A virtual standardized patient teaching model generation method and system

Abstract

The application relates to the technical field of medical teaching, in particular to a virtual standardized patient teaching model generation method and system. The method comprises the following steps: acquiring case data for model training; analyzing the case data to determine patient inquiry conditions and patient types; calling inquiry site monitoring; based on the patient types, analyzing the inquiry site monitoring to determine the patient inquiry state of the same patient type; based on the patient types, analyzing the patient inquiry conditions to determine the patient interaction level; determining model training parameters according to the patient inquiry conditions, the patient inquiry state and the patient interaction level, and generating a virtual standardized patient teaching model according to the model training parameters. The virtual standardized patient teaching model improves the authenticity in the visual and auditory dimensions. The virtual standardized patient teaching model can accurately reflect the interaction dynamics in real inquiry. The creation of a high-fidelity teaching tool is realized, and inquiry training in a simulated environment is ensured.

Description

Technical Field

[0001] This application relates to the field of medical teaching technology, and in particular to a method and system for generating virtual standardized patient teaching models. Background Technology

[0002] In the field of medical education, especially in the teaching of mental illnesses, virtual standardized patient teaching models are widely used as an efficient teaching tool to simulate real patient consultation scenarios in order to train medical students' diagnostic, communication and decision-making abilities.

[0003] However, existing methods often rely on isolated case data, resulting in models that fail to fully reflect the patient's true condition and are ineffective in simulating unprofessional expressions or hidden information. Summary of the Invention

[0004] This application provides a method and system for generating virtual standardized patient teaching models to solve the above-mentioned problems.

[0005] In a first aspect, this application provides a method for generating a virtual standardized patient teaching model, the method comprising:

[0006] Obtain case data for model training; analyze the case data to determine patient history and patient type;

[0007] Retrieve on-site monitoring data of the consultation; based on the patient type, analyze the on-site monitoring data to determine the consultation status of patients of the same patient type;

[0008] Based on the patient type, analyze the patient's consultation history to determine the level of patient interaction;

[0009] Based on the patient's consultation details, consultation status, and interaction level, model training parameters are determined, and a virtual standardized patient teaching model is generated based on the model training parameters.

[0010] This solution acquires case data for model training, ensuring that the generated virtual standardized patient teaching model is based on real clinical information, thereby improving the initial data coverage and trainability of the virtual standardized patient teaching model. Case data is analyzed to determine patient consultation details and patient type, ensuring customized processing for different disease types. Consultation site monitoring is retrieved to ensure the correlation between monitoring data and case data, thus expanding the data dimensions for virtual standardized patient teaching model training and avoiding the limitations of relying solely on text data. Based on patient type, consultation site monitoring is analyzed to determine the consultation status of patients of the same type, improving the realism of the virtual standardized patient teaching model in visual and auditory dimensions. Based on patient type, patient consultation details are analyzed to determine patient interaction levels, ensuring that the virtual standardized patient teaching model accurately reflects the dynamic interactions in real consultations. Based on patient consultation details, patient consultation status, and patient interaction levels, model training parameters are determined, and a virtual standardized patient teaching model is generated based on these parameters, realizing the creation of a high-fidelity teaching tool and ensuring consultation training in a simulated environment.

[0011] Optionally, the step of analyzing the on-site monitoring of the consultation based on the patient type to determine the patient's consultation status includes:

[0012] Analyze the case data to determine the timestamps and content of the consultation interactions;

[0013] Based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the real-time response behavior;

[0014] Based on the patient type, the consultation interaction content and the real-time response behavior are analyzed to determine the consultation status of patients of the same patient type.

[0015] This solution analyzes case data to determine the timestamps and content of patient consultations, ensuring accurate correlation between time, content, and behavior. This lays a data foundation for overall analysis and avoids ambiguity or subjective input. Based on the timestamps, it analyzes on-site monitoring to determine real-time response behaviors, capturing patient behavioral dimensions, compensating for the limitations of purely verbal data, enriching multidimensional information about patient performance, and ensuring that the determination of patient consultation status does not rely on a single data source. Based on patient type, it analyzes consultation interaction content and real-time response behaviors to determine the consultation status of patients of the same type, ensuring that the status description is accurate, adjustable, and suitable for teaching and assessment.

[0016] Optionally, the step of analyzing the patient's medical history based on the patient type to determine the level of patient interaction includes:

[0017] Based on the patient type, a patient feature database is determined, and based on the patient feature database, common patient characteristics corresponding to the patient type are determined;

[0018] Analyze the patient's medical history to determine the patient's response time and the content of their response;

[0019] Analyze the content of the response to determine the accuracy of the wording and the coherence of the thought process;

[0020] The patient interaction level is determined based on the general patient characteristics, patient response time, accuracy of wording, and coherence of thought.

[0021] This solution establishes a patient characteristic database based on patient type. Based on this database, it identifies common patient characteristics corresponding to each patient type, preventing analysis from deviating from the patient type. Patient interviews are analyzed to determine response times and content, ensuring objectivity and consistency in patient interaction levels. Response content is analyzed to determine word accuracy and coherence of thought, quantifying patient reaction speed and capturing verbal output, ensuring patient interaction levels cover both time and content dimensions. Based on common patient characteristics, response times, word accuracy, and coherence of thought, patient interaction levels are determined, providing actionable interaction assessment results.

[0022] Optionally, determining the patient interaction level based on the general patient characteristics, the patient response time, the accuracy of the wording, and the coherence of thought includes:

[0023] Based on the case data, environmental constraint factors were determined;

[0024] Analyze the on-site monitoring of the consultation to determine the suppression level of the patient's condition by the environmental constraint factors;

[0025] The patient interaction level is determined based on the general patient characteristics, the level of suppression, the patient response time, the accuracy of wording, and the coherence of thought.

[0026] This approach identifies environmental constraint factors based on case data, avoiding subjective judgment and parameter ambiguity. Analysis of on-site consultation monitoring determines the suppression level of the patient's condition by these environmental constraint factors, quantifies the degree to which the condition is concealed, and ensures that behavioral analysis is based on real data fusion, thus improving the accuracy of the virtual standardized patient teaching model in simulated environmental constraint scenarios. Based on common patient characteristics, suppression levels, patient response times, accuracy of wording, and coherence of thought, the patient interaction level is determined, avoiding parameter subjectivity and improving the repeatability and reliability of the virtual standardized patient teaching model.

[0027] Optionally, the analysis of the on-site monitoring during the consultation to determine the suppression level of the patient's condition by the environmental constraint factors includes:

[0028] Based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the behavioral and interaction changes of the environmental constraint factors;

[0029] Based on the behavioral changes and interaction changes, determine the linkage effect of the environmental constraint factors on patient behavior at the consultation interaction timestamp;

[0030] Based on the aforementioned linkage effect, the suppression level of the patient's condition by the environmental constraint factor is determined.

[0031] This solution analyzes on-site monitoring of patient consultations based on consultation interaction timestamps to determine behavioral and interactive changes in environmental constraint factors, ensuring that the suppression level calculation is based on objective data rather than vague descriptions. Based on these behavioral and interactive changes, it identifies the interconnected impact of environmental constraint factors on patient behavior at each consultation interaction timestamp, avoiding the one-sidedness of treating behavior or interaction in isolation and ensuring the precise capture of constraint effects. Based on these interconnected effects, it determines the suppression level of the patient's condition caused by environmental constraint factors, improving the accuracy and practicality of the virtual standardized patient teaching model.

[0032] Optionally, determining the model training parameters based on the patient's consultation details, consultation status, and interaction level includes:

[0033] Based on the consultation interaction timestamp, the patient's consultation status is analyzed to determine the patient's physical manifestations;

[0034] Analyze the patient's medical history to determine any tendency toward expression defects;

[0035] Based on the consultation interaction timestamp, and by combining the patient's consultation status and consultation details, the patient's emotional changes are determined.

[0036] Based on the general patient characteristics, patient interaction level, patient physical manifestations, tendency of expression deficits, and patient emotional changes, a disease severity level is established.

[0037] Based on the severity of the illness, the case data is segmented, and the model training parameters are determined.

[0038] This solution analyzes patient consultation status based on consultation interaction timestamps, determines patient physical manifestations, and ensures that the analysis results of patient consultation status are synchronized with the timestamps, achieving real-time data accuracy. It analyzes patient consultation situations to identify tendencies in expression deficits, ensuring that the analysis results of patient consultation situations are aligned with the timestamps, maintaining data consistency. Based on consultation interaction timestamps, it comprehensively analyzes patient consultation status and situations to determine patient emotional changes, ensuring data fusion with synchronized timestamps, enabling real-time assessment of patient status. Based on common patient characteristics, patient interaction levels, patient physical manifestations, tendencies in expression deficits, and patient emotional changes, it establishes disease severity levels, providing a unified basis for classifying case data. Based on disease severity levels, it classifies case data, determines model training parameters, and achieves structured grouping of case data, improving the adaptability and accuracy of the virtual standardized patient teaching model training.

[0039] Optionally, the analysis of the response content to determine the accuracy of wording and the coherence of thought includes:

[0040] Analyze the content of the response to determine its characteristics;

[0041] Based on the consultation interaction timestamp, determine the question data corresponding to the response content;

[0042] Determine the accuracy of wording based on the content characteristics and the question data;

[0043] Analyze the content of the response to determine the syntactic dependency tree and word vector distribution of the response text;

[0044] Analyze the syntactic dependency tree to determine the logical jump density;

[0045] Analyze the word vector distribution to determine the degree of deviation from the technical terminology;

[0046] Based on the logical jump density and the degree of deviation of technical terms, the coherence of thought is determined.

[0047] This solution analyzes response content, identifies its characteristics, and ensures that the analysis of wording accuracy and coherence of thought is based on solid evidence. Based on the consultation interaction timestamp, it identifies the corresponding question data for the response content, ensuring synchronization between question data and response content and avoiding analytical gaps. Based on content characteristics and question data, it determines wording accuracy, reflecting whether the response addresses the core of the problem, thus providing clear indicators of wording quality. Analyzing the response content, it determines the syntactic dependency tree and word vector distribution of the response text, enabling the analysis of logical jump density and technical terminology deviation based on syntactic and semantic representations. Analyzing the syntactic dependency tree determines logical jump density, reflecting logical structural defects in the response. Analyzing the word vector distribution determines technical terminology deviation, capturing defects in professional expression. Based on logical jump density and technical terminology deviation, it determines coherence of thought, eliminating the problem of parameter ambiguity.

[0048] Optionally, the analysis of the case data to determine the patient's medical history and patient type includes:

[0049] Analyze the case data to determine symptom description keywords and medical history timeline;

[0050] Based on the keywords described in the symptoms, the patient type is determined by matching them with a mental illness classification database.

[0051] Based on the aforementioned medical history timeline, the case data is analyzed to determine changes in the patient's condition;

[0052] Based on the changes in the patient's condition, determine the patient's medical history.

[0053] This solution analyzes case data to identify symptom description keywords and a medical history timeline, ensuring efficient processing of symptom and medical history information and avoiding data redundancy. Based on symptom description keywords, it matches the mental illness classification database to determine patient types, ensuring that patient types accurately reflect symptom characteristics and laying a typological foundation for analyzing changes in patient condition and patient interviews. Based on the medical history timeline, it analyzes case data to determine changes in patient condition, capture disease progression, and provide a time dimension for interviews, ensuring that patient status reflects historical changes. Based on changes in patient condition, it determines patient interview details, ensuring that the virtual standardized patient teaching model accurately presents real-time patient behavior in simulations.

[0054] Optionally, the step of analyzing the on-site monitoring of the consultation based on the consultation interaction timestamp to determine the real-time response behavior includes:

[0055] Based on the consultation interaction timestamp, the on-site monitoring of the consultation was analyzed to determine the facial key point coordinate sequence and limb posture data;

[0056] Analyze the facial key point coordinate sequence to determine the duration and intensity of micro-expressions;

[0057] Analyze the limb posture data to determine the spatiotemporal distribution pattern of unconventional movements;

[0058] Real-time reaction behavior is determined based on the duration of the micro-expression, the intensity of the micro-expression, and the unconventional action pattern.

[0059] This solution analyzes on-site monitoring of patient consultations based on consultation interaction timestamps to determine facial key point coordinate sequences and limb posture data, ensuring that behavioral analysis is based on temporal alignment and spatial localization. Analyzing the facial key point coordinate sequences determines the duration and intensity of micro-expressions, revealing the patient's transient emotional fluctuations. Analyzing limb posture data identifies the spatiotemporal distribution patterns of unconventional movements, highlighting abnormal patient behavior, capturing nonverbal cues, and adding spatial and temporal dimensions to real-time behavioral responses. Based on the duration and intensity of micro-expressions and unconventional movement patterns, real-time response behaviors are determined, reflecting the patient's real-time state.

[0060] Secondly, this application provides a virtual standardized patient teaching model generation system, the system comprising:

[0061] The data analysis module is used to acquire case data for model training; analyze the case data to determine patient consultation details and patient type;

[0062] The monitoring and analysis module is used to retrieve on-site monitoring data from the consultation process; based on the patient type, it analyzes the on-site monitoring data to determine the consultation status of patients of the same patient type;

[0063] The situation analysis module is used to analyze the patient's consultation situation based on the patient type and determine the patient's interaction level;

[0064] The model generation module is used to determine model training parameters based on the patient's consultation details, consultation status, and interaction level, and to generate a virtual standardized patient teaching model based on the model training parameters.

[0065] Optionally, the virtual standardized patient teaching model generation system further includes a state determination module, used for:

[0066] Analyze the case data to determine the timestamps and content of the consultation interactions;

[0067] Based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the real-time response behavior;

[0068] Based on the patient type, the consultation interaction content and the real-time response behavior are analyzed to determine the consultation status of patients of the same patient type.

[0069] Optionally, when the situation analysis module analyzes the patient's consultation history based on the patient type and determines the patient's interaction level, it is used for:

[0070] Based on the patient type, a patient feature database is determined, and based on the patient feature database, common patient characteristics corresponding to the patient type are determined;

[0071] Analyze the patient's medical history to determine the patient's response time and the content of their response;

[0072] Analyze the content of the response to determine the accuracy of the wording and the coherence of the thought process;

[0073] The patient interaction level is determined based on the general patient characteristics, patient response time, accuracy of wording, and coherence of thought.

[0074] Optionally, when the situation analysis module determines the patient interaction level based on the common patient characteristics, the patient response time, the accuracy of the wording, and the coherence of thought, it is used for:

[0075] Based on the case data, environmental constraint factors were determined;

[0076] Analyze the on-site monitoring of the consultation to determine the suppression level of the patient's condition by the environmental constraint factors;

[0077] The patient interaction level is determined based on the general patient characteristics, the level of suppression, the patient response time, the accuracy of wording, and the coherence of thought.

[0078] Optionally, when the situation analysis module analyzes the on-site monitoring of the consultation and determines the suppression level of the patient's condition due to the environmental constraint factors, it is used for:

[0079] Based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the behavioral and interaction changes of the environmental constraint factors;

[0080] Based on the behavioral changes and interaction changes, determine the linkage effect of the environmental constraint factors on patient behavior at the consultation interaction timestamp;

[0081] Based on the aforementioned linkage effect, the suppression level of the patient's condition by the environmental constraint factor is determined.

[0082] Optionally, when determining model training parameters based on the patient's consultation details, consultation status, and interaction level, the model generation module is used for:

[0083] Based on the consultation interaction timestamp, the patient's consultation status is analyzed to determine the patient's physical manifestations;

[0084] Analyze the patient's medical history to determine any tendency toward expression defects;

[0085] Based on the consultation interaction timestamp, and by combining the patient's consultation status and consultation details, the patient's emotional changes are determined.

[0086] Based on the general patient characteristics, patient interaction level, patient physical manifestations, tendency of expression deficits, and patient emotional changes, a disease severity level is established.

[0087] Based on the severity of the illness, the case data is segmented, and the model training parameters are determined.

[0088] Optionally, when the situation analysis module analyzes the response content to determine the accuracy of wording and the coherence of thought, it is used for:

[0089] Analyze the content of the response to determine its characteristics;

[0090] Based on the consultation interaction timestamp, determine the question data corresponding to the response content;

[0091] Determine the accuracy of wording based on the content characteristics and the question data;

[0092] Analyze the content of the response to determine the syntactic dependency tree and word vector distribution of the response text;

[0093] Analyze the syntactic dependency tree to determine the logical jump density;

[0094] Analyze the word vector distribution to determine the degree of deviation from the technical terminology;

[0095] Based on the logical jump density and the degree of deviation of technical terms, the coherence of thought is determined.

[0096] Optionally, when the data analysis module analyzes the case data to determine the patient's medical history and patient type, it is used for:

[0097] Analyze the case data to determine symptom description keywords and medical history timeline;

[0098] Based on the keywords described in the symptoms, the patient type is determined by matching them with a mental illness classification database.

[0099] Based on the aforementioned medical history timeline, the case data is analyzed to determine changes in the patient's condition;

[0100] Based on the changes in the patient's condition, determine the patient's medical history.

[0101] Optionally, when the status determination module analyzes the on-site monitoring of the consultation based on the consultation interaction timestamp to determine the real-time response behavior, it is used for:

[0102] Based on the consultation interaction timestamp, the on-site monitoring of the consultation was analyzed to determine the facial key point coordinate sequence and limb posture data;

[0103] Analyze the facial key point coordinate sequence to determine the duration and intensity of micro-expressions;

[0104] Analyze the limb posture data to determine the spatiotemporal distribution pattern of unconventional movements;

[0105] Real-time reaction behavior is determined based on the duration of the micro-expression, the intensity of the micro-expression, and the unconventional action pattern. Attached Figure Description

[0106] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0107] Figure 1 This is a schematic diagram illustrating an application scenario provided in one embodiment of this application;

[0108] Figure 2 A flowchart illustrating a method for generating a virtual standardized patient teaching model, as provided in an embodiment of this application;

[0109] Figure 3 This is a schematic diagram of a virtual standardized patient teaching model generation system provided in an embodiment of this application. Detailed Implementation

[0110] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0111] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0112] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0113] Existing methods often rely on isolated case data, resulting in models that fail to fully reflect the patient's true condition and are ineffective in simulating unprofessional expressions or hidden information.

[0114] Based on this, this application provides a method and system for generating a virtual standardized patient teaching model. It acquires case data for model training, ensuring that the generated virtual standardized patient teaching model is based on real clinical information, thereby improving the initial data coverage and trainability of the virtual standardized patient teaching model. The method analyzes case data to determine patient consultation details and patient type, ensuring customized processing for different disease types. It retrieves on-site consultation monitoring data to ensure correlation between monitoring data and case data, thereby expanding the data dimensions for training the virtual standardized patient teaching model and avoiding the limitations of relying solely on text data. Based on patient type, it analyzes on-site consultation monitoring to determine the consultation status of patients of the same type, improving the realism of the virtual standardized patient teaching model in visual and auditory dimensions. Based on patient type, it analyzes patient consultation details to determine patient interaction levels, ensuring that the virtual standardized patient teaching model accurately reflects the dynamic interactions in real consultations. Based on patient consultation details, patient consultation status, and patient interaction levels, it determines model training parameters and generates a virtual standardized patient teaching model based on these parameters, realizing the creation of a high-fidelity teaching tool and ensuring consultation training in a simulated environment.

[0115] Figure 1 This is a schematic diagram of an application scenario provided by this application, in which the method provided by this application is applied when generating a virtual standardized patient teaching model.

[0116] Specifically, the method provided in this application can be applied to any server, where the server interacts with monitoring equipment and a medical information system. By accessing the medical information system, it obtains case data for model training. The case data is analyzed to determine patient consultation details and patient type. The monitoring equipment retrieves on-site consultation footage. Based on patient type, the on-site monitoring footage is analyzed to determine the consultation status of patients of the same type, improving the realism of the virtual standardized patient teaching model in visual and auditory dimensions. Based on patient type, the patient consultation details are analyzed to determine the patient interaction level, ensuring that the virtual standardized patient teaching model accurately reflects the dynamic interactions in real consultations. Based on the patient consultation details, patient consultation status, and patient interaction level, model training parameters are determined, and based on these parameters, a virtual standardized patient teaching model is generated, realizing the creation of a high-fidelity teaching tool and ensuring consultation training in a simulated environment.

[0117] For specific implementation details, please refer to the following examples.

[0118] Figure 2 This is a flowchart illustrating a method for generating a virtual standardized patient teaching model according to an embodiment of this application. The method of this embodiment can be applied to the server in the above scenario. Figure 2 As shown, the method includes:

[0119] S201. Obtain case data for model training; analyze case data to determine patient history and patient type;

[0120] Case data can be structured raw medical data including basic patient information, symptom descriptions, diagnostic results, and consultation records.

[0121] Patient consultation details can be the specific information about the patient's behavior during the consultation process.

[0122] The patient type can be classified as a mental illness.

[0123] Specifically, the system accesses medical information devices via API to obtain patient case data, including basic patient information, symptom descriptions, diagnostic results, and consultation records, for model training. Then, natural language processing (NLP) technology is used to analyze the case data and identify patient consultation details. Finally, based on the diagnostic results and symptom descriptions in the case data, a machine learning classifier is applied to determine the patient type.

[0124] S202. Retrieve on-site monitoring data for patient consultation; analyze the on-site monitoring data based on patient type to determine the consultation status of patients of the same type;

[0125] On-site monitoring during a consultation can be real-time recording of the consultation process.

[0126] The same patient type can be a disease classification group that is the same as the identified patient type.

[0127] The patient's state during a consultation can be considered as their external behavioral state.

[0128] Specifically, by accessing a database storing video surveillance equipment and audio recording devices used in patient consultations, the system retrieves consultation monitoring data associated with case data. Based on patient type, computer vision and audio processing technologies are applied to analyze the consultation monitoring data and determine the consultation status of patients of the same type.

[0129] S203. Based on patient type, analyze patient consultation details and determine patient interaction level;

[0130] Patient interaction level can be an indicator of the degree of interaction with patients.

[0131] Specifically, based on patient type, natural language processing technology is used to assess patient consultation; then, the frequency of patient interaction, completeness of content, and coherence of expression are statistically analyzed to determine the level of patient interaction.

[0132] S204. Based on the patient's consultation details, consultation status, and interaction level, determine the model training parameters, and generate a virtual standardized patient teaching model based on the model training parameters.

[0133] The model training parameters can be a set of quantifiable indicators used to generate virtual standardized patient teaching models.

[0134] Virtual standardized patient teaching models can be computer simulation entities used to simulate real patient consultation behavior.

[0135] Specifically, the model training parameters are determined by integrating patient consultation information, patient consultation status, and patient interaction level through a weighted average algorithm. Finally, the model training parameters are used to drive the model generation engine to construct a virtual standardized patient teaching model.

[0136] This solution acquires case data for model training, ensuring that the generated virtual standardized patient teaching model is based on real clinical information, thereby improving the initial data coverage and trainability of the virtual standardized patient teaching model. Case data is analyzed to determine patient consultation details and patient type, ensuring customized processing for different disease types. Consultation site monitoring is retrieved to ensure the correlation between monitoring data and case data, thus expanding the data dimensions for virtual standardized patient teaching model training and avoiding the limitations of relying solely on text data. Based on patient type, consultation site monitoring is analyzed to determine the consultation status of patients of the same type, improving the realism of the virtual standardized patient teaching model in visual and auditory dimensions. Based on patient type, patient consultation details are analyzed to determine patient interaction levels, ensuring that the virtual standardized patient teaching model accurately reflects the dynamic interactions in real consultations. Based on patient consultation details, patient consultation status, and patient interaction levels, model training parameters are determined, and a virtual standardized patient teaching model is generated based on these parameters, realizing the creation of a high-fidelity teaching tool and ensuring consultation training in a simulated environment.

[0137] In some embodiments, case data is analyzed to determine the consultation interaction timestamps and consultation interaction content; based on the consultation interaction timestamps, on-site monitoring of the consultation is analyzed to determine real-time response behavior; based on patient type, consultation interaction content and real-time response behavior are analyzed to determine the consultation status of patients of the same patient type.

[0138] The consultation interaction timestamp can be a time marker for each interaction event during the consultation process.

[0139] The content of the consultation interaction can be the patient's verbal responses and interaction details during the consultation process.

[0140] Real-time responsive behavior can be the patient's external behavioral characteristics in real time.

[0141] Specifically, natural language processing (NLP) technology is applied to analyze the consultation records in the case data, automatically identifying and extracting consultation interaction timestamps; simultaneously, consultation interaction content is extracted from the same case data. Then, based on the consultation interaction timestamps, the corresponding time points in the consultation site monitoring are located; subsequently, computer vision and audio processing technologies are applied to analyze the located consultation site monitoring, identifying and quantifying the patient's real-time response behavior. Finally, a rule engine is used to analyze the language patterns in the consultation interaction content and the behavioral patterns in the real-time response behavior; furthermore, combined with patient type, the consultation status of patients of the same patient type is determined.

[0142] This solution analyzes case data to determine the timestamps and content of patient consultations, ensuring accurate correlation between time, content, and behavior. This lays a data foundation for overall analysis and avoids ambiguity or subjective input. Based on the timestamps, it analyzes on-site monitoring to determine real-time response behaviors, capturing patient behavioral dimensions, compensating for the limitations of purely verbal data, enriching multidimensional information about patient performance, and ensuring that the determination of patient consultation status does not rely on a single data source. Based on patient type, it analyzes consultation interaction content and real-time response behaviors to determine the consultation status of patients of the same type, ensuring that the status description is accurate, adjustable, and suitable for teaching and assessment.

[0143] In some embodiments, a patient feature database is determined based on patient type; based on the patient feature database, common patient characteristics corresponding to the patient type are determined; patient consultation is analyzed to determine patient response time and response content; response content is analyzed to determine wording accuracy and coherence of thought; and patient interaction level is determined based on common patient characteristics, patient response time, wording accuracy, and coherence of thought.

[0144] A patient profile database can be a predefined database that stores historical case data and behavioral characteristics of different patient types.

[0145] Common patient characteristics can be the general behavioral patterns of patient types in typical consultation scenarios.

[0146] Patient response time can be the time interval between a patient's answers to a doctor's questions.

[0147] The response can be a text of the patient's verbal response during the consultation.

[0148] Accuracy of wording can be defined as the degree to which the patient's wording is accurate relative to standard medical terminology.

[0149] Coherence of thought can refer to the logical fluency and consistency of a patient's answers.

[0150] Specifically, based on patient type, a patient feature database related to that patient type is retrieved. Then, data aggregation methods are applied to summarize different features from the patient feature database and identify high-frequency features, which are determined as common patient characteristics corresponding to the patient type. Subsequently, based on patient consultation details, consultation interaction timestamps are analyzed to calculate the time interval between the patient's answers to each question, generating the patient response time; simultaneously, the patient's verbal response content is extracted to generate reply content. Then, text processing rules are applied to parse the reply content, match it with a terminology database, and evaluate the accuracy of wording in the reply content; at the same time, the sentence structure and conjunctions in the reply content are analyzed to determine the coherence of thought. Furthermore, the common patient characteristics are used as a benchmark and compared with the actual patient response time, wording accuracy, and coherence of thought; finally, based on the comparison results, a deviation value is calculated, and combined with the evaluation values ​​of wording accuracy and coherence of thought, a weighted average is used to determine the patient interaction level.

[0151] This solution establishes a patient characteristic database based on patient type. Based on this database, it identifies common patient characteristics corresponding to each patient type, preventing analysis from deviating from the patient type. Patient interviews are analyzed to determine response times and content, ensuring objectivity and consistency in patient interaction levels. Response content is analyzed to determine word accuracy and coherence of thought, quantifying patient reaction speed and capturing verbal output, ensuring patient interaction levels cover both time and content dimensions. Based on common patient characteristics, response times, word accuracy, and coherence of thought, patient interaction levels are determined, providing actionable interaction assessment results.

[0152] In some embodiments, environmental constraint factors are identified based on case data; the suppression level of the patient's condition by environmental constraint factors is determined by analyzing on-site monitoring during consultations; and the patient interaction level is determined based on general patient characteristics, suppression level, patient response time, accuracy of wording, and coherence of thought.

[0153] Environmental constraint factors can be quantitative indicators of the degree of constraint of external influences extracted from case data, such as caregivers and light exposure.

[0154] The patient's condition can refer to the patient's disease state.

[0155] Suppression level can represent the degree to which a patient's condition is suppressed under environmental constraints.

[0156] Specifically, the external influence-related fields in the case data are analyzed and extracted. Then, the extracted environment-related fields are mapped to environmental constraint factors. For example, if the case data describes "the patient often looks at the caregiver," "the patient often holds the caregiver's hands tightly," or "the patient looks anxiously at the vase in the corner," then it can be considered that the patient is influenced by the caregiver or the vase in the corner, and these can be considered environmental constraint factors.

[0157] The process involves analyzing on-site monitoring during patient consultations to extract video and audio data. Image processing algorithms are then applied to analyze the video data, identifying patient body movements and facial expressions. Simultaneously, audio processing algorithms are used to analyze the audio data, recognizing tone changes. These findings are then integrated with the patient's body movements, facial expressions, and tone changes to form patient behavior indicators. Subsequently, reference behavioral benchmarks are established based on environmental constraints. The patient behavior indicators are then compared with the reference benchmarks to assess the degree of behavioral deviation. This degree of behavioral deviation is then mapped to the suppression level of the patient's condition under environmental constraints. Finally, a weighted sum is calculated based on general patient characteristics, suppression level, patient response time, word accuracy, and coherence of thought. The patient interaction level is then determined based on the calculation results.

[0158] This approach identifies environmental constraint factors based on case data, avoiding subjective judgment and parameter ambiguity. Analysis of on-site consultation monitoring determines the suppression level of the patient's condition by these environmental constraint factors, quantifies the degree to which the condition is concealed, and ensures that behavioral analysis is based on real data fusion, thus improving the accuracy of the virtual standardized patient teaching model in simulated environmental constraint scenarios. Based on common patient characteristics, suppression levels, patient response times, accuracy of wording, and coherence of thought, the patient interaction level is determined, avoiding parameter subjectivity and improving the repeatability and reliability of the virtual standardized patient teaching model.

[0159] In some embodiments, based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the behavioral and interactive changes of environmental constraint factors; based on the behavioral and interactive changes, the linkage effect of environmental constraint factors on patient behavior under the consultation interaction timestamp is determined; based on the linkage effect, the suppression level of the patient's condition by environmental constraint factors is determined.

[0160] Behavioral changes can refer to the degree of deviation in a patient's external behavior.

[0161] Changes in interaction can be the degree of deviation in the consultation interaction mode.

[0162] Patient behavior can refer to the outward behavioral characteristics exhibited by the patient during the consultation process.

[0163] The synergistic effect can be the comprehensive impact of environmental constraints on patient behavior at the time stamp of the consultation interaction.

[0164] Specifically, data segments corresponding to the timestamps of consultation interactions are extracted from the on-site monitoring. Then, video data is analyzed to identify the patient's external behavioral characteristics and compared with environmental states defined by environmental constraint factors (such as "family present" or "family absent") to calculate behavioral changes. Simultaneously, audio data is analyzed to identify consultation interaction characteristics and calculate interaction changes. Subsequently, based on behavioral and interaction changes, combined with consultation interaction timestamps, a weighted fusion algorithm is used to calculate the linkage effect of environmental constraint factors on patient behavior. Finally, based on this linkage effect, a linear mapping function is applied to convert the linkage effect into the suppression level of the patient's condition by the environmental constraint factors.

[0165] This solution analyzes on-site monitoring of patient consultations based on consultation interaction timestamps to determine behavioral and interactive changes in environmental constraint factors, ensuring that the suppression level calculation is based on objective data rather than vague descriptions. Based on these behavioral and interactive changes, it identifies the interconnected impact of environmental constraint factors on patient behavior at each consultation interaction timestamp, avoiding the one-sidedness of treating behavior or interaction in isolation and ensuring the precise capture of constraint effects. Based on these interconnected effects, it determines the suppression level of the patient's condition caused by environmental constraint factors, improving the accuracy and practicality of the virtual standardized patient teaching model.

[0166] In some embodiments, based on the consultation interaction timestamp, the patient's consultation status is analyzed to determine the patient's physical manifestations; the patient's consultation situation is analyzed to determine the tendency of expression defects; based on the consultation interaction timestamp, the patient's consultation status and consultation situation are combined to determine the patient's emotional changes; based on general patient characteristics, patient interaction level, patient's physical manifestations, tendency of expression defects, and patient's emotional changes, a disease severity level is established; based on the disease severity level, case data is divided to determine model training parameters.

[0167] A patient's physical manifestations can indicate the degree of abnormality in the patient's external behavioral characteristics.

[0168] A tendency toward expression deficit can be an indication of the degree of impairment in a patient's ability to express themselves.

[0169] Patient emotional changes can indicate the real-time direction of changes in a patient's emotions.

[0170] Disease severity level can be a grade value that indicates the degree of severity of a disease.

[0171] Specifically, based on video data corresponding to the consultation interaction timestamps, the patient's consultation status in the video data is analyzed. Then, computer vision algorithms are used to detect the amplitude, frequency, and type of limb movements to determine the patient's physical expression. Next, based on audio data corresponding to the consultation interaction timestamps, the patient's consultation situation in the audio data is analyzed. Then, audio processing technology is used to detect changes in speech rate, clarity, and terminology usage to determine tendencies in expression deficiencies. Then, based on the consultation interaction timestamps, emotion-related visual features are extracted from the patient's consultation status; simultaneously, emotion-related auditory features are extracted from the patient's consultation situation. Then, combining visual and auditory features, changes in the patient's emotional state are detected in the timestamp sequence. Then, a weighted sum of general patient characteristics, patient interaction level, patient physical expression, tendency in expression deficiencies, and patient emotional changes is calculated to establish a disease severity level. Finally, based on the disease severity level, the case data are grouped according to the severity level; then, model training parameters are determined for each group of case data.

[0172] This solution analyzes patient consultation status based on consultation interaction timestamps, determines patient physical manifestations, and ensures that the analysis results of patient consultation status are synchronized with the timestamps, achieving real-time data accuracy. It analyzes patient consultation situations to identify tendencies in expression deficits, ensuring that the analysis results of patient consultation situations are aligned with the timestamps, maintaining data consistency. Based on consultation interaction timestamps, it comprehensively analyzes patient consultation status and situations to determine patient emotional changes, ensuring data fusion with synchronized timestamps, enabling real-time assessment of patient status. Based on common patient characteristics, patient interaction levels, patient physical manifestations, tendencies in expression deficits, and patient emotional changes, it establishes disease severity levels, providing a unified basis for classifying case data. Based on disease severity levels, it classifies case data, determines model training parameters, and achieves structured grouping of case data, improving the adaptability and accuracy of the virtual standardized patient teaching model training.

[0173] In some embodiments, the response content is parsed to determine content features; the question data corresponding to the response content is determined based on the consultation interaction timestamp; the wording accuracy is determined based on the content features and question data; the response content is analyzed to determine the syntactic dependency tree and word vector distribution of the response text; the syntactic dependency tree is analyzed to determine the logical jump density; the word vector distribution is analyzed to determine the deviation of technical terms; and the coherence of thought is determined based on the logical jump density and the deviation of technical terms.

[0174] Content features can be structured data extracted from the response content.

[0175] Question data can be question records that correspond to the content of the answers.

[0176] The response text can be the answer provided by a virtual patient.

[0177] Syntactic dependency trees can be tree-structured representations of response text.

[0178] Word vector distribution can be a distributional representation of word vectors in the response text.

[0179] Logical jump density can represent the density of logical connectors in a syntactic dependency tree.

[0180] The deviation of technical terms can be used to represent the degree of deviation between the distribution of word vectors and the technical terminology database.

[0181] Specifically, the response text is segmented into word sequences. Part-of-speech tagging is then applied to identify the grammatical role of each word. Named entity recognition is performed to extract different entities, thereby determining content features. Next, based on the consultation interaction timestamp, the database storing historical consultation interaction records searches for question records with the same timestamp. The question records are then associated with the response content to determine the question data corresponding to the response content. Furthermore, the keywords in the content features are analyzed to determine if they address the core issue of the question data. Semantic consistency is then evaluated, using a semantic rule base to check if the response content avoids ambiguous language or ambiguity. A similarity score is then calculated to determine word accuracy. Finally, deep text analysis is performed on the response content, applying a dependency parser to process the response text and constructing a syntactic dependency tree. Simultaneously, a pre-defined word embedding model is built by training a word embedding algorithm on a large general corpus. The pre-defined word embedding model is then used to convert each word in the response text into a vector representation. Finally, the word vector distribution of the response text is calculated. Then, the syntactic dependency tree is traversed to identify the types of syntactic dependencies. Subsequently, the frequency of logical connectives in the syntactic dependency tree is statistically analyzed to determine the logical jump density. Next, a professional terminology database constructed from authoritative medical terminology is loaded and compiled, and the distance between each word vector in the response text and its corresponding vector in the professional terminology database is calculated. Then, the distance values ​​of several words are aggregated to determine the professional terminology deviation. Finally, the logical jump density and professional terminology deviation are normalized to a unified scale and fused using a weighted average formula to determine the coherence of thought.

[0182] This solution analyzes response content, identifies its characteristics, and ensures that the analysis of wording accuracy and coherence of thought is based on solid evidence. Based on the consultation interaction timestamp, it identifies the corresponding question data for the response content, ensuring synchronization between question data and response content and avoiding analytical gaps. Based on content characteristics and question data, it determines wording accuracy, reflecting whether the response addresses the core of the problem, thus providing clear indicators of wording quality. Analyzing the response content, it determines the syntactic dependency tree and word vector distribution of the response text, enabling the analysis of logical jump density and technical terminology deviation based on syntactic and semantic representations. Analyzing the syntactic dependency tree determines logical jump density, reflecting logical structural defects in the response. Analyzing the word vector distribution determines technical terminology deviation, capturing defects in professional expression. Based on logical jump density and technical terminology deviation, it determines coherence of thought, eliminating the problem of parameter ambiguity.

[0183] In some embodiments, case data is parsed to determine symptom description keywords and medical history timeline; based on the symptom description keywords, a mental illness classification database is matched to determine the patient type; based on the medical history timeline, case data is analyzed to determine changes in the patient's condition; and based on changes in the patient's condition, the patient's consultation details are determined.

[0184] Symptom description keywords can be a list of keywords related to the symptoms.

[0185] The medical history timeline can be a time series data structure for medical history.

[0186] A mental illness classification database can be a pre-stored database containing standard mental illness types and related symptom keywords.

[0187] Changes in a patient's condition can be described as a pattern of disease progression.

[0188] Specifically, a word segmentation tool is used to break down the text content in the case data into word sequences. Then, a keyword extraction algorithm is used to scan the segmented word sequences, identify and extract keywords related to symptoms, forming symptom description keywords. Simultaneously, the timestamp information in the case data is analyzed and sorted chronologically to construct a medical history timeline. Next, a mental illness classification database storing standard mental illness types and associated symptom keywords is accessed. Then, the extracted symptom description keywords are matched one by one with the symptom entries in the mental illness classification database, calculating the similarity score between each symptom description keyword and an entry in the database. Based on the calculation results, the disease type with the highest score is selected as the patient type. Then, based on the medical history timeline, the case data is reorganized chronologically. Next, the symptom intensity at different time points is analyzed. Subsequently, a trend analysis algorithm is used to identify the disease progression pattern and determine changes in the patient's condition. Finally, based on changes in the patient's condition, a pre-defined mapping rule is applied to map these changes to the patient's medical history.

[0189] This solution analyzes case data to identify symptom description keywords and a medical history timeline, ensuring efficient processing of symptom and medical history information and avoiding data redundancy. Based on symptom description keywords, it matches the mental illness classification database to determine patient types, ensuring that patient types accurately reflect symptom characteristics and laying a typological foundation for analyzing changes in patient condition and patient interviews. Based on the medical history timeline, it analyzes case data to determine changes in patient condition, capture disease progression, and provide a time dimension for interviews, ensuring that patient status reflects historical changes. Based on changes in patient condition, it determines patient interview details, ensuring that the virtual standardized patient teaching model accurately presents real-time patient behavior in simulations.

[0190] In some embodiments, based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the facial key point coordinate sequence and limb posture data; the facial key point coordinate sequence is analyzed to determine the duration and intensity of micro-expressions; the limb posture data is analyzed to determine the spatiotemporal distribution pattern of unconventional movements; and real-time response behavior is determined based on the duration, intensity, and unconventional movement pattern of micro-expressions.

[0191] A sequence of facial key point coordinates can be a sequence of changes in the coordinates of facial feature points over time.

[0192] Limb posture data can be a data sequence of the coordinates and angles of body joints changing over time.

[0193] The duration of a micro-expression can be the time difference from the start to the end of a micro-expression event.

[0194] The duration of micro-expressions can be based on the relative change in the displacement amplitude of key facial points.

[0195] Unconventional movements can be abnormal movements compared to a preset baseline of normal movements.

[0196] Spatiotemporal distribution patterns can be the characteristic patterns of unconventional actions in terms of spatial and temporal distribution.

[0197] Unconventional action patterns can be quantified output patterns that describe the characteristics of unconventional actions.

[0198] Specifically, time frames corresponding to the timestamps of consultation interactions are extracted from the on-site monitoring of the consultation. Then, a facial keypoint detection algorithm is applied to each time frame to identify facial feature points and generate a sequence of facial keypoint coordinates that changes over time. Simultaneously, a limb posture detection algorithm is applied to each time frame to identify the coordinates and angles of body joints, generating limb posture data. Subsequently, the dynamic changes between the facial keypoint coordinate sequences are calculated to detect micro-expression events. Then, for each micro-expression event, the duration of the micro-expression is calculated, and the intensity of the micro-expression is evaluated. Next, the limb posture data is compared with a preset baseline of conventional actions established using manually defined standard behavioral patterns. Based on the comparison results, unconventional actions are identified, and their spatiotemporal distribution patterns are analyzed. Finally, the duration of micro-expressions, the intensity of micro-expressions, and the unconventional action patterns are input into a feature combination, and a behavior mapping rule engine is applied to determine real-time response behavior.

[0199] This solution analyzes on-site monitoring of patient consultations based on consultation interaction timestamps to determine facial key point coordinate sequences and limb posture data, ensuring that behavioral analysis is based on temporal alignment and spatial localization. Analyzing the facial key point coordinate sequences determines the duration and intensity of micro-expressions, revealing the patient's transient emotional fluctuations. Analyzing limb posture data identifies the spatiotemporal distribution patterns of unconventional movements, highlighting abnormal patient behavior, capturing nonverbal cues, and adding spatial and temporal dimensions to real-time behavioral responses. Based on the duration and intensity of micro-expressions and unconventional movement patterns, real-time response behaviors are determined, reflecting the patient's real-time state.

[0200] Figure 3 This is a schematic diagram of the structure of a virtual standardized patient teaching model generation system provided in an embodiment of this application, as shown below. Figure 3 As shown, the virtual standardized patient teaching model generation system 300 of this embodiment includes: a data analysis module 301, a monitoring and analysis module 302, a situation analysis module 303, and a model generation module 304.

[0201] Data analysis module 301 is used to acquire case data for model training; analyze the case data to determine patient consultation details and patient type;

[0202] The monitoring and analysis module 302 is used to retrieve on-site monitoring data of the consultation; based on the patient type, it analyzes the on-site monitoring data of the consultation to determine the consultation status of patients of the same patient type;

[0203] The situation analysis module 303 is used to analyze the patient's consultation situation based on the patient type and determine the patient's interaction level;

[0204] The model generation module 304 is used to determine model training parameters based on the patient's consultation situation, the patient's consultation status and the patient's interaction level, and to generate a virtual standardized patient teaching model based on the model training parameters.

[0205] Optionally, the virtual standardized patient teaching model generation system further includes a state determination module 305, used for:

[0206] Analyze the case data to determine the timestamps and content of the consultation interactions;

[0207] Based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the real-time response behavior;

[0208] Based on the patient type, the consultation interaction content and the real-time response behavior are analyzed to determine the consultation status of patients of the same patient type.

[0209] Optionally, when the situation analysis module 303 analyzes the patient's medical history based on the patient type and determines the patient's interaction level, it is used for:

[0210] Based on the patient type, a patient feature database is determined, and based on the patient feature database, common patient characteristics corresponding to the patient type are determined;

[0211] Analyze the patient's medical history to determine the patient's response time and the content of their response;

[0212] Analyze the content of the response to determine the accuracy of the wording and the coherence of the thought process;

[0213] The patient interaction level is determined based on the general patient characteristics, patient response time, accuracy of wording, and coherence of thought.

[0214] Optionally, when the situation analysis module 303 determines the patient interaction level based on the common patient characteristics, the patient response time, the accuracy of the wording, and the coherence of thought, it is used for:

[0215] Based on the case data, environmental constraint factors were determined;

[0216] Analyze the on-site monitoring of the consultation to determine the suppression level of the patient's condition by the environmental constraint factors;

[0217] The patient interaction level is determined based on the general patient characteristics, the level of suppression, the patient response time, the accuracy of wording, and the coherence of thought.

[0218] Optionally, when the situation analysis module 303 analyzes the on-site monitoring of the consultation and determines the suppression level of the patient's condition due to the environmental constraint factors, it is used for:

[0219] Based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the behavioral and interaction changes of the environmental constraint factors;

[0220] Based on the behavioral changes and interaction changes, determine the linkage effect of the environmental constraint factors on patient behavior at the consultation interaction timestamp;

[0221] Based on the aforementioned linkage effect, the suppression level of the patient's condition by the environmental constraint factor is determined.

[0222] Optionally, when determining model training parameters based on the patient's consultation details, consultation status, and interaction level, the model generation module 304 is used to:

[0223] Based on the consultation interaction timestamp, the patient's consultation status is analyzed to determine the patient's physical manifestations;

[0224] Analyze the patient's medical history to determine any tendency toward expression defects;

[0225] Based on the consultation interaction timestamp, and by combining the patient's consultation status and consultation details, the patient's emotional changes are determined.

[0226] Based on the general patient characteristics, patient interaction level, patient physical manifestations, tendency of expression deficits, and patient emotional changes, a disease severity level is established.

[0227] Based on the severity of the illness, the case data is segmented, and the model training parameters are determined.

[0228] Optionally, when the situation analysis module 303 analyzes the response content and determines the accuracy of wording and the coherence of thought, it is used for:

[0229] Analyze the content of the response to determine its characteristics;

[0230] Based on the consultation interaction timestamp, determine the question data corresponding to the response content;

[0231] Determine the accuracy of wording based on the content characteristics and the question data;

[0232] Analyze the content of the response to determine the syntactic dependency tree and word vector distribution of the response text;

[0233] Analyze the syntactic dependency tree to determine the logical jump density;

[0234] Analyze the word vector distribution to determine the degree of deviation from the technical terminology;

[0235] Based on the logical jump density and the degree of deviation of technical terms, the coherence of thought is determined.

[0236] Optionally, when the data analysis module 301 analyzes the case data to determine the patient's medical history and patient type, it is used for:

[0237] Analyze the case data to determine symptom description keywords and medical history timeline;

[0238] Based on the keywords described in the symptoms, the patient type is determined by matching them with a mental illness classification database.

[0239] Based on the aforementioned medical history timeline, the case data is analyzed to determine changes in the patient's condition;

[0240] Based on the changes in the patient's condition, determine the patient's medical history.

[0241] Optionally, when the state determination module 305 analyzes the on-site monitoring of the consultation based on the consultation interaction timestamp to determine the real-time response behavior, it is used for:

[0242] Based on the consultation interaction timestamp, the on-site monitoring of the consultation was analyzed to determine the facial key point coordinate sequence and limb posture data;

[0243] Analyze the facial key point coordinate sequence to determine the duration and intensity of micro-expressions;

[0244] Analyze the limb posture data to determine the spatiotemporal distribution pattern of unconventional movements;

[0245] Real-time reaction behavior is determined based on the duration of the micro-expression, the intensity of the micro-expression, and the unconventional action pattern.

[0246] The system in this embodiment can be used to execute the methods of any of the above embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

Claims

1. A method for generating a virtual standardized patient teaching model, characterized in that, include: Obtain case data for model training; Analyze the case data to determine the patient's medical history and patient type; Access the surveillance footage of the consultation site; Based on the patient type, analyze the on-site monitoring of the consultation to determine the consultation status of patients of the same patient type; Based on the patient type, analyze the patient's consultation history to determine the level of patient interaction; Based on the patient's consultation details, consultation status, and interaction level, the model training parameters are determined, and a virtual standardized patient teaching model is generated based on the model training parameters. The step of determining model training parameters based on the patient's consultation history, consultation status, and interaction level includes: Based on the consultation interaction timestamp, the patient's consultation status is analyzed to determine the patient's physical manifestations; Analyze the patient's medical history to determine any tendency toward expression defects; Based on the consultation interaction timestamp, and by combining the patient's consultation status and consultation details, the patient's emotional changes are determined. Based on general patient characteristics, patient interaction levels, patient physical manifestations, tendency to expressive deficiencies, and patient emotional changes, a disease severity level is established. Based on the severity of the illness, the case data is segmented, and the model training parameters are determined.

2. The method according to claim 1, characterized in that, The step of analyzing the on-site monitoring of the consultation based on the patient type to determine the patient's consultation status includes: Analyze the case data to determine the timestamp and content of the consultation interaction; Based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the real-time response behavior; Based on the patient type, the consultation interaction content and the real-time response behavior are analyzed to determine the consultation status of patients of the same patient type.

3. The method according to claim 2, characterized in that, The process of analyzing patient consultation data based on the patient type to determine the level of patient interaction includes: Based on the patient type, a patient feature database is determined, and based on the patient feature database, the common patient characteristics corresponding to the patient type are determined; Analyze the patient's medical history to determine the patient's response time and the content of their response; Analyze the content of the response to determine the accuracy of the wording and the coherence of the thought process; The patient interaction level is determined based on the general patient characteristics, patient response time, accuracy of wording, and coherence of thought.

4. The method according to claim 3, characterized in that, The determination of patient interaction level based on the general patient characteristics, patient response time, accuracy of wording, and coherence of thought includes: Based on the case data, environmental constraint factors were determined; Analyze the on-site monitoring of the consultation to determine the suppression level of the patient's condition by the environmental constraint factors; The patient interaction level is determined based on the general patient characteristics, the level of suppression, the patient response time, the accuracy of wording, and the coherence of thought.

5. The method according to claim 4, characterized in that, The analysis of the on-site monitoring during the consultation to determine the suppression level of the patient's condition by the environmental constraint factors includes: Based on the consultation interaction timestamp, the on-site monitoring of the consultation is analyzed to determine the behavioral and interaction changes of the environmental constraint factors; Based on the behavioral changes and interaction changes, determine the linkage effect of the environmental constraint factors on patient behavior at the consultation interaction timestamp; Based on the aforementioned linkage effect, the suppression level of the patient's condition by the environmental constraint factor is determined.

6. The method according to claim 3, characterized in that, The analysis of the response content, determining the accuracy of wording and the coherence of thought, includes: Analyze the content of the response to determine its characteristics; Based on the consultation interaction timestamp, determine the question data corresponding to the response content; Determine the accuracy of wording based on the content characteristics and the question data; Analyze the content of the response to determine the syntactic dependency tree and word vector distribution of the response text; Analyze the syntactic dependency tree to determine the logical jump density; Analyze the word vector distribution to determine the degree of deviation from the technical terminology; Based on the logical jump density and the degree of deviation of technical terms, the coherence of thought is determined.

7. The method according to claim 1, characterized in that, The analysis of the case data, determining the patient's medical history and patient type, includes: Analyze the case data to determine symptom description keywords and medical history timeline; Based on the keywords described in the symptoms, the patient type is determined by matching them with a mental illness classification database. Based on the aforementioned medical history timeline, the case data is analyzed to determine changes in the patient's condition; Based on the changes in the patient's condition, determine the patient's medical history.

8. The method according to claim 2, characterized in that, The step of analyzing the on-site monitoring of the consultation based on the consultation interaction timestamp to determine real-time response behavior includes: Based on the consultation interaction timestamp, the on-site monitoring of the consultation was analyzed to determine the facial key point coordinate sequence and limb posture data; Analyze the facial key point coordinate sequence to determine the duration and intensity of micro-expressions; Analyze the limb posture data to determine the spatiotemporal distribution pattern of unconventional movements; Real-time reaction behavior is determined based on the duration of the micro-expression, the intensity of the micro-expression, and the unconventional action pattern.

9. A virtual standardized patient teaching model generation system, characterized in that, The method applied to any one of claims 1-8 includes: The data analysis module is used to acquire case data for model training; analyze the case data to determine patient consultation details and patient type; The monitoring and analysis module is used to retrieve on-site monitoring data from the consultation process; based on the patient type, it analyzes the on-site monitoring data to determine the consultation status of patients of the same patient type; The situation analysis module is used to analyze the patient's consultation situation based on the patient type and determine the patient's interaction level; The model generation module is used to determine model training parameters based on the patient's consultation details, consultation status, and interaction level, and to generate a virtual standardized patient teaching model based on the model training parameters.

Images