Medical image system intelligent management and control method based on MRI (Magnetic Resonance Imaging) technology

Abstract

The invention relates to the technical field of medical imaging, in particular to an intelligent management and control method for a medical imaging system based on an MRI (Magnetic Resonance Imaging) technology. Performing pre-test work on the MRI image system based on the target work task, and acquiring actual work data of each subsystem at a plurality of preset time nodes in a pre-test time period to obtain an actual work data set of each subsystem; comparing and analyzing the actual working data set of each subsystem with a corresponding preset working data set to obtain the working state of each subsystem; and if the working state of a certain subsystem is an abnormal state, further analyzing the subsystem based on the actual working data set of the subsystem to obtain probability values of various abnormal operation events of the subsystem. According to the invention, an efficient and automatic solution can be provided for fault diagnosis of the MRI image system, and the reliability and maintenance efficiency of the system can be improved.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging technology, and in particular to an intelligent management and control method for a medical imaging system based on MRI technology. Background Art

[0002] In the field of modern medicine, magnetic resonance imaging (MRI) technology, as a non-invasive and non-invasive diagnostic tool, has played a key role in many fields such as neuroscience, cardiovascular disease, and tumor detection. However, efficient management and maintenance of MRI equipment is crucial to ensure its stable operation. Traditional management models often rely on manual monitoring and intervention, which is inefficient and error-prone. In order to solve this problem, intelligent management and control methods based on artificial intelligence have emerged in recent years. Through deep learning and big data analysis, such methods can monitor the operating status of MRI systems in real time, predict equipment failures, and even perform preventive maintenance in advance. For example, by collecting multi-source data such as equipment operating parameters, image quality indicators, and user feedback, an intelligent model is constructed. The model will learn the normal behavior pattern of the equipment, and once an abnormality is found, it can automatically trigger an early warning or propose a solution. This intelligent management and control method not only significantly improves the availability of MRI systems, but also reduces operating costs, providing more accurate and efficient equipment management services for medical environments. Summary of the invention

[0003] In order to overcome the shortcomings of the prior art that manual monitoring and intervention are required, which are inefficient and error-prone, the present invention provides an intelligent management and control method for a medical imaging system based on MRI technology.

[0004] To achieve the above object, the present invention discloses an intelligent management and control method for a medical imaging system based on MRI technology, comprising the following steps:

[0005] Obtaining a target work task of the MRI imaging system, and matching the target work task in a work data knowledge base to obtain a preset work data set of each subsystem of the MRI imaging system when executing the target work task;

[0006] and performing a pre-test on the MRI imaging system based on the target work task, and acquiring actual work data of each subsystem at several preset time nodes within the pre-test time period to obtain an actual work data set of each subsystem;

[0007] Compare and analyze the actual working data set of each subsystem with the corresponding preset working data set to obtain the working status of each subsystem;

[0008] If the working state of a certain subsystem is abnormal, the subsystem is further analyzed based on the actual working data set of the subsystem to obtain the probability values ​​of various operating abnormal events occurring in the subsystem.

[0009] Specifically, the target work task of the MRI imaging system is obtained, and the preset work data sets of each subsystem of the MRI imaging system when performing the target work task are matched in the work data knowledge base according to the target work task, specifically:

[0010] Acquire performance characteristic information of the MRI imaging system, and pre-prepare preset working data sets of each subsystem of the MRI imaging system when performing various preset working tasks according to the performance characteristic information;

[0011] Constructing a knowledge base, and importing preset working data sets of each subsystem when the MRI imaging system performs various preset working tasks into the knowledge base, to obtain a working data knowledge base;

[0012] Obtaining the text of the target work task of the MRI imaging system, and the semantic relevance between the text of the target work task and the text of various preset work tasks in the work data knowledge base based on the semantic analysis method;

[0013] The preset work task corresponding to the maximum semantic association is obtained, and the preset work data set of each subsystem of the MRI imaging system when executing the target work task is obtained by pairing the preset work task corresponding to the maximum semantic association in the work data knowledge base.

[0014] Specifically, the actual working data set of each subsystem is compared and analyzed with the corresponding preset working data set to obtain the working status of each subsystem, specifically:

[0015] Calculate the Jaccard similarity between the actual working data set of each subsystem and the corresponding preset working data set respectively;

[0016] Compare the Jaccard similarity between the actual working data set of each subsystem and the corresponding preset working data set with the preset similarity threshold;

[0017] If the Jaccard similarity between the actual working data set of a subsystem and the corresponding preset working data set is greater than the preset similarity threshold, the working state of the subsystem is marked as normal;

[0018] If the Jaccard similarity between the actual working data set of a subsystem and the corresponding preset working data set is not greater than the preset similarity threshold, the working state of the subsystem is marked as an abnormal state.

[0019] Specifically, if the working state of a subsystem is abnormal, the subsystem is further analyzed based on the actual working data set of the subsystem to obtain the probability values ​​of various abnormal operation events occurring in the subsystem, specifically:

[0020] Obtaining operation log information of each subsystem, obtaining various abnormal operation events of each subsystem according to the operation log information, performing feature extraction on the abnormal operation events of each subsystem, and obtaining feature data corresponding to each subsystem when the abnormal operation event occurs;

[0021] A decision tree algorithm is introduced, and several root nodes are generated based on various abnormal operation events, and the characteristic data corresponding to each subsystem when an abnormal operation event occurs is mapped to the corresponding root node to complete the initialization processing of each characteristic data;

[0022] Iteratively split each root node based on the feature data in each root node to obtain several branch nodes; obtain the data status of the feature data in each branch node. If a branch node contains feature data of other categories, continue to iteratively split the branch node until there is no feature data of other categories in the branch node, and the iterative split ends to obtain a leaf node;

[0023] After the iterative splitting of each root node is completed, a decision tree model is constructed according to each root node, and the decision tree model is evaluated and corrected to obtain a corrected decision tree model;

[0024] The genetic algebra is set by a genetic algorithm, and the modified decision tree model is reversely genetically trained according to the characteristic data of each leaf node in each root node in combination with a crossover loss function until the genetic algebra reaches a preset number of times, and a conditional probability table of each root node is output;

[0025] The event probability of various abnormal operation events occurring at each root node is calculated by combining the Markov chain with the conditional probability table; and the probability value of various abnormal operation events occurring at each root node is assigned to each root node according to the event probability of various abnormal operation events occurring at each root node, so as to obtain the final decision tree model.

[0026] Specifically, the decision tree model is evaluated and corrected to obtain a corrected decision tree model, which is specifically:

[0027] For each leaf node in the decision tree model, the information entropy of each leaf node is calculated, and the information entropy represents the degree of confusion of the feature data in the leaf node; and the node features of each leaf node are obtained, and the node features include node purity, node depth, and the number of node feature data;

[0028] The information gain of each leaf node is calculated according to the node features of each leaf node; wherein the information gain represents the degree to which the information entropy of the leaf node is reduced after the feature data is used for splitting;

[0029] The information gain ratio of each leaf node is calculated based on the information gain and information entropy of each leaf node; and the information gain ratio of each leaf node is compared with a preset information gain ratio threshold;

[0030] If the information gain ratio of a leaf node is greater than the preset information gain ratio threshold, it means that there is no redundancy in the leaf node, and there is no need to modify the leaf node;

[0031] If the information gain ratio of a leaf node is greater than the preset information gain ratio threshold, it means that the leaf node does not have redundancy, and the leaf node needs to be corrected and marked as a redundant leaf node;

[0032] Calculate the Jaccard similarity between each redundant leaf node. If the Jaccard similarity between two redundant leaf nodes is greater than a preset threshold, merge the redundant leaf nodes whose Jaccard similarity is greater than the preset threshold.

[0033] If there is a situation where the Jaccard similarity between a redundant leaf node and each redundant leaf node is not greater than the preset threshold, the redundant leaf node is pruned to complete the pruning correction operation;

[0034] The final state of each leaf node in the decision tree model is obtained to obtain the revised decision tree model.

[0035] The following steps are also included:

[0036] Acquire an actual working data set of a subsystem whose working state is marked as an abnormal state; map all actual working data in the actual working data set to corresponding leaf nodes in the final decision tree model;

[0037] Obtain the total amount of actual work data mapped on each root node in the final decision tree model, and mark the root node with the largest total amount of actual work data mapped;

[0038] The probability values ​​of various abnormal operation events occurring at the marked root node are obtained, and the probability values ​​of various abnormal operation events occurring are sent to a preset terminal for display.

[0039] To achieve the above object, the present invention further discloses an intelligent management and control system for a medical imaging system based on MRI technology, wherein the intelligent management and control system for a medical imaging system comprises a memory and a processor, wherein a program for an intelligent management and control method for a medical imaging system is stored in the memory, and when the program for an intelligent management and control method for a medical imaging system is executed by the processor, the following steps are implemented:

[0040] Obtaining a target work task of the MRI imaging system, and matching the target work task in a work data knowledge base to obtain a preset work data set of each subsystem of the MRI imaging system when executing the target work task;

[0041] and performing a pre-test on the MRI imaging system based on the target work task, and acquiring actual work data of each subsystem at several preset time nodes within the pre-test time period to obtain an actual work data set of each subsystem;

[0042] Compare and analyze the actual working data set of each subsystem with the corresponding preset working data set to obtain the working status of each subsystem;

[0043] If the working state of a certain subsystem is abnormal, the subsystem is further analyzed based on the actual working data set of the subsystem to obtain the probability values ​​of various operating abnormal events occurring in the subsystem.

[0044] To achieve the above-mentioned objectives, the present invention also discloses a computer-readable storage medium, which includes a medical imaging system intelligent management and control method program. When the medical imaging system intelligent management and control method program is executed by a processor, the steps of any one of the medical imaging system intelligent management and control methods are implemented.

[0045] The present invention solves the technical defects existing in the background technology, and has the following beneficial effects: obtaining the target work task of the MRI imaging system, matching the target work task in the work data knowledge base to obtain the preset work data set of each subsystem of the MRI imaging system when executing the target work task; and pre-testing the MRI imaging system based on the target work task, and obtaining the actual work data of each subsystem at several preset time nodes within the pre-test time period to obtain the actual work data set of each subsystem; comparing and analyzing the actual work data set of each subsystem with the corresponding preset work data set to obtain the working status of each subsystem; if the working status of a subsystem is an abnormal state, further analyzing the subsystem based on the actual work data set of the subsystem to obtain the probability value of various abnormal operation events of the subsystem. The present invention can provide an efficient and automated solution for fault diagnosis of MRI imaging systems, which helps to improve the reliability and maintenance efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0047] Figure 1 It is a flow chart of the overall method of intelligent control method of medical imaging system based on MRI technology;

[0048] Figure 2 It is a partial method flow chart of an intelligent control method of a medical imaging system based on MRI technology;

[0049] Figure 3 This is a system block diagram of an intelligent management and control system for medical imaging systems based on MRI technology. DETAILED DESCRIPTION

[0050] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0051] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0052] like Figure 1 As shown, the present invention discloses an intelligent management and control method for a medical imaging system based on MRI technology, comprising the following steps:

[0053] S102: Acquire a target work task of the MRI imaging system, and match the target work task in a work data knowledge base to obtain a preset work data set of each subsystem of the MRI imaging system when executing the target work task;

[0054] S104: performing a pre-test on the MRI imaging system based on the target work task, and acquiring actual work data of each subsystem at a number of preset time nodes within the pre-test time period to obtain an actual work data set of each subsystem;

[0055] S106: Compare and analyze the actual working data set of each subsystem with the corresponding preset working data set to obtain the working status of each subsystem;

[0056] S108: If the working state of a certain subsystem is an abnormal state, further analyze the subsystem based on the actual working data set of the subsystem to obtain the probability values ​​of various operating abnormal events occurring in the subsystem.

[0057] It should be noted that the target work tasks of the MRI imaging system are formulated by relevant obligated personnel based on information such as the patient's lesion location and lesion condition, such as the sampling frequency, sampling path, sampling time and other information of the MRI imaging system.

[0058] Specifically, the target work task of the MRI imaging system is obtained, and the preset work data sets of each subsystem of the MRI imaging system when performing the target work task are matched in the work data knowledge base according to the target work task, specifically:

[0059] Acquire performance characteristic information of the MRI imaging system, and pre-prepare preset working data sets of each subsystem of the MRI imaging system when performing various preset working tasks according to the performance characteristic information;

[0060] Constructing a knowledge base, and importing preset working data sets of each subsystem when the MRI imaging system performs various preset working tasks into the knowledge base, to obtain a working data knowledge base;

[0061] Obtaining the text of the target work task of the MRI imaging system, and the semantic relevance between the text of the target work task and the text of various preset work tasks in the work data knowledge base based on the semantic analysis method;

[0062] The preset work task corresponding to the maximum semantic association is obtained, and the preset work data set of each subsystem of the MRI imaging system when executing the target work task is obtained by pairing the preset work task corresponding to the maximum semantic association in the work data knowledge base.

[0063] Among them, the MRI imaging system usually includes the following subsystems:

[0064] Magnet system, where the working data includes magnetic field strength, magnetic field uniformity, etc.;

[0065] Gradient system, whose working data include gradient strength, gradient switching rate, etc.;

[0066] RF system, whose working data includes RF power, frequency, etc.;

[0067] Control system, whose working data includes scanning parameters, etc.;

[0068] Image processing system, whose working data includes image contrast, resolution, etc.;

[0069] Cooling system, whose working data includes temperature, flow rate, etc.;

[0070] Power supply system, whose working data includes voltage, current, etc.

[0071] It should be noted that the performance characteristic information of the MRI imaging system refers to a series of key parameters and performance indicators exhibited by the system during the imaging process. This information includes but is not limited to image clarity (resolution), signal-to-noise ratio, magnetic field strength, scanning speed, magnetic field uniformity, dynamic range, and spatial and frequency encoding characteristics. These characteristics directly affect the quality, details, and diagnostic accuracy of the image. For example, high resolution means that smaller tissue structures can be captured, and a high signal-to-noise ratio means that the signal in the image is clearer and has less interference. In addition, the performance characteristics also cover the stability, consistency, and adaptability of the system to different types of imaging sequences. By understanding and analyzing this information, the performance of the MRI system can be evaluated, scanning parameters can be optimized, and its effectiveness in medical diagnosis can be ensured.

[0072] It should be noted that, through the performance characteristic information of the system, combined with expert experience, the preset working data sets of each subsystem are prefabricated in advance, and a knowledge base is constructed to realize intelligent management and control of the MRI imaging system when performing preset working tasks. Specifically, the method first presets the working data sets of each subsystem according to the performance characteristic information of the MRI system, and then matches the target working task with the preset working task in the knowledge base through semantic analysis, so as to obtain the preset working data set that matches the target task. By constructing a knowledge base and using semantic analysis, intelligent management of MRI system task execution is realized, and the automation level of the system is improved. Using semantic relevance analysis, the target task and the preset task can be accurately matched to ensure the accuracy of task execution. The prefabricated working data set can reduce on-site configuration time and improve the working efficiency of the MRI system.

[0073] Specifically, the actual working data set of each subsystem is compared and analyzed with the corresponding preset working data set to obtain the working status of each subsystem, specifically:

[0074] Calculate the Jaccard similarity between the actual working data set of each subsystem and the corresponding preset working data set respectively;

[0075] Compare the Jaccard similarity between the actual working data set of each subsystem and the corresponding preset working data set with the preset similarity threshold;

[0076] If the Jaccard similarity between the actual working data set of a subsystem and the corresponding preset working data set is greater than the preset similarity threshold, the working state of the subsystem is marked as normal;

[0077] If the Jaccard similarity between the actual working data set of a subsystem and the corresponding preset working data set is not greater than the preset similarity threshold, the working state of the subsystem is marked as an abnormal state.

[0078] It should be noted that by calculating the Jaccard similarity between the actual working data set and the preset working data set of each subsystem and comparing it with the preset similarity threshold, it is determined whether the working state of the subsystem is normal. If the Jaccard similarity between the actual working data set and the preset working data set is greater than the threshold, the subsystem is considered to be in a normal state; otherwise, the subsystem is considered to be in an abnormal state. The automated comparison process reduces manual intervention and improves the efficiency and accuracy of diagnosis. Based on the comparison between the actual data set and the preset data set, data support is provided for system maintenance and optimization. Using Jaccard similarity as an evaluation indicator, intelligent monitoring and diagnosis of the working state of the MRI imaging system subsystem is realized, which helps to improve the operating efficiency and stability of the system.

[0079] like Figure 2 Specifically, if the working state of a subsystem is abnormal, the subsystem is further analyzed based on the actual working data set of the subsystem to obtain the probability values ​​of various abnormal operation events occurring in the subsystem, specifically:

[0080] S202: Acquire operation log information of each subsystem, acquire various abnormal operation events of each subsystem according to the operation log information, extract features of the abnormal operation events of each subsystem, and obtain feature data corresponding to the abnormal operation events of each subsystem;

[0081] The operation log information of each subsystem in the MRI imaging system refers to the operation status, parameter settings, operation records and events of each subsystem at each time point, including parameter values ​​and executed instructions during normal operation, as well as error messages and warning prompts when abnormalities occur. This information can help analyze the system operation status, locate the cause of the fault, and perform diagnosis and maintenance;

[0082] S204: Introducing a decision tree algorithm, generating a number of root nodes based on various abnormal operation events, and mapping the feature data corresponding to each subsystem when an abnormal operation event occurs to the corresponding root node, so as to complete initialization processing of each feature data;

[0083] S206: performing iterative splitting processing on each root node based on the feature data in each root node to obtain a plurality of branch nodes; obtaining the data status of the feature data in each branch node, and if a branch node has feature data of other categories, continuing to iteratively split the branch node until there is no feature data of other categories in the branch node, the iterative splitting ends, and a leaf node is obtained;

[0084] S208: After the iterative splitting of each root node is completed, a decision tree model is constructed according to each root node, and the decision tree model is evaluated and corrected to obtain a corrected decision tree model;

[0085] S210: Setting the genetic generation through a genetic algorithm, and performing reverse genetic training on the modified decision tree model according to the characteristic data of each leaf node in each root node in combination with a crossover loss function, until the genetic generation reaches a preset number of times, and outputting a conditional probability table of each root node;

[0086] S212: Calculate the event probability of various abnormal operation events occurring at each root node by combining the Markov chain and the conditional probability table; assign probability values ​​of various abnormal operation events occurring at each root node according to the event probability of various abnormal operation events occurring at each root node, and obtain a final decision tree model.

[0087] It should be noted that, first, the system collects the operation logs of each subsystem and extracts the characteristic data of abnormal operation events. Then, the decision tree algorithm is used to process and classify these characteristic data to construct a decision tree model that can identify different abnormal patterns. Then, the model is optimized by genetic algorithm to further improve the accuracy of diagnosis. Finally, the probability of abnormal operation of each subsystem is calculated in combination with Markov chain to obtain the final decision tree model. Through the decision tree algorithm, the system can accurately identify the abnormal operation of the MRI subsystem and provide reliable data support for fault diagnosis; the fault diagnosis model can be continuously optimized to improve the self-learning and adaptability of the system; and combined with the Markov chain to calculate the probability of abnormal operation, the system can predict potential fault risks and take preventive measures in time, making fault diagnosis and processing faster and more efficient. In general, this method realizes intelligent monitoring and management of MRI imaging systems by constructing an effective and reliable fault diagnosis model, reduces dependence on manual intervention, and improves the level of automation of management.

[0088] Specifically, the decision tree model is evaluated and corrected to obtain a corrected decision tree model, which is specifically:

[0089] For each leaf node in the decision tree model, the information entropy of each leaf node is calculated, and the information entropy represents the degree of confusion of the feature data in the leaf node; and the node features of each leaf node are obtained, and the node features include node purity, node depth, and the number of node feature data;

[0090] The information gain of each leaf node is calculated according to the node features of each leaf node; wherein the information gain represents the degree to which the information entropy of the leaf node is reduced after the feature data is used for splitting;

[0091] The information gain ratio of each leaf node is calculated based on the information gain and information entropy of each leaf node; and the information gain ratio of each leaf node is compared with a preset information gain ratio threshold;

[0092] If the information gain ratio of a leaf node is greater than the preset information gain ratio threshold, it means that there is no redundancy in the leaf node, and there is no need to modify the leaf node;

[0093] If the information gain ratio of a leaf node is greater than the preset information gain ratio threshold, it means that the leaf node does not have redundancy, and the leaf node needs to be corrected and marked as a redundant leaf node;

[0094] Calculate the Jaccard similarity between each redundant leaf node. If the Jaccard similarity between two redundant leaf nodes is greater than a preset threshold, merge the redundant leaf nodes whose Jaccard similarity is greater than the preset threshold.

[0095] If there is a situation where the Jaccard similarity between a redundant leaf node and each redundant leaf node is not greater than the preset threshold, the redundant leaf node is pruned to complete the pruning correction operation;

[0096] The final state of each leaf node in the decision tree model is obtained to obtain the revised decision tree model.

[0097] It should be noted that when the redundancy of leaf nodes obtained by iterative splitting in the decision tree model is too high, the decision tree model used for fault diagnosis will have the following defects: First, redundant leaf nodes will make the model too complex and difficult to understand and explain. Too many leaf nodes mean that the structure of the decision tree is too large and has too many branches, making it difficult to intuitively grasp the decision-making process and logical relationship of the model. This will reduce the interpretability of the model and the reliability of the prediction results, making it difficult to perform fault diagnosis and analysis. Secondly, redundant leaf nodes will cause the model to overfit the training data and reduce the generalization ability. Too many leaf nodes mean that the model pays too much attention to the details in the training data and ignores the general rules in the data. When encountering new data, the model may not be able to accurately predict the fault type, resulting in inaccurate diagnosis results. In addition, redundant leaf nodes will also reduce the efficiency of the model. Too many leaf nodes will increase the amount of calculation of the model and slow down the prediction speed, which will affect the real-time and efficiency of fault diagnosis and is not conducive to timely detection and resolution of faults. Therefore, when building a decision tree model, it is necessary to avoid redundant leaf nodes.

[0098] It should be noted that the information entropy of each leaf node in the decision tree is first calculated to evaluate the degree of confusion of the data in the leaf node. Then, the node features of the leaf node, including node purity, depth and number of feature data, are obtained, and the information gain of each leaf node is calculated accordingly. The information gain reflects the degree to which the information entropy of the leaf node is reduced after splitting using the feature data. By comparing the information gain of each leaf node with the preset information gain ratio threshold, redundant leaf nodes can be identified. If the similarity between two redundant leaf nodes exceeds the preset threshold, they are merged to reduce redundancy. For special leaf nodes that are not highly similar to other redundant leaf nodes, pruning is performed. Finally, through these steps, a modified decision tree model can be obtained, in which the status of each leaf node is updated. This method can effectively reduce the complexity of the decision tree and improve the generalization ability of the model. Through pruning, the decision tree can avoid overfitting of the training data, so that the model can also perform well on new and unknown data. In addition, by merging similar redundant leaf nodes, the model structure can be simplified, and the interpretability and maintenance efficiency of the model can be improved. At the same time, this method can also reduce the consumption of computing resources and speed up the prediction speed of the model, which is particularly important for application scenarios that require real-time or near real-time fault diagnosis. In general, the present invention can effectively improve the performance and practicality of decision tree models in applications such as fault diagnosis by pruning decision trees.

[0099] The following steps are also included:

[0100] Acquire an actual working data set of a subsystem whose working state is marked as an abnormal state; map all actual working data in the actual working data set to corresponding leaf nodes in the final decision tree model;

[0101] Obtain the total amount of actual work data mapped on each root node in the final decision tree model, and mark the root node with the largest total amount of actual work data mapped;

[0102] The probability values ​​of various abnormal operation events occurring at the marked root node are obtained, and the probability values ​​of various abnormal operation events occurring are sent to a preset terminal for display.

[0103] It should be noted that the actual working data set of the subsystem under abnormal conditions is collected and mapped to the corresponding leaf nodes of the final decision tree model. Next, the total number of actual working data mapped on each root node in the decision tree is counted, and the root node with the most data is marked. Finally, the probability values ​​of various abnormal operation events occurring at the root node are obtained, and these probability values ​​are displayed on the preset terminal. By automatically mapping the actual working data to the decision tree model, the need for manual analysis is reduced, the efficiency of fault diagnosis is improved, and a quantitative fault risk assessment can be provided for maintenance personnel. The probability of abnormal events is displayed in real time on the terminal, which is convenient for operators to promptly discover and respond to potential system problems, and provide decision support for system maintenance and management personnel. It provides an efficient and automated solution for subsystem fault diagnosis, which helps to improve system reliability and maintenance efficiency.

[0104] In addition, the method further comprises the following steps:

[0105] Obtaining radio frequency strength values ​​of nodes at each preset position in the radio frequency system, and constructing an actual radio frequency strength distribution map of the radio frequency system according to the radio frequency strength values ​​of nodes at each preset position;

[0106] and obtaining a preset radio frequency intensity distribution map of the radio frequency system, and dividing the actual radio frequency intensity distribution map and the preset radio frequency intensity distribution map into a plurality of radio frequency nodes;

[0107] Calculating the structural similarity index between the same radio frequency nodes in the actual radio frequency intensity distribution map and the preset radio frequency intensity distribution map, and constructing a structural similarity index matrix according to the calculated structural similarity index;

[0108] Traversing all elements in the structural similarity index matrix, and comparing all elements in the structural similarity index matrix with preset values;

[0109] If an element in the structural similarity index matrix is ​​greater than a preset value, the corresponding RF node in the actual RF intensity distribution map is rendered black; if an element in the structural similarity index matrix is ​​not greater than a preset value, the corresponding RF node in the actual RF intensity distribution map is rendered white;

[0110] Obtaining area values ​​of black areas and white areas in the actual radio frequency intensity distribution map, and calculating similarity between the actual radio frequency intensity distribution map and the preset radio frequency intensity distribution map according to the area values ​​of the black areas and white areas in the actual radio frequency intensity distribution map;

[0111] If the similarity between the actual RF intensity distribution map and the preset RF intensity distribution map is not greater than the preset similarity, it means that the performance of the RF system does not meet the standard, and the rendered actual RF intensity distribution map is transmitted to the preset terminal for display.

[0112] It should be noted that the RF strength values ​​of each preset position node in the RF system are first obtained, and the actual RF strength distribution map is constructed. Then, the actual distribution map is compared with the preset RF strength distribution map, and the structural similarity index between the same RF nodes is calculated by dividing the two into several RF nodes, and the corresponding matrix is ​​constructed. By traversing this matrix and comparing it with the preset value, it can be determined which areas have RF strengths that are consistent with expectations and which are not. The matching areas (i.e., areas with a structural similarity index greater than the preset value) are rendered black in the actual RF strength distribution map, otherwise they are rendered white. By calculating the area of ​​the black and white areas, the similarity between the actual RF strength distribution map and the preset RF strength distribution map can be evaluated. If the similarity does not meet the standard, it is considered that the RF system performance does not meet the standard, and the result is sent to the preset terminal for display. This method provides a means of quantitatively evaluating the performance of the RF system. By visually displaying the conformance of the RF intensity distribution, it can quickly identify possible problem areas in the RF system. These areas may be signs of signal transmission problems, equipment failures, or performance degradation, thereby discovering potential problems as early as possible and making targeted optimizations and adjustments, such as adjusting the antenna position, improving the signal path, or upgrading the hardware, thereby improving the performance of the entire system.

[0113] In addition, the method further comprises the following steps:

[0114] Obtain the probability values ​​of various abnormal operation events occurring in the subsystem whose working state is abnormal, and mark the abnormal operation events whose probability values ​​are greater than the preset probability values ​​as Class I abnormal operation events; mark the abnormal operation events whose probability values ​​are not greater than the preset probability values ​​as Class II abnormal operation events;

[0115] Based on the operation log information, the historical fault parts corresponding to a type of abnormal operation event caused by the subsystem in the historical working process are defined as suspicious parts, and the real-time electrical parameter information of the suspicious parts is obtained based on the Internet of Things;

[0116] Importing the real-time electrical parameter information of the suspicious part into the Markov model to perform fault deduction, and obtaining the state transition probability of the suspicious part;

[0117] If the state transition probability of the suspicious part is greater than the preset probability value, the suspicious part is marked as a faulty part.

[0118] It should be noted that the Markov model is a mathematical tool that is based on the transition probability between states and is used to describe the dynamic process of a system changing over time. In fault deduction, if we regard the working state of the part as each state in the Markov chain, then each state has a specific probability of transitioning from the adjacent state (such as from normal to a minor fault, or from a minor fault to a serious fault). These probabilities reflect the possibility of a part transitioning from one state to another. For suspicious parts, by collecting their electrical parameter data under different working conditions, the relationship between these parameter changes and state transitions is analyzed. For example, electrical parameters may show certain patterns before a fault occurs. Then, through statistical learning and modeling, the transition probability from one electrical parameter value to another, or from one working state to another, can be calculated. In this way, it is possible to predict the possible fault state and its possibility of the part under given conditions, so as to carry out effective fault prediction and management. Through this method, the location of the faulty part that caused the fault can be further located based on the probability values ​​of various abnormal operation events occurring in the subsystem.

[0119] To achieve the above purpose, if Figure 3 As shown, the present invention also discloses an intelligent management and control system for a medical imaging system based on MRI technology, the intelligent management and control system for a medical imaging system comprises a memory 20 and a processor 60, the memory 20 stores a program for an intelligent management and control method for a medical imaging system, and when the program for an intelligent management and control method for a medical imaging system is executed by the processor 60, the following steps are implemented:

[0120] Obtaining a target work task of the MRI imaging system, and matching the target work task in a work data knowledge base to obtain a preset work data set of each subsystem of the MRI imaging system when executing the target work task;

[0121] and performing a pre-test on the MRI imaging system based on the target work task, and acquiring actual work data of each subsystem at several preset time nodes within the pre-test time period to obtain an actual work data set of each subsystem;

[0122] Compare and analyze the actual working data set of each subsystem with the corresponding preset working data set to obtain the working status of each subsystem;

[0123] If the working state of a certain subsystem is abnormal, the subsystem is further analyzed based on the actual working data set of the subsystem to obtain the probability values ​​of various operating abnormal events occurring in the subsystem.

[0124] To achieve the above-mentioned objectives, the present invention also discloses a computer-readable storage medium, which includes a medical imaging system intelligent management and control method program. When the medical imaging system intelligent management and control method program is executed by a processor, the steps of any one of the medical imaging system intelligent management and control methods are implemented.

[0125] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0126] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0127] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0128] Those skilled in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.

[0129] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods of each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0130] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent management and control method for a medical imaging system based on MRI technology, characterized in that: The following steps are involved: Obtaining a target work task of the MRI imaging system, and matching the target work task in a work data knowledge base to obtain a preset work data set of each subsystem of the MRI imaging system when executing the target work task; and performing a pre-test on the MRI imaging system based on the target work task, and acquiring actual work data of each subsystem at several preset time nodes within the pre-test time period to obtain an actual work data set of each subsystem; Compare and analyze the actual working data set of each subsystem with the corresponding preset working data set to obtain the working status of each subsystem; If the working state of a certain subsystem is abnormal, the subsystem is further analyzed based on the actual working data set of the subsystem to obtain the probability values ​​of various operating abnormal events occurring in the subsystem.

2. According to claim 1, a medical imaging system intelligent management and control method based on MRI technology is characterized in that: The target work task of the MRI imaging system is obtained, and the preset work data sets of each subsystem of the MRI imaging system when performing the target work task are matched in the work data knowledge base according to the target work task, specifically: Acquire performance characteristic information of the MRI imaging system, and pre-prepare preset working data sets of each subsystem of the MRI imaging system when performing various preset working tasks according to the performance characteristic information; Constructing a knowledge base, and importing preset working data sets of each subsystem when the MRI imaging system performs various preset working tasks into the knowledge base, to obtain a working data knowledge base; Obtaining the text of the target work task of the MRI imaging system, and the semantic relevance between the text of the target work task and the text of various preset work tasks in the work data knowledge base based on the semantic analysis method; The preset work task corresponding to the maximum semantic association is obtained, and the preset work data set of each subsystem of the MRI imaging system when executing the target work task is obtained by pairing the preset work task corresponding to the maximum semantic association in the work data knowledge base.

3. The intelligent management and control method of a medical imaging system based on MRI technology according to claim 1, characterized in that: Compare and analyze the actual working data set of each subsystem with the corresponding preset working data set to obtain the working status of each subsystem, specifically: Calculate the Jaccard similarity between the actual working data set of each subsystem and the corresponding preset working data set respectively; Compare the Jaccard similarity between the actual working data set of each subsystem and the corresponding preset working data set with the preset similarity threshold; If the Jaccard similarity between the actual working data set of a subsystem and the corresponding preset working data set is greater than the preset similarity threshold, the working state of the subsystem is marked as normal; If the Jaccard similarity between the actual working data set of a subsystem and the corresponding preset working data set is not greater than the preset similarity threshold, the working state of the subsystem is marked as an abnormal state.

4. The intelligent management and control method of a medical imaging system based on MRI technology according to claim 1, characterized in that: If the working state of a subsystem is abnormal, the subsystem is further analyzed based on the actual working data set of the subsystem to obtain the probability values ​​of various abnormal operation events occurring in the subsystem, specifically: Obtaining operation log information of each subsystem, obtaining various abnormal operation events of each subsystem according to the operation log information, performing feature extraction on the abnormal operation events of each subsystem, and obtaining feature data corresponding to each subsystem when the abnormal operation event occurs; A decision tree algorithm is introduced, and several root nodes are generated based on various abnormal operation events, and the characteristic data corresponding to each subsystem when an abnormal operation event occurs is mapped to the corresponding root node to complete the initialization processing of each characteristic data; Iteratively split each root node based on the feature data in each root node to obtain several branch nodes; obtain the data status of the feature data in each branch node. If a branch node contains feature data of other categories, continue to iteratively split the branch node until there is no feature data of other categories in the branch node, and the iterative split ends to obtain a leaf node; After the iterative splitting of each root node is completed, a decision tree model is constructed according to each root node, and the decision tree model is evaluated and corrected to obtain a corrected decision tree model; The genetic algebra is set by a genetic algorithm, and the modified decision tree model is reversely genetically trained according to the characteristic data of each leaf node in each root node in combination with a crossover loss function until the genetic algebra reaches a preset number of times, and a conditional probability table of each root node is output; The event probability of various abnormal operation events occurring at each root node is calculated by combining the Markov chain with the conditional probability table; and the probability value of various abnormal operation events occurring at each root node is assigned to each root node according to the event probability of various abnormal operation events occurring at each root node, so as to obtain the final decision tree model.

5. The intelligent management and control method of medical imaging system based on MRI technology according to claim 4, characterized in that: The decision tree model is evaluated and corrected to obtain a corrected decision tree model, specifically: For each leaf node in the decision tree model, the information entropy of each leaf node is calculated, and the information entropy represents the degree of confusion of the feature data in the leaf node; and the node features of each leaf node are obtained, and the node features include node purity, node depth, and the number of node feature data; The information gain of each leaf node is calculated according to the node features of each leaf node; wherein the information gain represents the degree to which the information entropy of the leaf node is reduced after the feature data is used for splitting; The information gain ratio of each leaf node is calculated based on the information gain and information entropy of each leaf node; and the information gain ratio of each leaf node is compared with a preset information gain ratio threshold; If the information gain ratio of a leaf node is greater than the preset information gain ratio threshold, it means that there is no redundancy in the leaf node, and there is no need to modify the leaf node; If the information gain ratio of a leaf node is greater than the preset information gain ratio threshold, it means that the leaf node does not have redundancy, and the leaf node needs to be corrected and marked as a redundant leaf node; Calculate the Jaccard similarity between each redundant leaf node. If the Jaccard similarity between two redundant leaf nodes is greater than a preset threshold, merge the redundant leaf nodes whose Jaccard similarity is greater than the preset threshold. If there is a situation where the Jaccard similarity between a redundant leaf node and each redundant leaf node is not greater than the preset threshold, the redundant leaf node is pruned to complete the pruning correction operation; The final state of each leaf node in the decision tree model is obtained to obtain the revised decision tree model.

6. The intelligent management and control method of a medical imaging system based on MRI technology according to claim 4, characterized in that: The following steps are also included: Acquire an actual working data set of a subsystem whose working state is marked as an abnormal state; map all actual working data in the actual working data set to corresponding leaf nodes in the final decision tree model; Obtain the total amount of actual work data mapped on each root node in the final decision tree model, and mark the root node with the largest total amount of actual work data mapped; The probability values ​​of various abnormal operation events occurring at the marked root node are obtained, and the probability values ​​of various abnormal operation events occurring are sent to a preset terminal for display.

7. An intelligent management and control system for medical imaging systems based on MRI technology, characterized in that: The medical imaging system intelligent management and control system comprises a memory and a processor, wherein the memory stores a medical imaging system intelligent management and control method program, and when the medical imaging system intelligent management and control method program is executed by the processor, the following steps are implemented: Obtaining a target work task of the MRI imaging system, and matching the target work task in a work data knowledge base to obtain a preset work data set of each subsystem of the MRI imaging system when executing the target work task; and performing a pre-test on the MRI imaging system based on the target work task, and acquiring actual work data of each subsystem at several preset time nodes within the pre-test time period to obtain an actual work data set of each subsystem; Compare and analyze the actual working data set of each subsystem with the corresponding preset working data set to obtain the working status of each subsystem; If the working state of a certain subsystem is abnormal, the subsystem is further analyzed based on the actual working data set of the subsystem to obtain the probability values ​​of various operating abnormal events occurring in the subsystem.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a medical imaging system intelligent management and control method program. When the medical imaging system intelligent management and control method program is executed by a processor, the steps of the medical imaging system intelligent management and control method according to any one of claims 1 to 6 are implemented.

Images