Ultrasound examination method, system, device, storage medium and program product

Abstract

The application relates to the technical field of artificial intelligence, and provides an ultrasonic examination method, system, device, storage medium and program product, the method comprising the following steps: acquiring instruction text; performing intention recognition on the instruction text based on an ultrasonic examination model to obtain an interface function corresponding to the instruction text, the ultrasonic examination model being obtained by training on the basis of sample instruction text and sample interface functions corresponding to the sample instruction text; performing parameter analysis on the interface function to obtain an execution parameter, and controlling an ultrasonic robot to perform ultrasonic examination based on the execution parameter.The ultrasonic examination method, system, device, storage medium and program product provided by the application train an ultrasonic examination model on the basis of a large language model, use the trained ultrasonic examination model as a leading intelligent system, drive an ultrasonic robot used for ultrasonic examination, and provide more accurate, comfortable and widely applicable ultrasonic scanning schemes in an automated, intelligent and convenient manner.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and in particular to an ultrasonic examination method, system, device, storage medium and program product. BACKGROUND

[0002] In the medical field, ultrasonic examination is a widely used imaging technology, especially in the examination of carotid arteries. Currently, ultrasonic examination is usually manually operated by professional medical personnel, relying on their professional knowledge and skills to acquire and interpret images. Handheld ultrasonic instruments require operators to manually operate, which may be affected by the skills and experience of the operators, resulting in subjectivity and errors.

[0003] In recent years, with the advancement of technology, robot-assisted ultrasonic systems have emerged to improve the accuracy and efficiency of scanning. However, these systems mostly rely on preset programs and limited adaptive capabilities. SUMMARY

[0004] The present application provides an ultrasonic examination method, system, device, storage medium and program product to solve the defects of subjective errors caused by manual operation of medical personnel for ultrasonic examination and poor adaptive ability of robot-assisted ultrasonic systems in the prior art.

[0005] The present application provides an ultrasonic examination method, comprising:

[0006] acquiring instruction text;

[0007] performing intent recognition on the instruction text based on an ultrasonic examination model to obtain an interface function corresponding to the instruction text, the ultrasonic examination model being obtained by training a large language model based on sample instruction text and its corresponding sample interface function;

[0008] performing parameter analysis on the interface function to obtain execution parameters, and controlling an ultrasonic robot to perform ultrasonic examination based on the execution parameters.

[0009] According to the ultrasonic examination method provided by the present application, the training step of the ultrasonic examination model comprises:

[0010] obtaining corpus text related to the ultrasonic robot, and pre-training the large language model based on the corpus text to obtain a pre-trained ultrasonic examination model;

[0011] determining sample instruction text and its corresponding sample interface function based on the corpus text;

[0012] inputting the sample instruction text into the pre-trained ultrasonic examination model to obtain a predicted interface function output by the pre-trained ultrasonic examination model;

[0013] Fine-tune the pre-training ultrasound examination model based on the difference between the sample interface function and the predicted interface function to obtain the ultrasound examination model.

[0014] According to the ultrasound examination method provided by the application, the sample instruction text and the corresponding sample interface function are determined based on the corpus text, which includes:

[0015] The corpus text is standardized, and the initial sample instruction text and the corresponding initial sample interface function are determined based on the standardized corpus text.

[0016] The initial sample instruction text and the corresponding initial sample interface function are format-converted to obtain the sample instruction text and the corresponding sample interface function.

[0017] According to the ultrasound examination method provided by the application, the interface function corresponding to the instruction text is obtained by performing intent recognition on the instruction text based on an ultrasound examination model, which includes:

[0018] Based on a knowledge base, relevant information retrieval is performed on the instruction text to obtain a retrieval result.

[0019] The instruction text and the retrieval result are input into the ultrasound examination model to obtain the corresponding interface function output by the ultrasound examination model.

[0020] According to the ultrasound examination method provided by the application, the ultrasound robot is controlled to perform ultrasound examination based on the execution parameter, which includes:

[0021] The ultrasound robot is controlled to perform ultrasound examination based on the execution parameter to obtain a current examination result.

[0022] The current examination result is returned to the ultrasound examination model, so that the ultrasound examination model makes a decision based on the current examination result.

[0023] According to the ultrasound examination method provided by the application, the method further includes:

[0024] The instruction text and the ultrasound examination result are displayed on the user interface.

[0025] The application also provides an ultrasound examination system, which includes:

[0026] A text acquisition unit is configured to acquire an instruction text.

[0027] An interface determination unit is configured to perform intent recognition on the instruction text based on an ultrasound examination model to obtain an interface function corresponding to the instruction text, wherein the ultrasound examination model is obtained by training a sample instruction text and a corresponding sample interface function on the basis of a large language model.

[0028] The ultrasound examination unit is configured to perform parameter resolution on the interface function to obtain execution parameters, and control the ultrasound robot to perform the ultrasound examination based on the execution parameters.

[0029] The application further provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the ultrasound examination method according to any one of the above when executing the program.

[0030] The application further provides a non-transitory computer readable storage medium, which stores a computer program, and the computer program implements the ultrasound examination method according to any one of the above when executed by a processor.

[0031] The application further provides a computer program product, which includes a computer program, and the computer program implements the ultrasound examination method according to any one of the above when executed by a processor.

[0032] The ultrasound examination method, system, device, storage medium and program product provided by the application are based on a large language model to train an ultrasound examination model, and the trained ultrasound examination model is applied as a leading intelligent system to drive an ultrasound robot for ultrasound examination. The ultrasound robot is a carrier for executing intelligent decisions made by the ultrasound examination model, and provides a more accurate, comfortable and widely applicable ultrasound scanning method in an automated, intelligent and convenient manner. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0034] Figure 1 is one of the flowcharts of the ultrasound examination method provided by the application.

[0035] Figure 2 is another flowchart of the ultrasound examination method provided by the application.

[0036] Figure 3 is a structural schematic diagram of the ultrasound examination system provided by the application.

[0037] Figure 4 is a structural schematic diagram of the electronic device provided by the application. DETAILED DESCRIPTION

[0038] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in conjunction with the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work belong to the protection scope of the present application.

[0039] Existing neck vessel ultrasound examination methods, whether manual or robot-assisted, have certain degree of accuracy and consistency problems. Manual operation is limited by the skills and experience of the operator, and the quality of ultrasound examination depends largely on the professional skills and experience of the operator; and the robot-assisted system is difficult to adapt to complex medical scenarios due to the lack of high-level cognitive ability.

[0040] In addition, traditional ultrasound scanning equipment has limitations in operation flexibility and adaptability, especially when facing patients with complex anatomical structures or significant individual differences, resulting in poor patient comfort and satisfaction.

[0041] To solve the above problems, the embodiments of the present application provide an ultrasound examination method, in which first, the instruction text is obtained, then the intention of the instruction text is recognized based on the ultrasound examination model to obtain the interface function corresponding to the instruction text, the ultrasound examination model is obtained by training based on the large language model and the sample interface function corresponding to the sample instruction text; the interface function is parsed to obtain the execution parameter, and the ultrasound robot is controlled to execute the ultrasound examination based on the execution parameter.

[0042] The ultrasound examination method provided by the embodiments of the present application trains the ultrasound examination model based on the large language model, applies the trained ultrasound examination model as the leading intelligent system to drive the ultrasound robot for ultrasound examination. The ultrasound robot serves as the carrier of the intelligent decision made by the ultrasound examination model, and provides more accurate, comfortable and widely applicable ultrasound scanning methods in an automated, intelligent and convenient manner.

[0043] The embodiments of the present application can be applied to scenarios requiring ultrasound examination, such as whole-body ultrasound examination, such as carotid artery, abdomen, musculoskeletal system, and other medical imaging and diagnosis scenarios; can also be applied to remote medical services, performing ultrasound examination through remote control of an ultrasound robot; can also be applied to medical education and professional training, helping students and doctors learn ultrasound examination techniques and interpret results; can also be used during surgery to provide real-time ultrasound imaging support for surgeons, increasing the safety and success rate of surgery. The execution subject of the method can be a terminal device, a computer, a server, a server cluster, or a specially designed ultrasound examination device, etc. electronic device, or an ultrasound examination system provided in the electronic device, which can be realized by the combination of software and hardware.

[0044] Figure 1 is one of the flowcharts of the ultrasound examination method provided by the present application, as shown in Figure 1 The ultrasound examination method comprises the following steps:

[0045] Step 110, obtaining instruction text.

[0046] Specifically, the instruction text refers to text that can express user requirements, which can be natural language instructions input by the user.

[0047] The instruction text can be text input directly by the user; it can also be obtained by transcribing audio collected after the user inputs voice; it can also be obtained by performing OCR (Optical Character Recognition) or intent recognition on an image input by the user, and the present application does not make specific limitations on this.

[0048] The instruction text can be any language, and the ultrasound examination method can support multiple languages, such as Chinese, English, Russian, and French, etc. For example, the instruction text can be "Please perform a carotid-ultrasound scan of the neck" or "Please perform a carotid-ultrasound scan of the neck", etc.

[0049] It should be noted that the requirement text can be a deterministic instruction or an uncertain instruction, and the present application does not make specific limitations on this.

[0050] Step 120, based on an ultrasound examination model, performing intent recognition on the instruction text to obtain an interface function corresponding to the instruction text, wherein the ultrasound examination model is obtained by training sample instruction texts and their corresponding sample interface functions based on a large language model.

[0051] Specifically, the embodiment of the present application applies a trained ultrasound examination model as a leading intelligent system to drive an ultrasound robot for ultrasound examination. Therefore, after obtaining the instruction text, it is necessary to perform intent recognition on the instruction text to obtain the interface function corresponding to the instruction text.

[0052] It should be noted that the ultrasound examination model can be trained before step 120 is performed. In the embodiment, the ultrasound examination model is trained based on a large language model by applying sample instruction texts and their corresponding sample interface functions.

[0053] A large language model (LLM) refers to a natural language processing (NLP) model with a large number of parameters. The model processes large-scale text data during training and has the ability to understand and generate natural language. For example, the large language model can include ChatGLM-6B and its derivative models.

[0054] In some embodiments, the training step of the ultrasound examination model includes:

[0055] Obtain corpus text related to the ultrasound robot, and pre-train the large language model based on the corpus text to obtain a pre-trained ultrasound examination model;

[0056] Determine sample instruction texts and their corresponding sample interface functions based on the corpus text;

[0057] Input the sample instruction texts into the pre-trained ultrasound examination model to obtain the predicted interface functions output by the pre-trained ultrasound examination model;

[0058] Based on the difference between the sample interface functions and the predicted interface functions, fine-tune the pre-trained ultrasound examination model to obtain the ultrasound examination model.

[0059] Specifically, the large language model is retrained and supervised fine-tuned according to the specific needs of the ultrasound robot. The retraining here refers to a transitional stage between pre-training and supervised fine-tuning, which retains the universality of pre-training and adds the professionalism of specific tasks.

[0060] For the ultrasound examination scene of the ultrasound robot, first, collect corpus text related to the operation of the ultrasound robot. The corpus text can include the operation manual of the robot, the code repository, medical terminology, operation instructions, and the code of the robot written, etc.

[0061] Based on the corpus text, the large language model is pre-trained to obtain a pre-trained ultrasound examination model. The collected corpus text is combined in the following format: [

[0063] {

[0064] "text": "xxx"

[0065] },

[0066] {

[0067] "text": "xxx"

[0068] }, ... ]

[0071] According to the method of "next word prediction", secondary pre-training is performed. The loss calculated for each text is used for gradient backpropagation. The loss calculation method can be cross-entropy loss.

[0072] Then, based on the corpus text, sample instruction texts and their corresponding sample interface functions are determined, and the pre-trained ultrasound examination model is supervised fine-tuned. Here, supervised fine-tuning refers to using supervised learning methods to enhance the ability of the pre-trained ultrasound examination model to follow specific format instructions in this scenario. The sample instruction texts and their corresponding sample interface functions should cover various instructions and corresponding operations that may be encountered in ultrasound examinations.

[0073] First, the corpus text can be standardized, including data cleaning, word segmentation processing, and irrelevant data removal to ensure data quality and relevance. Based on the standardized corpus text, initial sample instruction texts and their corresponding initial sample interface functions are determined. The initial sample instruction texts and their corresponding initial sample interface functions can be obtained by combining the standardized corpus text according to a pre-set format.

[0074] On this basis, the initial sample instruction texts and their corresponding initial sample interface functions are converted to sample instruction texts and their corresponding sample interface functions.

[0075] For the ultrasound examination scenario, first determine the data format. Then convert the "initial sample instruction text - corresponding initial sample interface function" data into the corresponding data format. Then train according to the method of "next word prediction". Note that only the loss of the output is used for gradient backpropagation, and the loss calculation method can be cross-entropy loss.

[0076] The supervised fine-tuning data is combined according to the following format: [

[0078] {

[0079] "instruction": "xxxx", "input": "xxxx", "output": "xxxx"

[0080] },

[0081] {

[0082] "instruction": "xxxx", "input": "xxxx", "output": "xxxx"

[0083] }, ... ]

[0086] The ultrasound examination model trained by the above method can realize efficient natural language understanding and processing, and generate an API (Application Programming Interface) of an ultrasound robot that needs to be called. That is, an interface function, which can be a best interface calling method.

[0087] For example, the instruction text is "please perform carotid artery ultrasound scan", and the interface function corresponding to the instruction text can be "Run robot".

[0088] In some other embodiments, based on the ultrasound examination model, the instruction text is subjected to intent recognition to obtain an interface function corresponding to the instruction text, including:

[0089] Based on the knowledge base, the instruction text is subjected to relevant information retrieval to obtain a retrieval result;

[0090] The instruction text and the retrieval result are input into the ultrasound examination model to obtain the interface function output by the ultrasound examination model.

[0091] Specifically, in order to make the ultrasound detection model more accurately identify the user's intention, ensure the accurate analysis of the instruction, and make the generated interface function more in line with the user's intention, in this embodiment, first, the instruction text is subjected to relevant information retrieval based on the knowledge base to obtain a retrieval result. Through data retrieval of multiple knowledge bases, comprehensive information support can be provided.

[0092] On this basis, the instruction text and the retrieval result are input into the ultrasound examination model to obtain the interface function output by the ultrasound examination model.

[0093] In step 130, the interface function is subjected to parameter analysis to obtain an execution parameter, and the ultrasound robot is controlled to perform ultrasound examination based on the execution parameter.

[0094] Specifically, the ultrasound examination model and the ultrasound robot can achieve effective docking and collaborative work. After the ultrasound examination model generates the interface function, the interface function can be further parsed for parameters to obtain execution parameters. The ultrasound robot executes specific ultrasound scanning operations according to the parsed execution parameters.

[0095] In some other embodiments, the control of the ultrasound robot to perform the ultrasound examination based on the execution parameters comprises:

[0096] The control of the ultrasound robot to perform the ultrasound examination based on the execution parameters obtains a current examination result.

[0097] The current examination result is returned to the ultrasound examination model, so that the ultrasound examination model makes a decision based on the current examination result.

[0098] Specifically, the ultrasound robot obtains a current examination result after each execution of the ultrasound examination, and returns the current examination result to the ultrasound examination model for the ultrasound examination model to make a decision based on the current examination result. The decision here may be, for example, a re-execution operation or an end operation.

[0099] If the decision of the ultrasound examination model is the re-execution operation, the ultrasound robot is controlled to perform the ultrasound examination again; if the decision of the ultrasound examination model is the end operation, the ultrasound robot is controlled to end the ultrasound examination.

[0100] Based on any of the above embodiments, the ultrasound examination method further comprises:

[0101] The instruction text, the interface function and the ultrasound examination result are displayed on the user interface.

[0102] Specifically, in order to enable medical professionals to easily monitor and guide the operation of the robot, an intuitive user interface can also be designed. The user can be interacted with by displaying the instruction text, the interface function and the ultrasound examination result on the user interface.

[0103] Based on any of the above embodiments, Figure 2 is a flowchart of the ultrasound examination method provided by the present application, as shown in Figure 2 The method comprises:

[0104] S1, data preparation and preprocessing.

[0105] An efficient data collection tool is used to focus on collecting corpus texts related to the operation of the ultrasound robot, including medical terms, operation instructions, etc.

[0106] The corpus text is standardized, and based on the standardized corpus text, the initial sample instruction text and its corresponding initial sample interface function are determined; the initial sample instruction text and its corresponding initial sample interface function are format-converted to obtain the sample instruction text and its corresponding sample interface function, which prepares for subsequent model training.

[0107] S2, model training.

[0108] Secondary pre-training of large language model: based on the general large language model, secondary pre-training is performed using specific ultrasonic robot operation corpus to enhance the professionalism of the model in this field.

[0109] Supervised fine-tuning of large language model: further fine-tune the model through structured data to accurately adapt to the specific needs of ultrasonic scanning.

[0110] S3, model evaluation and optimization.

[0111] Comprehensive testing of the model with validation data sets to evaluate its performance in accuracy, recall rate, F1 score and other key indicators. Based on the evaluation results, adjust the model parameters and conduct meticulous optimization to improve the overall performance and accuracy of the model. Implement continuous training cycles to adapt to new data and changing medical environments.

[0112] S4, model deployment.

[0113] Export the trained model into a format compatible with the actual application environment. Integrate the exported model into the ultrasonic robot system to ensure seamless collaboration between the model and the robot. The model can process natural language instructions in real time and provide corresponding operation guidance.

[0114] S5, combined with embodied system.

[0115] Effectively interface the large language model with the robot system to ensure accurate communication of instructions. Optimize the natural language input mechanism to improve the accuracy and efficiency of language understanding. Optimize the interface calling process to reduce latency and improve response speed. Based on actual application performance, further optimize the system to ensure efficient and stable operation.

[0116] S6, operation execution and feedback.

[0117] The data retrieval is combined with multiple knowledge bases to provide comprehensive information support. The user intent is identified by using the large language model to ensure accurate parsing of the instructions. The best interface calling method is determined by the large language model. The embodiment of the application parses the parameters provided by the large language model to prepare for the operation. The robot performs specific ultrasonic scanning operations according to the parsed parameters. After the robot completes the operation, the result is returned to the large language model intent recognition module for decision-making by the large language model. The entire ultrasonic scanning process is completed, and necessary diagnostic information is provided.

[0118] The embodiment of the application provides an innovative solution for ultrasonic examination by combining the large language model and the ultrasonic robot. This integration brings significant performance improvement, especially in accuracy, efficiency, adaptability and user experience. By utilizing the advanced data processing and analysis capabilities of the large language model and the precise and flexible operation of the robot, the embodiment of the application can optimize the ultrasonic examination process and provide more accurate and reliable diagnostic results.

[0119] Specifically, automated robot operation reduces the time for manual setup and adjustment, speeds up the scanning process, and improves overall examination efficiency. The system can automatically adjust according to the specific conditions of different patients, adapt to various anatomical structures and specific medical needs. It reduces the dependence on advanced ultrasonic operation skills of operators, enabling more medical staff to perform high-quality ultrasonic examinations. By reducing examination time and improving diagnostic accuracy, patient comfort and satisfaction with medical services are enhanced. The continuous learning ability of the large language model ensures that the performance of the system will continuously improve over time, adapting to new data sets. By processing and analyzing large amounts of data using the large language model, the system provides strong data-based support for doctors, assisting them in making more accurate diagnostic decisions.

[0120] The method provided by the embodiment of the application can provide fast and non-invasive examination, reduce the pressure on the bodies of elderly patients, and is particularly suitable for elderly patients. At the same time, as a research case of artificial intelligence application in the medical field, it can promote the application and exploration of artificial intelligence in more medical fields.

[0121] Based on any of the above embodiments, Figure 3 is a structural schematic diagram of an ultrasonic examination system provided by the embodiment of the application, as Figure 3 shown, the ultrasonic examination system comprises:

[0122] The text acquisition unit 310 is configured to acquire instruction text.

[0123] The interface determination unit 320 is configured to perform intent recognition on the instruction text based on an ultrasonic examination model to obtain an interface function corresponding to the instruction text, wherein the ultrasonic examination model is obtained by training a large language model based on sample instruction text and sample interface functions corresponding to the sample instruction text.

[0124] an ultrasonic examination unit 330, configured to perform parameter resolution on the interface function to obtain an execution parameter, and control the ultrasonic robot to perform ultrasonic examination based on the execution parameter.

[0125] The system provided by the embodiment of the present application is based on a large language model to train an ultrasonic examination model, and applies the trained ultrasonic examination model as a leading intelligent system to drive an ultrasonic robot for ultrasonic examination. The ultrasonic robot serves as a carrier for implementing intelligent decisions made by the ultrasonic examination model, and provides more accurate, comfortable and widely applicable ultrasonic scanning schemes in an automated, intelligent and convenient manner.

[0126] Based on any of the above embodiments, further comprising a model training unit configured to:

[0127] obtain corpus text related to the ultrasonic robot, and pre-train the large language model based on the corpus text to obtain a pre-trained ultrasonic examination model;

[0128] determine sample instruction text and corresponding sample interface functions based on the corpus text;

[0129] input the sample instruction text into the pre-trained ultrasonic examination model to obtain predicted interface functions output by the pre-trained ultrasonic examination model;

[0130] fine-tune the pre-trained ultrasonic examination model based on differences between the sample interface functions and the predicted interface functions to obtain the ultrasonic examination model.

[0131] Based on any of the above embodiments, the model training unit is specifically configured to:

[0132] perform standardization processing on the corpus text, and determine initial sample instruction text and corresponding initial sample interface functions based on the standardized corpus text;

[0133] perform format conversion on the initial sample instruction text and the corresponding initial sample interface functions to obtain the sample instruction text and the corresponding sample interface functions.

[0134] Based on any of the above embodiments, the interface determination unit is specifically configured to:

[0135] perform relevant information retrieval on the instruction text based on a knowledge base to obtain a retrieval result;

[0136] input the instruction text and the retrieval result into the ultrasonic examination model to obtain corresponding interface functions output by the ultrasonic examination model.

[0137] Based on any of the above embodiments, the ultrasonic examination unit is specifically configured to:

[0138] control the ultrasound robot to perform an ultrasound examination based on the execution parameters to obtain a current examination result;

[0139] return the current examination result to the ultrasound examination model to enable the ultrasound examination model to make a decision based on the current examination result.

[0140] Based on any of the above embodiments, further comprising a display unit for:

[0141] displaying the instruction text and the ultrasound examination result on a user interface.

[0142] Based on any of the above embodiments, an ultrasound examination system is provided, comprising:

[0143] a data processing layer: responsible for collecting, cleaning and preprocessing corpus data related to ultrasound robots, including medical terminology, operation instructions, etc.

[0144] a model training layer: using the cleaned and preprocessed data to perform secondary pre-training and supervised fine-tuning on the large language model, specifically for ultrasound robot application scenarios.

[0145] a model evaluation and optimization layer: evaluating the performance of the trained model, including accuracy, recall rate and F1 score, etc., and optimizing and adjusting according to the evaluation results.

[0146] a model deployment layer: integrating the optimized model into the ultrasound robot system to ensure real-time processing and natural language understanding capabilities of the model.

[0147] an embodied system integration layer: realizing the interface connection between the large language model and the ultrasound robot, optimizing the interface call link, and performing performance optimization.

[0148] an operation execution and feedback layer: including knowledge base merging and retrieval, intent recognition, interface function determination, parameter analysis, robot operation execution and return of operation results.

[0149] Figure 4 An example of an entity structure diagram of an electronic device is shown in Figure 4As shown, the electronic device can include a processor 410, a communications interface 420, a memory 430, and a communications bus 440, wherein the processor 410, the communications interface 420, and the memory 430 communicate with each other through the communications bus 440. The processor 410 can invoke a logical instruction in the memory 430 to execute an ultrasonic examination method, which includes: obtaining instruction text; performing intent recognition on the instruction text based on an ultrasonic examination model to obtain an interface function corresponding to the instruction text, the ultrasonic examination model being obtained by training based on sample instruction text and sample interface functions corresponding thereto; performing parameter analysis on the interface function to obtain execution parameters, and controlling an ultrasonic robot to perform ultrasonic examination based on the execution parameters.

[0150] In addition, the logical instructions in the memory 430 described above can be implemented in the form of a software functional unit and sold or used as an independent product, which can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0151] On the other hand, the present application also provides a computer program product, which includes a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program can be executed by a processor to enable a computer to execute the ultrasonic examination method provided by the above-mentioned methods, which includes: obtaining instruction text; performing intent recognition on the instruction text based on an ultrasonic examination model to obtain an interface function corresponding to the instruction text, the ultrasonic examination model being obtained by training based on sample instruction text and sample interface functions corresponding thereto; performing parameter analysis on the interface function to obtain execution parameters, and controlling an ultrasonic robot to perform ultrasonic examination based on the execution parameters.

[0152] In yet another aspect, the present application also provides a non-transitory computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the ultrasonic examination method provided by any of the above methods, and the method comprises: obtaining an instruction text; performing intent recognition on the instruction text based on an ultrasonic examination model to obtain an interface function corresponding to the instruction text, the ultrasonic examination model being obtained by training based on sample instruction texts and sample interface functions corresponding to the sample instruction texts; performing parameter analysis on the interface function to obtain an execution parameter, and controlling an ultrasonic robot to perform ultrasonic examination based on the execution parameter.

[0153] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0154] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software plus necessary universal hardware platforms, and of course can also be realized by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0155] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some 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 application.

Claims

1. An ultrasonic examination method, characterized in that, The method comprises the following steps: acquiring instruction text; performing intent recognition on the instruction text based on an ultrasound examination model to obtain an interface function corresponding to the instruction text, the ultrasound examination model being obtained by training a large language model based on sample instruction text and sample interface functions corresponding to the sample instruction text; performing parameter analysis on the interface function to obtain execution parameters, and controlling an ultrasound robot to perform ultrasound examination based on the execution parameters; the training steps of the ultrasound examination model comprise: acquiring corpus text related to the ultrasound robot, and pre-training the large language model based on the corpus text to obtain a pre-trained ultrasound examination model; determining sample instruction text and sample interface functions corresponding to the sample instruction text based on the corpus text; inputting the sample instruction text into the pre-trained ultrasound examination model to obtain predicted interface functions output by the pre-trained ultrasound examination model; fine-tuning the pre-trained ultrasound examination model based on the difference between the sample interface functions and the predicted interface functions to obtain the ultrasound examination model; the step of determining sample instruction text and sample interface functions corresponding to the sample instruction text based on the corpus text comprises: performing standardization processing on the corpus text, and determining initial sample instruction text and initial sample interface functions corresponding to the initial sample instruction text based on the standardized corpus text; performing format conversion on the initial sample instruction text and the initial sample interface functions corresponding to the initial sample instruction text to obtain the sample instruction text and the sample interface functions corresponding to the sample instruction text.

2. The ultrasonic examination method of claim 1, wherein, the step of performing intent recognition on the instruction text based on the ultrasound examination model to obtain an interface function corresponding to the instruction text comprises: performing relevant information retrieval on the instruction text based on a knowledge base to obtain retrieval results; inputting the instruction text and the retrieval results into the ultrasound examination model to obtain the interface function output by the ultrasound examination model.

3. The ultrasonic examination method of claim 1, wherein, the step of controlling the ultrasound robot to perform ultrasound examination based on the execution parameters comprises: controlling the ultrasound robot to perform ultrasound examination based on the execution parameters to obtain a current examination result; returning the current examination result to the ultrasound examination model to enable the ultrasound examination model to make a decision based on the current examination result.

4. The ultrasonic examination method of claim 1, wherein, The method further comprises: displaying the instruction text and the ultrasound examination result on a user interface.

5. An ultrasonic examination system, characterized by The device comprises: a text acquisition unit configured to acquire instruction text; an interface determination unit configured to perform intent recognition on the instruction text based on an ultrasound examination model to obtain an interface function corresponding to the instruction text, the ultrasound examination model being obtained by training a large language model based on sample instruction text and sample interface functions corresponding to the sample instruction text; an ultrasound examination unit configured to perform parameter analysis on the interface function to obtain execution parameters, and control an ultrasound robot to perform ultrasound examination based on the execution parameters; the training steps of the ultrasound examination model comprise: acquiring corpus text related to the ultrasound robot, and pre-training the large language model based on the corpus text to obtain a pre-trained ultrasound examination model; determining sample instruction text and sample interface functions corresponding to the sample instruction text based on the corpus text; inputting the sample instruction text into the pre-trained ultrasound examination model to obtain predicted interface functions output by the pre-trained ultrasound examination model; inputting the sample instruction text into the pre-trained ultrasound examination model to obtain a predicted interface function output by the pre-trained ultrasound examination model; fine-tuning the pre-trained ultrasound examination model based on a difference between the sample interface function and the predicted interface function to obtain the ultrasound examination model; the determining of the sample instruction text and the corresponding sample interface function based on the corpus text comprises: standardizing the corpus text, and determining an initial sample instruction text and an initial sample interface function corresponding to the initial sample instruction text based on the standardized corpus text; performing format conversion on the initial sample instruction text and the initial sample interface function corresponding to the initial sample instruction text to obtain the sample instruction text and the sample interface function corresponding to the sample instruction text.

6. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the ultrasound examination method according to any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the ultrasound examination method according to any one of claims 1 to 4.

8. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the ultrasound examination method according to any one of claims 1 to 4.

Images