate whether the monitoring target can perform colonoscopy after executing the intestinal preparation program through the size of the second quantification parameter. The present invention provides an objective data basis for intestinal preparation quality assessment by quantifying the features before and after intestinal preparation.

[0016] Further, the feature quantification standard is: the feature label of the first feature or the second feature is a type A feature, and the corresponding first quantization parameter or second quantization parameter is a, a is a positive number; the feature label of the first feature or the second feature is a type B feature, and the corresponding first quantization parameter or second quantization parameter is b, b>a; the feature label of the first feature or the second feature is a type C feature, and the corresponding first quantization parameter or second quantization parameter is c, c>b; the feature label of the first feature or the second feature is a type D feature, and the corresponding first quantization parameter or second quantization parameter is d, d>c; the feature label of the first feature or the second feature is an E feature, and the corresponding first quantization parameter or second quantization parameter is e, e>c; the feature label of the first feature or the second feature is an F feature, and the corresponding first quantization parameter or second quantization parameter is f, f>e; the feature label of the first feature or the second feature is a G feature, and the corresponding first quantization parameter or second quantization parameter is g, g>f. The present invention can not only provide a clear numerical basis for the evaluation of intestinal preparation quality through a standardized quantification method, but also provide reliable data support for the subsequent adjustment of intervention measures, which is helpful to improve the quality of intestinal preparation.

[0017] The intestinal preparation assessment and detection system and method for elderly patients before colonoscopy provided by the present invention have the following benefits: the present invention uses image recognition technology to perform feature analysis on the stool shape of elderly patients before intestinal preparation, and generates intervention measures suitable for elderly patients in combination with their health conditions, thereby ensuring adequate cleaning of the intestines, helping to reduce examination failures and repeated examinations due to insufficient preparation, thereby reducing the medical burden on elderly patients, and providing elderly patients with a safer and more efficient intestinal preparation solution, which has broad clinical application value and market potential. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of the intestinal preparation assessment and detection system for elderly patients before colonoscopy provided by the present invention;

[0019] Figure 2 This is a flow chart of the intestinal preparation assessment and detection method for elderly patients before colonoscopy provided by the present invention. DETAILED DESCRIPTION

[0020] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are only for illustration and are not intended to limit the present invention. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it is obvious to those of ordinary skill in the art that these specific details do not need to be adopted to implement the present invention. In other examples, in order to avoid confusing the present invention, known circuits, software or methods are not specifically described.

[0021] In one embodiment, Figure 1 As shown, the intestinal preparation assessment and detection system for elderly patients before colonoscopy provided by the present invention includes: an image acquisition module, a feature extraction module, a solution generation module and a quality assessment module; wherein the image acquisition module is connected to the feature extraction module, the feature extraction module is connected to the solution generation module, and the solution generation module is connected to the quality assessment module.

[0022] Furthermore, the image acquisition module is used to acquire a first image and a second image, the first image being a stool image of the monitored target before the intestinal preparation plan is executed, and the second image being a stool image of the monitored target after the intestinal preparation plan is executed; the feature extraction module is used to extract a first feature from the first image, and to extract a second feature from the second image, the first feature being a stool shape feature of the monitored target before the intestinal preparation plan is executed, and the second feature being a stool shape feature of the monitored target after the intestinal preparation plan is executed; the plan generation module is used to generate an intestinal preparation plan based on the first feature, and the intestinal preparation plan is a drug intervention measure, an exercise intervention measure and / or an abdominal vibration intervention measure when the monitored target performs intestinal preparation; the quality assessment module is used to assess the intestinal preparation quality of the monitored target based on the first feature and the second feature.

[0023] It is understood that the monitoring target described in the present invention refers to the subject who receives intestinal preparation assessment and intervention, usually an elderly patient who is preparing for a colonoscopy. Since elderly patients have particularities in intestinal function, drug absorption, motility, etc., their intestinal preparation effect requires special attention.

[0024] In this embodiment, the intestinal preparation assessment detection system for elderly patients before colonoscopy can implement the intestinal preparation assessment detection method for elderly patients before colonoscopy in the following steps to achieve intestinal preparation guidance and intestinal preparation quality assessment for elderly patients before colonoscopy surgery. Figure 2As shown, the intestinal preparation assessment and detection method for elderly patients in front of the mirror includes the following steps:

[0025] S01. The image acquisition module acquires a first image.

[0026] Furthermore, the image acquisition module has a camera with a data transmission interface, a smart phone, a smart watch with a camera function, and other devices, so as to respectively obtain stool images of the monitored target before and after the bowel preparation program is executed. It should be noted that privacy protection is required during the image acquisition process, and privacy protection is also required when the first image and the identity information of the patient corresponding to the first image are collected.

[0027] S02. The feature extraction module obtains a first feature through the first image.

[0028] In this embodiment, the step S02 of obtaining the first feature through the first image includes the following steps: using an image feature recognition model to recognize the first feature of the first image.

[0029] Furthermore, the stool shape characteristics are specifically divided into 7 categories, from A to G. Different types of stool shape characteristics show different intestinal motility states of the monitored targets. Therefore, the stool shape characteristics can intuitively show the activity level of intestinal function and can serve as the basis for the selection of subsequent intestinal preparation plans.

[0030] In this embodiment, the features of category A to category G are specifically shown in the following Table 1:

[0031]

[0032]

[0033] In this embodiment, the image feature recognition model is a machine learning or deep learning model that uses a computer algorithm to automatically extract and analyze features from image data, such as an image feature recognition model based on a convolutional neural network, an image feature recognition model based on a convolutional neural network and a self-attention mechanism, etc.

[0034] Specifically, the image feature recognition model is an image feature recognition model based on a convolutional neural network (CNN); further, the image feature recognition model is composed of an input layer, a convolution layer, an activation layer, a pooling layer, a fully connected layer and an output layer connected in sequence.

[0035] Further, the input layer is used to receive the first image. In this embodiment, any first image corresponds to a three-dimensional tensor P, and the size of the three-dimensional tensor P is H×W×C, where H represents the height of the first image, W represents the width of the first image, and C represents the number of channels of the first image: if the first image is an RGB image, C is 3; if the first image is a black and white image, C is 1.

[0036] Further, the convolution layer is connected to the input layer, and the convolution layer is used to extract local features in the first image. In this embodiment, the convolution layer includes a first convolution layer, a second convolution layer and a third convolution layer; the first convolution layer is connected to the input layer, the second convolution layer is connected to the first convolution layer, and the third convolution layer is connected to the second convolution layer; the first convolution layer is used to generate a first feature map based on the first image, the second convolution layer is used to generate a second feature map based on the first feature map, and the third convolution layer is used to generate a third feature map based on the second feature map.

[0037] In this embodiment, in order to capture the details in the first image and ensure computational efficiency, the convolution kernel size of the first convolution layer is 3×3×C, where C is the number of channels of the convolution kernel, which is the same as the number of channels of the first image. The number of convolution kernels is K1, the stride is S1, and the padding is P1. The output first feature map satisfies the following expression: in, represents the eigenvalue of the first feature map generated by the kth convolution kernel in the first convolutional layer at position (i, j), represents the weight of the kth convolution kernel at position (m, n) and input channel c, P i+m-1，j+n-1 represents the pixel value of the first image at position (i+ms, j+ns) and channel c, represents the bias term of the kth convolution sum in the first convolutional layer; the size of the first feature map is H1×W1×K1, where H1 represents the height of the first feature map, W1 represents the width of the first feature map, and K1 represents the number of channels of the first feature map: the height of the first feature map Width of the first feature map The number of channels of the first feature map is equal to the number of convolution kernels of the first convolutional layer.

[0038] Similarly, in order to capture the details of the first feature map and ensure computational efficiency, the convolution kernel size of the second convolution layer is 3×3×K1, where K1 is the number of channels of the convolution kernel, which is the same as the number of channels of the first feature map. The number of convolution kernels is K2, the stride is S2, and the padding is P2. The expression and size of the second feature map output by the second convolution layer can be found in the expression and size of the first feature map output by the first convolution layer.

[0039] Similarly, in order to capture the details of the second feature map and ensure computational efficiency, the convolution kernel size of the third convolution layer is 3×3×K2, where K2 is the number of channels of the convolution kernel, which is the same as the number of channels of the second feature map. The number of convolution kernels is K3, the stride is S3, and the padding is P3. The expression and size expression of the third feature map output by the third convolution layer can be found in the expression and size expression of the first feature map output by the first convolution layer.

[0040] Further, the activation layer is connected to the convolution layer, and the activation layer is used to perform a non-deformation transformation on the output of the convolution layer. In this embodiment, the activation layer includes a first activation layer, a second activation layer, and a third activation layer, wherein the first activation layer is connected to the first convolution layer, the second activation layer is connected to the second convolution layer, and the third activation layer is connected to the third convolution layer.

[0041] Furthermore, in order to improve the computational efficiency, the activation function selected by the first activation layer is the ReLU function. In other embodiments, functions such as Leaky ReLU can also be selected as activation functions. Specifically, after any eigenvalue in the first feature map is input into the ReLU function, the activated eigenvalue satisfies the following expression: in, represents the activation value of the first feature map generated by the kth convolution kernel in the first convolutional layer at position (i, j), The eigenvalue of the first feature map generated by the kth convolution kernel in the first convolutional layer at position (i, j).

[0042] Similarly, the activation functions selected for the second activation layer and the third activation layer are the same as the activation function of the first activation layer, and the corresponding activated eigenvalue expressions are not repeated here.

[0043] Further, the pooling layer is used to reduce the spatial dimension of the activation layer output. In this embodiment, the pooling layer includes a first pooling layer, a second pooling layer and a third pooling layer, wherein the first pooling layer is connected to the first activation layer, the second pooling layer is connected to the second activation layer, and the third pooling layer is connected to the third activation layer.

[0044] More specifically, in order to extract the main features in the feature image and reduce the risk of overfitting, any pooling window in the first pooling layer selects the maximum activation value as the output value. In other embodiments, the average value can also be selected as the output value. Similarly, the second pooling layer and the third pooling layer also select the maximum activation value in the corresponding area as the output value.

[0045] Furthermore, the fully connected layer is used to combine the output of the pooling layer and map the output of the pooling layer to the classification space. In this embodiment, the fully connected layer is connected to the first pooling layer, the second pooling layer and the third pooling layer respectively, and the classification space corresponding to the fully connected layer includes class A feature space, class B feature space, class C feature space, class D feature space, class E feature space, class F feature and class G feature space.

[0046] The score of any type of feature space satisfies the following expression: τ =W τ ·x+b τ , where τ is the feature space number, τ∈{A,B,C,D,E,F,G}, y τ represents the score of mapping the one-dimensional vector obtained by flattening the output of the pooling layer to the feature space of class A, W τ represents the weight vector of the τ-class feature space, x represents the one-dimensional vector obtained by flattening the output of the pooling layer, and b τ Represents the bias of the τ-like feature space.

[0047] Furthermore, the output layer is used to generate and output a feature classification result according to the mapping result of the fully connected layer. In this embodiment, the output layer is connected to the fully connected layer, and the feature result output by the output layer satisfies the following expression: Among them, Predicted Class represents the feature result output by the output layer. Indicates that the feature with the highest prediction probability is the feature result output by the output layer. represents the predicted probability of the τ-class feature space.

[0048] In this embodiment, the image feature recognition module extracts shape features in the first image to determine the first feature; in one or some other embodiments, the image feature recognition model can extract other features such as size features, color features, etc. in the first image to assist in the recognition and judgment of the first feature.

[0049] Furthermore, the image feature recognition model includes the following acquisition steps: using a sample data set to train and verify the image feature recognition model, wherein any sample in the sample data set is a stool image with a feature label, that is, the feature label of any sample is a feature label among class A features, class B features, class C features, class D features, class E features, class F features, and class G features.

[0050] S03. The plan generation module generates a bowel preparation plan based on the first feature.

[0051] In this embodiment, the bowel preparation program is a drug intervention measure, an exercise intervention measure and / or an abdominal vibration intervention measure when the monitoring target is performing bowel preparation.

[0052] Furthermore, the drug intervention measures refer to the type, dosage, time of taking and method of use of the drugs that the monitoring target needs to take during bowel preparation. Common drugs include laxatives such as polyethylene glycol and electrolyte solutions, as well as prokinetic drugs such as domperidone and bisacodyl.

[0053] Furthermore, the exercise intervention measures refer to exercise measures that the monitored target needs to perform during intestinal preparation to promote intestinal motility, such as walking, light jogging, gymnastics, etc.

[0054] Furthermore, the abdominal vibration refers to an auxiliary measure in which the monitored target needs to use abdominal vibration equipment (such as a massager, an abdominal vibrator) to help stimulate intestinal peristalsis during intestinal preparation.

[0055] Conclusions: Abdominal vibration is a non-invasive physical therapy suitable for elderly patients with limited exercise capacity, or as an alternative or adjunct to exercise intervention.

[0056] In this embodiment, the step S03 of generating a bowel preparation plan based on the first feature includes the following steps:

[0057] S031. Based on the first feature, match the intestinal preparation plan with the first feature label from the intestinal preparation plan library.

[0058] It can be understood that the intestinal preparation program library is a structured database for storing multiple intestinal preparation programs, and each intestinal preparation program in the intestinal preparation program library is composed of at least one dimension of intervention measures, including drug intervention measures, exercise intervention measures, abdominal vibration intervention measures, etc., and each intestinal preparation program has at least one feature label.

[0059] Furthermore, the matching model of step S031 satisfies the following expression: BP = arg{(e1, E ε )=1,forε∈{1,2,3,...,Φ}},where BP represents the Ω Among the Ω feature labels, there is a feature label E ε For the intestinal preparation scheme whose matching value with the first feature E1 of the monitoring target is 1, forε∈{1,2,3,...,Φ} represents a cyclic matching, that is, the first feature e1 is matched one by one with Φ feature labels in an intestinal preparation scheme.

[0060] Among them, (e1,E ε )=1 means e1 and E εcorresponds to the same feature label, such as a feature label among class A features, class B features, class C features, class D features, class E features, class F features, and class G features; (e1,E ε )=0 means e1 and E ε The corresponding feature labels.

[0061] Specifically, step S031 matches the stool shape feature, any one of the features from Class A to Class G, with the intestinal preparation program label based on the same feature label in the intestinal preparation program library, and further outputs the matching result, such as "Class A feature; take 4 liters of polyethylene glycol solution daily, divided into 2 times, 2 liters each time, taking time: 2 pm and 8 pm; walk for 45 minutes daily, divided into 2 times, each time walking for 22-23 minutes, 8 am and 4 pm; abdominal vibration twice a day, each vibration time is 15 minutes, the vibration frequency is 50 Hz, use time: 10 am and 9 pm".

[0062] Furthermore, the step S03 of generating a bowel preparation plan based on the first feature also includes the following steps: obtaining a bowel preparation plan library, wherein any bowel preparation plan in the bowel preparation plan library has at least one feature tag.

[0063] In this embodiment, any intestinal preparation plan is a plan sample with a stool characteristic label index designed by medical experts (such as gastroenterologists, nutritionists, etc.) based on medical test data, clinical data, etc.

[0064] In one or some other embodiments, to ensure the compatibility of each bowel preparation plan with the monitoring target, any bowel preparation plan also has a status tag and other personalized tags. The step S03 of generating a bowel preparation plan based on the first feature further includes the following steps:

[0065] S0321. Obtain a third feature, where the third feature is a state feature of the detection target.

[0066] In this or these embodiments, the third characteristic refers to a state characteristic of the monitored target, such as basic physiological characteristics, medical history characteristics, drug tolerance characteristics, health status characteristics, etc.

[0067] It is understandable that these status characteristics not only take into account the patient's health status, but also include information related to individual patient differences, medical history, drug tolerance, etc.

[0068] S0322. Based on the third feature, match the intestinal preparation plan with the third feature from the intestinal preparation plans with the first feature label.

[0069] In step S0322, by combining the third characteristic of the patient, a plan that best meets the patient's needs is further screened out from the matched intestinal preparation plans with the first characteristic label.

[0070] Furthermore, the matching model of step S0322 satisfies the following expression: BP = arg{(e1, E ε )=1,(e3,E ∈ )=1,forε∈{1,2,3,...,Φ},∈∈{1,2,3,...,Φ}},where BP represents the Ω Among the Φ feature labels, there is a feature label E ε The matching value with the first feature E1 of the monitoring target is 1, and there is a feature label E ∈ The intestinal preparation plan whose matching value with the first feature E3 of the monitoring target is 1; similarly, for ε∈{1,2,3,...,Φ}, ∈∈{1,2,3,...,Φ} represents a circular matching, that is, the first feature e1 matches Φ feature labels in an intestinal preparation plan one by one, and the third feature e3 matches Ω feature labels in an intestinal preparation plan one by one.

[0071] Similarly, (e3,E ∈ )=1 means e3 and E ∈ corresponds to the same feature label; (e3,E ∈ )=0 means e3 and E ∈ The corresponding feature labels.

[0072] S04. The image acquisition module acquires a second image.

[0073] The specific method for obtaining the second image in step S04 can refer to the specific method for obtaining the first image in step S01, which will not be repeated here.

[0074] S05. The feature extraction module obtains a second feature through the second image.

[0075] Similarly, the specific identification method of the second feature in the second image can refer to the specific identification method of the first feature in step S02, which will not be repeated here.

[0076] It should be noted that in any intestinal preparation quality assessment process, the identification of the first feature and the second feature uses the same image feature recognition model to ensure the consistency, accuracy and continuity of the two feature results, thereby reducing the errors and deviations caused by different image feature recognition models, and thus ensuring the reliability of intestinal preparation quality assessment.

[0077] S06. The quality assessment module assesses the intestinal preparation quality of the monitored target based on the first feature and the second feature.

[0078] In this embodiment, the step S06 of evaluating the intestinal preparation quality of the monitoring target according to the first feature and the second feature includes the following steps:

[0079] S061. Based on a feature quantization standard, quantize the first feature and the second feature respectively to obtain a first quantization parameter and a second quantization parameter.

[0080] It can be understood that the characteristic quantification standard is a rule and method for converting the shape characteristics of stool (such as Class A to Class G) into numerical parameters, which can assign a quantitative value to different stool shapes, thereby achieving quantitative evaluation of intestinal preparation quality based on the characteristic quantification value.

[0081] Further, the feature quantization standard is: the feature label of the first feature or the second feature is a type A feature, and the corresponding first quantization parameter or second quantization parameter is a; the feature label of the first feature or the second feature is a type B feature, and the corresponding first quantization parameter or second quantization parameter is b, b>a; the feature label of the first feature or the second feature is a type C feature, and the corresponding first quantization parameter or second quantization parameter is c, c>b; the feature label of the first feature or the second feature is a type D feature, and the corresponding first quantization parameter or second quantization parameter is d, d>c; the feature label of the first feature or the second feature is an E feature, and the corresponding first quantization parameter or second quantization parameter is e, e>c; the feature label of the first feature or the second feature is an F feature, and the corresponding first quantization parameter or second quantization parameter is f, f>e; the feature label of the first feature or the second feature is a G feature, and the corresponding first quantization parameter or second quantization parameter is g, g>f; further, any quantization parameter value in the feature quantization standard is a positive integer, that is, a, b, c, d, e, f, g, are all positive integers.

[0082] Specifically, step S061 satisfies the following quantization model: Wherein, P represents a quantization parameter, e represents a feature, including a first feature or a second feature, A represents a class A feature, B represents a class B feature, C represents a class C feature, D represents a class D feature, E represents an class E feature, F represents an F feature, and G represents a class G feature. Specifically, the feature quantization parameters are shown in Table 2 below:

[0083]

[0084]

[0085] As shown in Table 2, from type A features to type G features, the quantitative parameter gradually increases from 0.1 to 0.7. Different quantitative parameters reflect the changes in intestinal motility and defecation difficulty: the smaller the quantitative parameter, the drier and harder the stool, the more difficult the defecation, and the slower the intestinal motility; the larger the quantitative parameter, the looser and softer the stool, the easier the defecation, and the faster the intestinal motility.

[0086] S062. Evaluate the effectiveness of the intestinal preparation program performed by the monitoring target through the difference between the first quantitative parameter and the second quantitative parameter.

[0087] Furthermore, the effectiveness of the bowel preparation program satisfies the following evaluation model: Wherein, Ω represents the effectiveness of the intestinal preparation plan, Ω=1 represents that the intestinal preparation plan is effective, Ω=0 represents that the intestinal preparation plan is ineffective, P1 represents the first quantitative parameter, and P2 represents the second quantitative parameter.

[0088] S063. Evaluate, by the size of the second quantitative parameter, whether the monitored target can undergo colonoscopy after executing the intestinal preparation plan.

[0089] Furthermore, the adaptability of the colonoscopy surgery satisfies the judgment model: Among them, Γ represents the adaptability of colonoscopy, Γ=1 means that the monitored target can undergo colonoscopy after executing the intestinal preparation plan, Γ=0 means that the monitored target cannot undergo colonoscopy after executing the intestinal preparation plan, P2 represents the second quantitative parameter, and g represents the quantitative parameter corresponding to the G-type feature.

[0090] In one or some other embodiments, if the monitored target cannot directly undergo colonoscopy after executing the bowel preparation plan, bowel preparation is further re-performed based on the bowel preparation evaluation result.

[0091] like Figure 2 As shown, the intestinal preparation assessment and detection method for elderly patients before colonoscopy also includes the following steps: S07, the plan generation module regenerates the intestinal preparation plan according to the intestinal preparation quality assessment result.

[0092] Furthermore, after the monitored target executes the initial intestinal preparation plan, a second feature is identified through a second image to reflect the current intestinal cleansing status of the monitored target, for example, the stool is still dry and hard (such as Class A feature, Class B feature or Class C feature) or relatively soft (such as Class D feature, Class E feature or Class F feature).

[0093] When the criteria for colonoscopy are not met (G-type features), the plan generation module is further used to automatically generate a new intestinal preparation plan based on the current stool shape feature (i.e., the second feature), and further optimize the patient's intestinal cleansing.

[0094] It should be noted that when the intestinal preparation quality assessment result shows that the first feature is the same as the second feature, the intestinal preparation plan matched based on the second feature is different from the intestinal preparation plan matched based on the first feature to ensure the effectiveness of the implementation of the intestinal preparation plan.

[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents.

Claims

1. A system for evaluating and detecting intestinal preparation of elderly patients before colonoscopy, characterized in that: include: An image acquisition module, the image acquisition module is used to acquire a first image and a second image, the first image is a stool image of the monitoring target before the intestinal preparation program is performed, and the second image is a stool image of the monitoring target after the intestinal preparation program is performed; a feature extraction module, the feature extraction module being used to extract a first feature from the first image and a second feature from the second image, the first feature being a stool shape feature of the monitoring target before the intestinal preparation program is executed, and the second feature being a stool shape feature of the monitoring target after the intestinal preparation program is executed; A scheme generating module, the scheme generating module is used to generate a bowel preparation scheme based on the first feature, the bowel preparation scheme being a drug intervention measure, an exercise intervention measure and / or an abdominal vibration intervention measure when the monitoring target performs bowel preparation; A quality assessment module is used to assess the intestinal preparation quality of the monitored target based on the first feature and the second feature.

2. A method for evaluating intestinal preparation of elderly patients before colonoscopy, characterized in that: The intestinal preparation assessment and detection method for elderly patients before colonoscopy is applicable to the intestinal preparation assessment and detection system for elderly patients before colonoscopy according to claim 1, and comprises the following steps: The image acquisition module acquires a first image; The feature extraction module obtains a first feature through the first image; The plan generation module generates a bowel preparation plan based on the first feature; The image acquisition module acquires a second image; The feature extraction module obtains a second feature through the second image; The quality assessment module assesses the intestinal preparation quality of the monitoring target according to the first feature and the second feature.

3. The method for evaluating intestinal preparation of elderly patients before colonoscopy according to claim 2, characterized in that: The feature extraction module obtains a first feature through the first image, comprising the following steps: An image feature recognition model is used to identify a first feature of the first image, where the first feature is one of a type A feature, a type B feature, a type C feature, a type D feature, a type E feature, a type F feature, and a type G feature.

4. The method for evaluating intestinal preparation of elderly patients before colonoscopy according to claim 2, characterized in that: The feature extraction module obtains the second feature through the second image, comprising the following steps: An image feature recognition model is used to identify a second feature of the second image, where the second feature is one of the following: A feature, B feature, C feature, D feature, E feature, F feature, or G feature.

5. The method for evaluating intestinal preparation of elderly patients before colonoscopy according to any one of claims 2-3, characterized in that: The following steps are also included: The feature extraction module uses a sample data set to train and verify an image feature recognition model, wherein any sample in the sample data set is a stool image with a feature label, and the feature label is a feature label selected from the group consisting of A feature, B feature, C feature, D feature, E feature, F feature, and G feature.

6. The method for evaluating intestinal preparation of elderly patients before colonoscopy according to claim 5, characterized in that: The plan generation module generates a bowel preparation plan based on the first feature, including the following steps: Based on the first feature, a bowel preparation regimen having the first feature label is matched from a bowel preparation regimen library.

7. The method for evaluating intestinal preparation of elderly patients before colonoscopy according to claim 6, characterized in that: The following steps are also included: The solution generation module obtains a bowel preparation solution library, and any bowel preparation solution in the bowel preparation solution library has at least one characteristic tag.

8. The method for evaluating intestinal preparation of elderly patients before colonoscopy according to claim 6, characterized in that: The plan generation module generates a bowel preparation plan based on the first feature, and further includes the following steps: Acquire a third feature, where the third feature is a state feature of the detection target; Based on the third feature, a bowel preparation plan having the third feature is matched from bowel preparation plans having the first feature tag.

9. The method for evaluating intestinal preparation of elderly patients before colonoscopy according to claim 5, characterized in that: The quality assessment module assesses the intestinal preparation quality of the monitoring target according to the first feature and the second feature, including the following steps: Based on a feature quantization standard, quantizing the first feature and the second feature respectively to obtain a first quantization parameter and a second quantization parameter; evaluating the effectiveness of the bowel preparation program performed by the monitoring target through the difference between the first quantitative parameter and the second quantitative parameter; The size of the second quantitative parameter is used to evaluate whether the monitored target can undergo colonoscopy after executing the intestinal preparation plan.

10. The method for evaluating intestinal preparation of elderly patients before colonoscopy according to claim 9, characterized in that: The characteristic quantification standard is: The feature label of the first feature or the second feature is a class A feature, and the corresponding first quantization parameter or the second quantization parameter is a, where a is a positive number; The feature label of the first feature or the second feature is a class B feature, and the corresponding first quantization parameter or the second quantization parameter is b, where b>a; The feature label of the first feature or the second feature is a C-type feature, and the corresponding first quantization parameter or the second quantization parameter is c, where c>b; The feature label of the first feature or the second feature is a D-type feature, and the corresponding first quantization parameter or the second quantization parameter is d, d>c; The feature label of the first feature or the second feature is a class E feature, and the corresponding first quantization parameter or the second quantization parameter is e, e>c; The feature label of the first feature or the second feature is F-type feature, and the corresponding first quantization parameter or second quantization parameter is f, f>e; the feature label of the first feature or the second feature is G-type feature, and the corresponding first quantization parameter or second quantization parameter is g, g>f.

Images