Gastrorrhagia monitoring system and method

Abstract

The invention relates to the technical field of gastroenterology department medical instruments, in particular to a gastrorrhagia monitoring system and method.According to the gastrorrhagia monitoring system and method, the hemoglobin concentration in gastric juice is obtained through a sensor, and then the hemoglobin concentration is input into a gastrorrhagia monitoring model to be used for achieving nearly real-time monitoring of gastrorrhagia; meanwhile, a deep learning model is introduced for gastrorrhagia monitoring, so that the accuracy of gastrorrhagia monitoring is improved; meanwhile, when the gastrorrhagia monitoring sample set is amplified, the sample set amplification multiple is determined according to the number of the elements in the sample set, namely, if the number of the elements in the sample set is large, the smaller sample set amplification multiple is set, and if the number of the elements in the sample set is small, the larger sample set amplification multiple is set, so that the situation that when the sample set is large, the sample set is not amplified is avoided. Meanwhile, more sample sets are amplified, so that the training process consumes too long time, and the training efficiency of the gastrorrhagia monitoring model is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of gastroenterology medical instruments, and in particular to a gastric bleeding monitoring system and method. Background Art

[0002] Gastric bleeding refers to non-varicose gastrointestinal bleeding in the stomach and other areas above the Treitz ligament, with gastric bleeding as the main clinical manifestation. Patients with large amounts of bleeding are often accompanied by clinical manifestations such as insufficient or failure of peripheral circulation, resulting in low perfusion of important organs and resulting in dysfunction of corresponding organs. In clinical practice, acute gastric mucosal lesions are the most common form of gastric bleeding.

[0003] In the prior art, when monitoring gastric bleeding, during gastrointestinal decompression, although gastric juice is continuously drained, when the amount of bleeding is small, it is impossible to distinguish it with the naked eye, and it can only be discovered when the color of the gastric juice changes significantly (brown, black, etc.); enteral nutrition support routinely withdraws the gastric tube every 2h or 4h to observe the gastric juice and other conditions. If bleeding occurs within the interval of 2h or 4h, it cannot be discovered in time; for the treatment of gastrointestinal bleeding, intermittent tests are required to understand the treatment effect, the improvement or deterioration of the bleeding situation, and its real-time changes cannot be grasped; therefore, the real-time performance of gastric bleeding monitoring is often not strong; at the same time, the prior art also has the problem of low accuracy and efficiency of gastric bleeding monitoring. Summary of the invention

[0004] In view of this, the present invention proposes a gastric bleeding monitoring system and method, which are used to improve the real-time monitoring, monitoring accuracy and monitoring efficiency of gastric bleeding.

[0005] In order to achieve the above object, a gastric bleeding monitoring system is provided, the system comprising the following modules: A collection module, used for collecting data of gastric bleeding indicators to be monitored; A transmission module, connected to the acquisition module, for transmitting data of the gastric bleeding indicators to be monitored acquired by the acquisition module; A data preprocessing module, connected to the transmission module, performs preprocessing operations on the gastric bleeding indicator value data to be monitored; A gastric bleeding monitoring module, connected to the transmission module, for inputting the monitoring index value data preprocessed by the data preprocessing module into the gastric bleeding monitoring model for detecting gastric bleeding; The sample set amplification of the gastric bleeding monitoring model is used to improve the monitoring accuracy of the gastric bleeding monitoring model. The sample set amplification of the gastric bleeding monitoring model is specifically as follows: determining the amplification multiple of the sample set according to the size of the preset number; The alarm module is connected to the gastric bleeding monitoring module and is used to realize an alarm according to the monitoring result of the gastric bleeding monitoring module.

[0006] Preferably, the collection module is specifically: a hemoglobin sensor is arranged at the front end of the gastric tube to constitute the collection module.

[0007] Preferably, the calculation formula for the amplification multiple M of the sample set is:

[0008] Wherein, round() is a rounding function, and N is a preset number of monitoring index value data input into the gastric bleeding monitoring model.

[0009] Preferably, the operating logic of the alarm module is as follows: if the gastric bleeding monitoring module detects that there is no hemoglobin in the gastric juice, the display screen shows normal and the green light is on; if the gastric bleeding monitoring module detects that there is hemoglobin in the gastric juice, but the content is less than or equal to the set value, the display screen shows suspicious and the yellow light is on; when the gastric bleeding monitoring module detects that the hemoglobin content in the sensed gastric juice is greater than the set value, the red light is on and an alarm sound is issued.

[0010] Preferably, in S2, the preprocessing operation includes outlier processing and missing value processing.

[0011] Preferably, the missing value processing is specifically: using mean interpolation method to process the missing values.

[0012] Preferably, the gastric bleeding monitoring model is an artificial neural network model.

[0013] Preferably, the structure of the artificial neural network is divided into three parts: an input layer, a hidden layer and an output layer.

[0014] Preferably, the number of hidden layers is not fixed.

[0015] According to another aspect of the present invention, a method for monitoring gastric bleeding is provided, wherein the method adopts the above-mentioned gastric bleeding monitoring system, and the method comprises: S1: Obtaining monitoring index value data in the gastric fluid of the patient to be tested; S2: performing preprocessing operations on the monitoring indicator value data; S3: inputting the monitoring index value data pre-processed by S2 into a gastric bleeding monitoring model for detecting gastric bleeding; Among them, the sample set amplification of the gastric bleeding monitoring model is used to improve the monitoring accuracy of the gastric bleeding monitoring model. The sample set amplification of the gastric bleeding monitoring model is specifically: determining the amplification multiple of the sample set according to the size of the preset number.

[0016] The advantages and beneficial effects of the present invention are: The present invention obtains the hemoglobin concentration in gastric juice through a sensor, and then inputs the hemoglobin concentration into a gastric bleeding monitoring model to achieve near real-time monitoring of gastric bleeding; at the same time, the present invention introduces a deep learning model for gastric bleeding monitoring, thereby improving the accuracy of gastric bleeding monitoring; At the same time, when the present invention amplifies the sample set for gastric bleeding monitoring, the sample set amplification multiple is determined according to the number of elements in the sample set, that is, if the number of elements in the sample set is large, a smaller sample set amplification multiple is set, and if the number of elements in the sample set is small, a larger sample set amplification multiple is set. This avoids amplifying more sample sets at the same time when the sample set is large, resulting in a training process that takes too long, thereby improving the training efficiency of the gastric bleeding monitoring model. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the description of the present invention or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A schematic diagram of a gastric bleeding monitoring system provided by an embodiment of the present invention; Figure 2 The present invention provides a flow chart of a method for monitoring gastric bleeding. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] As attached Figure 1 As shown, a gastric bleeding monitoring system comprises the following modules: A collection module, used for collecting data of gastric bleeding indicators to be monitored; wherein the monitoring indicator is the hemoglobin concentration in gastric juice; Specifically, the hemoglobin concentration in the gastric juice is monitored by a hemoglobin sensor. After the hemoglobin concentration in the gastric juice is obtained, the hemoglobin concentration is transmitted to a data receiver; the data receiver is used to receive the hemoglobin concentration in the gastric juice; A transmission module, connected to the acquisition module, for transmitting data of the gastric bleeding indicators to be monitored acquired by the acquisition module; A data preprocessing module, connected to the transmission module, performs preprocessing operations on the gastric bleeding indicator value data to be monitored; Wherein, the preprocessing operation includes outlier processing, missing value processing, etc.; The outlier processing is to remove the values ​​that are obviously deviated from the reference value; For example, errors in data collection or data transmission may cause abnormal values ​​in the acquired data. For example, the hemoglobin concentration is 1600 g / L, which is obviously deviated from the baseline value and should be treated as an abnormal value and deleted; In reality, there may be varying degrees of missing values ​​in variables in data due to many reasons, which is particularly common in medical data. For example, errors in the data collection and storage process may also lead to missing values ​​in the data.

[0021] The missing value processing is specifically: using mean interpolation method to process the missing values; A gastric bleeding monitoring module, connected to the transmission module, for inputting the monitoring index value data preprocessed by the data preprocessing module into a gastric bleeding monitoring model for detecting gastric bleeding; wherein the gastric bleeding monitoring model is an artificial neural network model; Artificial Neural Network (ANN) is a hot research topic in the field of artificial intelligence since the 1980s. It abstracts the human brain neural network from the perspective of information processing, establishes a simple model, and forms different networks according to different connection methods. In engineering and academia, it is often referred to as neural network or quasi-neural network. Neural network is a computational model composed of a large number of nodes (or neurons) connected to each other. Each node represents a specific output function, called activation function. The connection between each two nodes represents a weighted value for the signal passing through the connection, called weight, which is equivalent to the memory of artificial neural network. The output of the network varies depending on the connection method, weight value and activation function of the network. The network itself is usually an approximation of a certain algorithm or function in nature, or it may be an expression of a logical strategy. An artificial neural network is a computational model that simulates the structure and function of the human brain's nervous system. It learns and trains a large amount of data to achieve a corresponding mapping relationship from input data to output data. The structure of the artificial neural network is divided into three parts: input layer, hidden layer and output layer. The number of hidden layers is not fixed. Each layer is composed of multiple neurons, and the neurons are connected by weights. Artificial neural network is a nonlinear, adaptive information processing system composed of a large number of interconnected processing units. It is proposed based on the research results of modern neuroscience and attempts to process information by simulating the way brain neural networks process and memorize information. Artificial neural networks have four basic characteristics: (1) Nonlinearity Nonlinear relationships are universal characteristics of nature. The intelligence of the brain is a nonlinear phenomenon. Artificial neurons are in two different states: activated or inhibited. This behavior is mathematically expressed as a nonlinear relationship. A network composed of neurons with thresholds has better performance and can improve fault tolerance and storage capacity.

[0022] (2) Non-locality A neural network is usually composed of multiple neurons that are widely connected. The overall behavior of a system depends not only on the characteristics of individual neurons, but may also be mainly determined by the interactions and connections between units. The non-locality of the brain is simulated through a large number of connections between units. Associative memory is a typical example of non-locality.

[0023] (3) Very qualitative artificial neural networks have the ability to adapt, self-organize, and self-learn. Not only can the information processed by the neural network be subject to various changes, but while processing information, the nonlinear dynamic system itself is also constantly changing. Iterative processes are often used to describe the evolution of dynamic systems.

[0024] (4) Non-convexity The evolution direction of a system will depend on a certain state function under certain conditions. For example, the energy function, whose extreme value corresponds to the relatively stable state of the system. Non-convexity means that this function has multiple extreme values, so the system has multiple relatively stable equilibrium states, which will lead to the diversity of system evolution.

[0025] According to the different learning environments, the learning methods of neural networks can be divided into supervised learning and unsupervised learning. In supervised learning, the data of the training samples are added to the network input, and the corresponding expected output is compared with the network output to obtain an error signal, which is used to control the adjustment of the weight connection strength. After multiple trainings, it converges to a certain weight. When the sample situation changes, the weight can be modified to adapt to the new environment through learning. Neural network models that use supervised learning include back propagation networks and perceptrons. In unsupervised learning, standard samples are not given in advance, and the network is directly placed in the environment, and the learning stage and the working stage become one. At this time, the change of learning rules obeys the evolution equation of the connection weight. The simplest example of unsupervised learning is the Hebb learning rule. The competitive learning rule is a more complex example of unsupervised learning, which adjusts the weights according to the established clusters. Self-organizing maps, adaptive resonance theory networks, etc. are typical models related to competitive learning.

[0026] Compared with traditional linear models, the artificial neural network has excellent processing capabilities for complex nonlinear relationships and can better adapt to problems in medical classification. In addition, artificial neural networks are widely adaptable because they can continuously optimize their performance through learning and training, so that they can cope with different tasks and data. Similar to the human brain, due to the characteristics of distributed storage and processing of information by artificial neural networks, artificial neural network models can still operate normally when some neurons or connections fail.

[0027] In this embodiment, anaconda3-5.3.1 and python3.7.13 are used to build the machine learning environment required for this embodiment. The system adopts Windows 10 Enterprise Edition G.21H1, and the selected hardware is a 16-core, Intel (R) Xeon (R) Gold 62268 CPU @2.90GHz CPU. The numpy version is 1.20.1, the pandas version is 1.2.4, the scikit-learn version is 1.0.2, the joblib version is 1.0.1, the matplotlib version is 3.2.2, and the pyqt5 version is 5.15.40.

[0028] In this embodiment, the hyperparameters of the gastric bleeding monitoring model include activation function, loss function, and learning rate; Among them, the calculation process of artificial neural network is the operation between matrices, and matrix operation is still linear operation no matter how many times it is superimposed. In real life, there is basically no pure linear problem. In order to enable the network model to solve nonlinear problems, it is often necessary to add activation functions of nonlinear factors to the artificial neural network, so as to improve the nonlinear mapping ability of the neural network, that is, the approximation ability; In this embodiment, the activation function is a sigmoid function; The loss function is a non-negative function that measures the difference between the predicted value and the true value in the artificial neural network. It is the basis for updating parameters in the artificial neural network model. The smaller the loss function, the more accurate the model's prediction. The essence of network model training is to minimize the loss by continuously calculating the loss function and updating the network weights and biases. In this embodiment, the loss function is a cross entropy loss function; The learning rate is a hyperparameter that adjusts the rate of change of the gradient descent of the neural network through the loss function. The larger the learning rate, the faster the gradient of the loss function changes, and the smaller the learning rate, the slower the gradient of the loss function changes. If the learning rate is too small, the network model will take more time to train, which is extremely inefficient, while if the learning rate is too large, it will easily cause the network model to be unstable or even unable to converge. Therefore, it is very important to set a suitable learning rate.

[0029] In this embodiment, the learning rate is set to 0.01; Specifically, the input of the artificial neural network model is a preset number of monitoring index value data preprocessed by S2, and the output of the artificial neural network model is a gastric bleeding monitoring result, including two results: the presence of gastric bleeding and the absence of gastric bleeding; Among them, the preset number is one of 10, 15, 20, and 50. This embodiment improves the prediction accuracy of the model by inputting multiple monitoring indicator value data into the deep learning model.

[0030] As a preferred embodiment, in order to improve the monitoring accuracy of gastric bleeding, in this embodiment, when training the gastric bleeding monitoring model, since the gastric bleeding monitoring technology appeared relatively late, it is difficult to obtain the sample set. Therefore, this embodiment needs to expand the sample set of the gastric bleeding monitoring model to improve the generalization ability of the gastric bleeding monitoring model, thereby improving the prediction accuracy of the gastric bleeding monitoring model; Specifically, the sample set amplification of the gastric bleeding monitoring model is specifically as follows: determining the amplification multiple of the sample set according to the size of the preset number; Specifically, the calculation formula for the amplification multiple M of the sample set is:

[0031] Wherein, round() is a rounding function, and N is a preset number of monitoring index value data input into the gastric bleeding monitoring model; In this embodiment, when the sample set of the gastric bleeding monitoring model is amplified, the sample set amplification multiple is determined according to the number of elements in the sample set, that is, if the number of elements in the sample set is large, a smaller sample set amplification multiple is set; if the number of elements in the sample set is small, a larger sample set amplification multiple is set. This avoids amplifying more sample sets at the same time when the sample set is large, resulting in a long training process, thereby improving the training efficiency of the gastric bleeding monitoring model.

[0032] In this embodiment, the accuracy, precision, recall and F1 value are used to measure whether the gastric bleeding monitoring model training is completed; The accuracy is the ratio of correctly classified samples to the total number of samples, and is the simplest and most intuitive classification indicator. A high accuracy does not mean good model performance, because when there is an imbalance in the categories of the data set, the accuracy may be misleadingly high.

[0033] The precision rate measures the proportion of samples predicted as positive by the model that are actually positive. That is, how many samples predicted as positive are actually positive. The precision rate can measure the accuracy of the model, especially when the classification result is positive.

[0034] The recall rate measures the proportion of samples that are actually positive and are predicted as positive by the model. That is, how many of the real positive samples are predicted as positive. The recall rate can measure the ability of the model to find positive samples.

[0035] The F1 value takes into account both precision and recall and is the harmonic mean of precision and recall. The higher the F1 value, the better the performance of the model.

[0036] An alarm module, connected to the gastric bleeding monitoring module, and used to realize an alarm according to the monitoring result of the gastric bleeding monitoring module; The operating logic of the alarm module is specifically as follows: if the gastric bleeding monitoring module detects that there is no hemoglobin in the gastric juice, the display screen shows normal and the green light is on; if the gastric bleeding monitoring module detects that there is hemoglobin in the gastric juice, but the content is less than or equal to the set value, the display screen shows suspicious and the yellow light is on; when the gastric bleeding monitoring module detects that the hemoglobin content in the sensed gastric juice is greater than the set value, the red light is on and an alarm sound is issued.

[0037] Embodiment 2, as attached Figure 2 As shown, this embodiment includes a gastric bleeding monitoring method, the method adopts a gastric bleeding monitoring system of embodiment 1, and the method includes: S1: Obtaining monitoring index value data in the gastric fluid of the patient to be tested; S2: performing preprocessing operations on the monitoring indicator value data; S3: inputting the monitoring index value data pre-processed by S2 into a gastric bleeding monitoring model for detecting gastric bleeding; Among them, the sample set amplification of the gastric bleeding monitoring model is used to improve the monitoring accuracy of the gastric bleeding monitoring model. The sample set amplification of the gastric bleeding monitoring model is specifically: determining the amplification multiple of the sample set according to the size of the preset number.

[0038] Embodiment 3: This embodiment includes a computer-readable storage medium, on which a data processing program is stored, and the data processing program is executed by a processor to implement a method for monitoring gastric bleeding in embodiment 2.

[0039] It should be understood by those skilled in the art that the embodiments of this article can be provided as methods, devices (equipment), or computer program products. Therefore, this article can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Including but not limited to RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, tape, disk storage or other magnetic storage device, or any other medium that can be used to store desired information and can be accessed by a computer. In addition, it is well known to those of ordinary skill in the art that communication media generally contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0040] This document is described with reference to flowcharts and / or block diagrams of methods, apparatuses (devices), and computer program products according to the embodiments of this document. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0041] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0042] The embodiments and / or implementation methods described above are only used to illustrate the preferred embodiments and / or implementation methods of the technology of the present invention, and are not intended to impose any formal restrictions on the implementation methods of the technology of the present invention. Any person skilled in the art may make some changes or modifications to other equivalent embodiments without departing from the scope of the technical means disclosed in the content of the present invention, but they should still be regarded as the technology or embodiments that are essentially the same as the present invention. Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described here, and that various obvious changes, readjustments and substitutions can be made to those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention is described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may also include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A gastric bleeding monitoring system, characterized in that: The system includes the following modules: A collection module, used for collecting data of gastric bleeding indicators to be monitored; A transmission module, connected to the acquisition module, for transmitting data of the gastric bleeding indicators to be monitored acquired by the acquisition module; A data preprocessing module, connected to the transmission module, performs preprocessing operations on the gastric bleeding indicator value data to be monitored; A gastric bleeding monitoring module, connected to the transmission module, for inputting the monitoring index value data preprocessed by the data preprocessing module into the gastric bleeding monitoring model for detecting gastric bleeding; The sample set amplification of the gastric bleeding monitoring model is used to improve the monitoring accuracy of the gastric bleeding monitoring model. The sample set amplification of the gastric bleeding monitoring model is specifically as follows: determining the amplification multiple of the sample set according to the size of the preset number; The alarm module is connected to the gastric bleeding monitoring module and is used to realize an alarm according to the monitoring result of the gastric bleeding monitoring module.

2. A gastric bleeding monitoring system according to claim 1, characterized in that: The collection module is specifically: a hemoglobin sensor is arranged at the front end of the gastric tube to constitute the collection module.

3. A gastric bleeding monitoring system according to claim 1, characterized in that: The calculation formula for the amplification multiple M of the sample set is: Wherein, round() is a rounding function, and N is a preset number of monitoring index value data input into the gastric bleeding monitoring model.

4. A gastric bleeding monitoring system according to claim 3, characterized in that: The operating logic of the alarm module is specifically as follows: if the gastric bleeding monitoring module detects that there is no hemoglobin in the gastric juice, the display screen shows normal and the green light is on; if the gastric bleeding monitoring module detects that there is hemoglobin in the gastric juice, but the content is less than or equal to the set value, the display screen shows suspicious and the yellow light is on; when the gastric bleeding monitoring module detects that the hemoglobin content in the sensed gastric juice is greater than the set value, the red light is on and an alarm sound is issued.

5. A gastric bleeding monitoring system according to claim 1, characterized in that: The preprocessing operation includes outlier processing and missing value processing.

6. A gastric bleeding monitoring system according to claim 5, characterized in that: The missing value processing is specifically: using the mean interpolation method to process the missing values.

7. A gastric bleeding monitoring system according to claim 1, characterized in that: The gastric bleeding monitoring model is an artificial neural network model.

8. A gastric bleeding monitoring system according to claim 7, characterized in that: The structure of the artificial neural network is divided into three parts: input layer, hidden layer and output layer.

9. A gastric bleeding monitoring system according to claim 8, characterized in that: The number of hidden layers is not fixed.

10. A method for monitoring gastric bleeding, characterized in that: The method adopts a gastric bleeding monitoring system according to any one of claims 1 to 9, and the method comprises: S1: Obtaining monitoring index value data in the gastric fluid of the patient to be tested; S2: performing preprocessing operations on the monitoring indicator value data; S3: inputting the monitoring index value data pre-processed by S2 into a gastric bleeding monitoring model for detecting gastric bleeding; Among them, the sample set amplification of the gastric bleeding monitoring model is used to improve the monitoring accuracy of the gastric bleeding monitoring model. The sample set amplification of the gastric bleeding monitoring model is specifically: determining the amplification multiple of the sample set according to the size of the preset number.

Images