Model training method and device, injection parameter generation method and device and electronic equipment

Abstract

The invention discloses a model training method and an injection parameter generation method, the model training method is used for training an injection parameter generation model, a multiple linear regression model is used as a training network, and clinical characteristic parameters of a patient are used as input of the network for training. The training network learns the mapping relation between the injection parameters and the clinical characteristic parameters, an injection parameter prediction value is generated, a loss function of the training network is established according to the injection parameter labels of the training samples and the injection parameter prediction value, and therefore network parameters are updated to complete model training. Clinical characteristic parameters of a patient are input into the model to generate target injection parameters, reference can be provided for injection operators before injection, and compared with an existing single model, the method provided by the invention considers the coupling relation among multiple parameters, and reflects the influence of the multiple parameters on the injection parameters through the multiple linear regression model.

Description

Technical Field

[0001] The present invention belongs to the field of medical devices, and in particular relates to a model training method, an injection parameter generating method, a device and an electronic device. Background Art

[0002] High-pressure injection technology has important application value in many fields such as medicine and industry. For example, in scenarios such as medical imaging, contrast agent infusion, drug delivery, and industrial material injection, high-pressure injection can ensure that the liquid reaches the target location quickly and accurately, thereby improving the effect of detection, treatment, or processing. In these applications, the determination of reasonable injection parameters (such as injection dose, flow rate, pressure, etc.) is of decisive significance for ensuring operational safety, improving injection efficiency, and reducing risks.

[0003] In practical applications, individual differences in patients or process systems (such as physiological parameters, medical history, allergy history, etc.) often make injection parameters have complex nonlinear relationships. How to accurately predict and generate injection parameters that meet various requirements has always been a key technical issue that needs to be solved in this field.

[0004] Traditional injection systems mostly rely on expert experience and preset rules. The injection parameters are determined by operators manually setting them or using simple mathematical formulas. They can meet basic needs to a certain extent, but their robustness and personalized control capabilities are relatively limited when faced with complex and changeable clinical or industrial conditions.

[0005] With the application of statistical modeling, some existing technologies have begun to adopt models for prediction. Although a single model is used to predict the target parameters, data-driven parameter recommendations have been achieved to a certain extent, they often face the problem of insufficient modeling accuracy when dealing with the coupling effects between multiple parameters. There is an inherent coupling relationship between multiple parameters in the injection process, and a single model is difficult to capture the mutual influence of these parameters at the same time, which may lead to problems such as mismatched parameter predictions or inconsistent overall effects in practical applications. Summary of the invention

[0006] Based on this, the present invention aims to propose a model training method, an injection parameter generation method, a device and an electronic device, which adopts a multivariate linear regression model as the framework of the parameter generation model, and uses the patient's individual information to train the model, so that the model can generate injection parameters according to the input data when it is actually used, providing a scientific reference for injection operators.

[0007] In a first aspect, the present invention provides a model training method, wherein the trained model is used to generate injection parameters, comprising:

[0008] A training set is obtained, where the training set includes clinical characteristic parameters with injection parameter labels;

[0009] A multiple linear regression model is used to construct a training network. The clinical feature parameters are used as the input of the training network for training, enabling the training network to learn the mapping relationship between the injection parameters and the clinical feature parameters, and generating predicted values of the injection parameters.

[0010] Based on the injection parameter labels and the predicted values of the injection parameters, a loss function of the training network is established. The network parameters of the training network are updated according to the loss function, and the trained training network is used as an injection parameter generation model.

[0011] Furthermore, the mathematical expression of the injection parameter generation model is as follows:

[0012] ,

[0013] where, represents the injection parameters, represents the input matrix composed of the clinical feature parameters, represents the transpose of the matrix, represents the vector of partial regression coefficients.

[0014] Furthermore, the loss function of the training network established based on the injection parameter labels and the predicted values of the injection parameters includes:

[0015] Using the mean square error as the loss function, the loss function is expressed as follows:

[0016]

[0017] where, represents the number of training samples in the training set, represents the injection parameter label of the i-th training sample, represents the predicted value of the injection parameter of the i-th training sample.

[0018] Furthermore, updating the network parameters of the training network according to the loss function includes:

[0019] Taking the minimization of the loss function as the optimization objective, the batch gradient descent algorithm is used to update the network parameters of the training network.

[0020] Furthermore, the clinical feature parameters at least include patient physiological parameters, patient past medical history, and patient treatment parameters.

[0021] Furthermore, the patient treatment parameters include the first injection parameter. Using the clinical feature parameters as the input of the training network for training, enabling the training network to learn the mapping relationship between the injection parameters and the clinical feature parameters, and generating the predicted values of the injection parameters includes:

[0022] Use the first injection parameter as the input to train the network, so that the trained network learns the mapping relationship between the first injection parameter and the second injection parameter, and generates a predicted value of the second injection parameter.

[0023] In a second aspect, the present invention provides an injection parameter generation method, including:

[0024] Obtain the clinical characteristic parameters of the patient;

[0025] Input the clinical characteristic parameters into the injection parameter generation model trained in the first aspect to generate injection parameters corresponding to the clinical characteristic parameters.

[0026] Further, inputting the clinical characteristic parameters into the injection parameter generation model trained in the first aspect to generate injection parameters corresponding to the clinical characteristic parameters includes:

[0027] Input the first clinical characteristic parameter into the first injection parameter generation model trained in the first aspect to generate the first injection parameter corresponding to the first clinical characteristic parameter;

[0028] Combine the first injection parameter and the first clinical characteristic parameter into the second clinical characteristic parameter, and input the second clinical characteristic parameter into the second injection parameter generation model trained in the first aspect to generate the second injection parameter corresponding to the second clinical characteristic parameter.

[0029] In a third aspect, the present invention provides a model training device for training an injection parameter generation model, including:

[0030] A training set acquisition module for acquiring a training set, where the training set includes clinical characteristic parameters with injection parameter labels;

[0031] A network training module for constructing a training network using a multiple linear regression model, training with the clinical characteristic parameters as the input of the training network, so that the training network learns the mapping relationship between the injection parameters and the clinical characteristic parameters, and generates injection parameter predicted values;

[0032] A network update module for establishing a loss function of the training network based on the injection parameter labels and the injection parameter predicted values, updating the network parameters of the training network according to the loss function, and using the trained training network as the injection parameter generation model.

[0033] In a fourth aspect, the present invention provides an injection parameter generation device, including:

[0034] A clinical parameter acquisition module for acquiring the clinical characteristic parameters of the patient;

[0035] An injection parameter generation module for inputting the clinical characteristic parameters into the injection parameter generation model trained by the device in the third aspect to generate injection parameters corresponding to the clinical characteristic parameters.

[0036] In a fifth aspect, the present invention provides an electronic device, including a memory storing computer-executable instructions and a processor. When the computer-executable instructions are executed by the processor, the device performs the model training method provided in the first aspect and / or each step of the injection parameter generation method provided in the second aspect.

[0037] In a sixth aspect, the present invention provides a readable storage medium storing a computer-executable program. When the program is executed, it can implement each step of the model training method provided in the first aspect and / or the injection parameter generation method provided in the second aspect.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] The present invention provides a model training method and an injection parameter generation method. The model training method is used to train an injection parameter generation model. A multiple linear regression model is used as the training network, and the clinical characteristic parameters of patients are used as the input of the network for training, so that the training network learns the mapping relationship between injection parameters and clinical characteristic parameters, generates injection parameter prediction values, and establishes a loss function of the training network according to the injection parameter labels and injection parameter prediction values of the training samples, thereby updating the network parameters to complete model training. When the model is actually used for injection parameter generation, only the clinical characteristic parameters of the patient need to be input into the model to generate the target injection parameters. The generated values can provide reference for injection operators before injection. Compared with the existing single model, the method provided by the present invention considers the coupling relationship between multiple parameters and reflects the influence of multiple parameters on injection parameters through a multiple linear regression model. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0041] Figure 1 is a flowchart of the implementation of the model training method provided by the embodiment of the present invention;

[0042] Figure 2 is a flowchart of the implementation of the injection parameter generation method provided by the embodiment of the present invention;

[0043] Figure 3 is a structural diagram of the model training device provided by the embodiment of the present invention;

[0044] Figure 4It is a structural diagram of an injection parameter generation device provided by an embodiment of the present invention;

[0045] Figure 5 It is an architecture diagram of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0047] Refer to Figure 1 , an embodiment of the present invention provides a model training method, and the trained model is used to generate injection parameters, including the following steps:

[0048] Step S110. Obtain a training set, where the training set includes clinical feature parameters with injection parameter labels.

[0049] In this step, the clinical feature parameters serve as the input of the model, enabling the model to generate injection parameters suitable for the patient according to the input patient clinical feature parameters.

[0050] Specifically, the clinical feature parameters include the patient's physiological characteristics (such as weight, height, gender, age, examination site, etc.), past medical history information (such as allergy history, past diseases, etc.), and treatment-related parameters (such as injection drug concentration, past injection history, etc.). In this embodiment, the clinical feature parameters can be obtained through multiple channels such as electronic medical records, examination reports, and medical device sensors, and after preprocessing, standardization, and normalization processing, they are stored in a unified format. The injection parameter labels can be considered as the injection parameters actually determined in clinical practice, such as injection dose, injection flow rate, injection pressure, etc. The injection parameter labels can be determined through clinical verification and standards formulated by experts and provided to the training network as supervision signals.

[0051] In a further embodiment, since the clinical feature parameters are often multi-data source and multi-modal, the original data can be cleaned to remove missing or abnormal data, ensuring the accuracy and consistency of the training data, and converting the data from different sources into a unified format to form a standardized training set, ensuring the compatibility of the input data during the subsequent model training process.

[0052] Step S120. Use a multiple linear regression model to construct a training network, take the clinical feature parameters as the input of the training network for training, so that the training network learns the mapping relationship between the injection parameters and the clinical feature parameters, and generates injection parameter prediction values.

[0053] In this step, considering the actual injection parameter generation process, although there may be certain non-linear components in the relationship between input features and output parameters, in many clinical scenarios, the influence of patients' physiological characteristics and treatment parameters on injection parameters often approximately satisfies a linear relationship, or can be well approximated as a linear relationship after appropriate feature engineering (such as normalization, standardization, and logarithmic transformation). The multiple linear regression model can capture this linear mapping to a certain extent while maintaining the simplicity and stability of the calculation process.

[0054] Specifically, the multiple linear regression model can be defined as:

[0055]

[0056] where, represents the injection parameters predicted by the model, which can be, for example, injection dose, injection flow rate, etc. represents the input clinical feature parameters. represents the intercept term. represents the partial regression coefficients of each parameter, reflecting the marginal influence of each independent variable on the predicted value when other variables remain unchanged.

[0057] Furthermore, by matrixizing the above definition formula, it can be expressed as:

[0058]

[0059] where, represents the injection parameters. represents the input matrix composed of clinical feature parameters. represents the transpose of the matrix. represents the vector of partial regression coefficients. .

[0060] Input the clinical feature parameters of each sample into the multiple linear regression model, calculate the predicted value of the injection parameters according to the current parameters, and gradually adjust each parameter through model training so that the predicted value output by the model is as close as possible to the actual injection parameter label, thereby learning the mapping relationship between the input features and the injection parameters.

[0061] Step S130. Establish a loss function for the training network based on the injection parameter labels and the predicted values of the injection parameters, update the network parameters of the training network according to the loss function, and use the trained training network as the injection parameter generation model.

[0062] In this step, the multiple linear regression model adopts the principle of linear mapping, that is, by solving the best-fitting straight line (or hyperplane) to establish the relationship between input variables and output variables. By minimizing the prediction error (such as the mean squared error), the model parameters are continuously updated until the preset convergence criterion is reached. By establishing a loss function based on the error between the injection parameter labels and the predicted values, and using the gradient descent algorithm to update the model parameters, the training of the training network is completed.

[0063] Specifically, to measure the difference between the predicted injection parameters and the actual labels, this embodiment uses the mean squared error (MSE) as the loss function, and its mathematical expression is:

[0064]

[0065] The above matrix form can be expressed as follows:

[0066]

[0067] Among them, represents the number of training samples in the training set, represents the injection parameter label of the i-th training sample, represents the predicted value of the injection parameter of the i-th training sample.

[0068] Furthermore, using the gradient descent algorithm, according to the preset learning rate, each parameter is gradually updated, and the partial derivative of the j-th partial regression coefficient is calculated:

[0069]

[0070] Among them, represents the value of the i-th row and j-th column in the matrix X.

[0071] Using the above partial derivatives to adjust the network parameters, the update rule of each partial regression coefficient can be expressed as follows:

[0072]

[0073] This process is iterated until the loss function converges to the preset threshold or reaches the maximum number of iterations, so that the training network finally achieves the best prediction performance.

[0074] By continuously adjusting the model parameters, the mean square error between the predicted value and the actual injection parameter label is continuously reduced, so as to accurately learn the mapping relationship between the input clinical feature parameters and the output injection parameters. After all parameter updates are completed, the obtained trained network is the injection parameter generation model. After being fully trained, this model has a high prediction accuracy and can automatically generate injection parameters that meet the actual requirements according to different clinical features of patients during actual use.

[0075] Furthermore, considering the dependency relationship between injection parameters, the determination of some injection parameters depends not only on the patient's clinical features but also on other injection parameters. Therefore, the patient treatment parameters in the clinical feature parameters can also include the first injection parameter. Thus, the first injection parameter can be used as a new input (either combined with the clinical features or alone as an input) and fed into another training network to learn the mapping relationship between the first injection parameter and the second injection parameter, and predict the second injection parameter.

[0076] In a more preferred embodiment, when training the injection parameter generation model for generating the second injection parameter, the first injection parameter as the input can be the data obtained during the training set acquisition stage, or the generated value of the first injection parameter generation model after training or the predicted value during the training stage. This hierarchical relationship ensures that the dependency and coupling effects between injection parameters can be reflected during the injection parameter generation process.

[0077] In a further embodiment, when two injection parameter generation models with a dependency relationship need to be trained, a joint loss function can be constructed. For example, the loss function of the second injection parameter generation model is used to feedback and update the network parameters of the first injection parameter generation model.

[0078] Refer to Figure 2 , an embodiment of the present invention provides an injection parameter generation method, including the following steps:

[0079] Step S210. Obtain the clinical feature parameters of the patient.

[0080] Step S220. Input the clinical feature parameters into the trained injection parameter generation model to generate injection parameters corresponding to the clinical feature parameters.

[0081] In a further embodiment, the clinical characteristic parameters input in step S220 may include a first injection parameter, which may be predicted by a first injection parameter generation model based on the first clinical characteristic parameters, or may be known determined data. The first injection parameter generation model and the second injection parameter generation model may share some of the same clinical characteristic parameters during training, the difference being that the input of the second injection parameter generation model includes the first injection parameter, and therefore the first injection parameter and the first clinical characteristic parameter may be combined into a second clinical characteristic parameter, and the second clinical characteristic parameter is input into the second injection parameter generation model obtained by training in the first aspect to generate a second injection parameter corresponding to the second clinical characteristic parameter.

[0082] The present invention is further described below through a scenario of injection parameter generation for high-pressure injection of contrast agent.

[0083] In one embodiment of the present invention, it is necessary to train an injection dose generation model and an injection flow rate generation model. The training process of these two models will be described below.

[0084] For the training of the injection dose generation model, the clinical characteristic parameters defined as input include the patient's physiological parameters, medical history and injection concentration, where the physiological parameters include height, weight, gender, age and examination site; the medical history includes the patient's allergy history.

[0085] Then the training set of the injection dose generation model can be expressed as:

[0086]

[0087] The architecture of the generation model is constructed using a multivariate linear regression model, and the mathematical expression of the injection dose generation model can be expressed as follows:

[0088]

[0089] The model training process can refer to the aforementioned description of the embodiment of the model training method, which will not be repeated here.

[0090] Similarly, the training set of the injection flow rate generation model can be supplemented with the dose parameter based on the above training set, and the training set of the injection flow rate generation model is expressed as:

[0091]

[0092] The architecture of the generation model is constructed using a multivariate linear regression model, and the mathematical expression of the injection dose generation model can be expressed as follows:

[0093]

[0094] The training of the injection dose generation model and the injection flow rate generation model can be carried out simultaneously, or the injection dose generation model can be trained first, and the predicted value of the injection dose generation model is supplemented into the training set and then the injection flow rate generation model is trained.

[0095] In actual use, the clinical characteristic parameters of the patient are first input into the injection dose generation model to generate a predicted injection dose, and then the predicted injection dose and other clinical characteristic parameters are input into the injection flow rate generation model together to obtain the predicted value of the injection flow rate.

[0096] Using the injection parameter generation model to generate parameters can provide parameter recommendation references for injection personnel before injection operations.

[0097] The above-mentioned disclosed method can be implemented by devices in various forms. Therefore, the present invention also discloses a device corresponding to the above method, and specific embodiments are given below for detailed description.

[0098] As Figure 3 shown, an embodiment of the present invention provides a model training device for training an injection parameter generation model, including:

[0099] A training set acquisition module 302, configured to acquire a training set, where the training set includes clinical characteristic parameters with injection parameter labels;

[0100] A network training module 304, configured to construct a training network using a multiple linear regression model, train by using the clinical characteristic parameters as the input of the training network, so that the training network learns the mapping relationship between the injection parameters and the clinical characteristic parameters, and generates injection parameter predicted values;

[0101] A network update module 306, configured to establish a loss function of the training network based on the injection parameter labels and the injection parameter predicted values, update the network parameters of the training network according to the loss function, and use the trained training network as the injection parameter generation model.

[0102] Refer to Figure 4 , an embodiment of the present invention provides an injection parameter generation device, including:

[0103] A clinical parameter acquisition module 402, configured to acquire the clinical characteristic parameters of the patient;

[0104] An injection parameter generation module 404, configured to input the clinical characteristic parameters into the injection parameter generation model trained by the device of the third aspect to generate injection parameters corresponding to the clinical characteristic parameters.

[0105] The device provided by the embodiments of the present application has the same implementation principle and the same technical effects as the foregoing method embodiments. For a brief description, for the parts not mentioned in the device embodiments, reference may be made to the corresponding contents in the foregoing method embodiments.

[0106] The methods and related devices mentioned in the above embodiments are described with reference to the method flowcharts and / or structural schematic diagrams provided in the embodiments of the present application. Specifically, they can be implemented by computer program instructions for each process and / or block in the method flowchart and / or structural schematic diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in Figure 1 one process or multiple processes and / or structural schematic Figure 1 one block or multiple blocks. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that implements the functions specified in Figure 1 one process or multiple processes and / or structural schematic Figure 1 one block or multiple blocks. These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one process or multiple processes and / or structural schematic one block or multiple blocks.

[0107] The following embodiments are described by taking the application of the method to a computer device as an example. It can be understood that the computer device can be any device with computing and processing functions, and can be, but is not limited to, a server or a personal laptop computer, etc. In one of the embodiments, the computer device can be an application server, and the application server can be a server for running an application under test.

[0108] Referring to Figure 5 , which shows a hardware structure block diagram of an electronic device. The electronic device is intended to represent various forms of digital computers, such as, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, a personal digital processor, a cellular phone, a smart phone, a wearable device, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described herein and / or claimed.

[0109] AsFigure 5 As shown, the electronic device includes: at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4;

[0110] In the embodiments of the present application, the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 complete mutual communication through the communication bus 4;

[0111] The processor 1 may be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention, etc.;

[0112] The memory 3 may include a high-speed RAM memory, and may also include a non-volatile memory, etc., such as at least one disk memory;

[0113] Among them, the memory stores a program, and the processor can call the program stored in the memory. The program is used to: implement the foregoing model training method and / or each step of the injection parameter generation method.

[0114] The embodiments of the present invention also provide a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements each step of the model training method and / or the injection parameter generation method provided by any possible implementation manner of the above embodiments and / or the combined embodiments.

[0115] The above embodiments have described the present invention in particular detail with respect to possible scenarios. Those skilled in the art will recognize that the present invention can be practiced through other embodiments. The specific naming of components, the case of terms, attributes, data structures, or any other programming or structural aspects are not mandatory or important. The mechanism or its features for implementing the present invention can have different names, forms, or procedures. The system can be implemented through a combination of hardware and software (as described), entirely through hardware elements, or entirely through software elements. The specific division of functions between various system components described in the text is merely exemplary and not mandatory; on the contrary, the functions performed by a single system component can be performed by multiple components, or the functions performed by multiple components can be performed by a single component.

[0116] Those skilled in the art should understand that each step of the above-disclosed method can be implemented by a general-purpose computing device. They can be centralized on a single computing device or distributed over a network composed of multiple computing devices. Optionally, they can be implemented with program code executable by the computing device, so that they can be stored in a storage device and executed by the computing device, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module for implementation. Thus, the disclosure of the embodiments of the present invention is not limited to any specific combination of hardware and software.

[0117] These programs executable by the computing device (also referred to as programs, software, software applications, or code) include machine instructions for a programmable processor and can implement these computing programs using high-level procedures and / or object-oriented programming languages and / or assembly / machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or apparatus (e.g., a magnetic disk, an optical disk, a memory, a programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0118] Certain aspects of the present invention include process steps and instructions described herein in the form of algorithms. It should be noted that the process steps and instructions of the present invention can be implemented in software, firmware, and / or hardware. When implemented by software, it can be downloaded and thus stored on different platforms used by various operating systems and operated from said platforms.

[0119] Those skilled in the art can understand that the structures shown in the respective drawings are merely block diagrams of some of the structures related to the solution of the present application and do not constitute a limitation on the terminal devices to which the solution of the present application is applied. The specific terminal devices may include more or fewer components than those shown in the figures, or combine some components, or have different component arrangements.

[0120] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "possible design", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0121] Finally, it should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the said element.

[0122] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A model training method, characterized in that, The model trained by the method is used to generate injection parameters, including: Obtain a training set, where the training set includes clinical feature parameters with injection parameter labels; Construct a training network using a multiple linear regression model, and use the clinical feature parameters as the input of the training network for training, so that the training network learns the mapping relationship between injection parameters and clinical feature parameters, and generates predicted values of injection parameters; Establish a loss function for the training network based on the injection parameter labels and the predicted values of injection parameters, update the network parameters of the training network according to the loss function, and use the trained training network as an injection parameter generation model.

2. The method according to claim 1, wherein The mathematical expression of the injection parameter generation model is as follows: ， Among them, represents the injection parameters, represents the input matrix composed of clinical characteristic parameters, represents the transpose of the matrix, represents the vector of partial regression coefficients.

3. The method according to claim 1, wherein The loss function for establishing the training network based on the injection parameter labels and the predicted values of injection parameters includes: Using the mean square error as the loss function, the loss function is expressed as follows: Among them, represents the number of training samples in the training set, represents the injection parameter label of the i-th training sample, represents the predicted value of the injection parameter of the i-th training sample.

4. The method according to claim 1, wherein The clinical feature parameters at least include patient physiological parameters, patient past medical history, and patient treatment parameters.

5. The method according to claim 4, characterized in that, The patient treatment parameters include first injection parameters. Using the clinical feature parameters as the input of the training network for training includes: Using the first injection parameter as the input of the training network for training, so that the training network learns the mapping relationship between the first injection parameter and the second injection parameter, and generates a predicted value of the second injection parameter.

6. A method for generating injection parameters, characterized in that, Including: Step S1. Obtain the clinical feature parameters of the patient; Step S2. Input the clinical feature parameters into the injection parameter generation model trained by the method according to any one of claims 1 to 5 to generate injection parameters corresponding to the clinical feature parameters.

7. The method according to claim 6, wherein The step S2 includes: Input the first clinical feature parameter into the first injection parameter generation model trained by the method according to any one of claims 1 to 5 to generate the first injection parameter corresponding to the first clinical feature parameter; Combine the first injection parameter and the first clinical feature parameter into a second clinical feature parameter, and input the second clinical feature parameter into the second injection parameter generation model trained by the method according to any one of claims 1 to 5 to generate the second injection parameter corresponding to the second clinical feature parameter.

8. A model training device, characterized in that, For training an injection parameter generation model, including: A training set acquisition module for obtaining a training set, where the training set includes clinical feature parameters with injection parameter labels; A network training module for constructing a training network using a multiple linear regression model, using the clinical feature parameters as the input of the training network for training, so that the training network learns the mapping relationship between the injection parameters and the clinical feature parameters, and generates predicted values of injection parameters; A network update module for establishing a loss function for the training network based on the injection parameter labels and the predicted values of injection parameters, updating the network parameters of the training network according to the loss function, and using the trained training network as an injection parameter generation model.

9. An injection parameter generation device, characterized in that, Including: A clinical parameter acquisition module for obtaining the clinical feature parameters of the patient; An injection parameter generation module for inputting the clinical feature parameters into the injection parameter generation model trained by the model training device according to claim 8 to generate injection parameters corresponding to the clinical feature parameters.

10. An electronic device, characterized in that, Comprising a memory storing computer-executable instructions and a processor, when the computer-executable instructions are executed by the processor, the device is caused to execute the model training method according to any one of claims 1 to 5, and / or the injection parameter generation method according to any one of claims 6 to 7.

Images