Traditional Chinese medicine pulse diagnosis digital system based on VR and AI technologies

Abstract

The invention relates to the technical field of medical intellectualization, and discloses a traditional Chinese medicine pulse diagnosis digital system based on VR and AI technologies, which comprises a basic knowledge learning module, a machine analysis learning module and an inquiry training module, the basic knowledge learning module demonstrates eight elements including a pulse diagnosis origin, a pulse map / pulse condition forming mechanism, a pulse taking method and a pulse condition in a mode of combining animation, voice and / or image-text; the machine analysis learning module demonstrates the working principle of the pulse instrument and the pulse signal acquisition and analysis process in a mode of combining animation, voice and / or image-text; and the inquiry training module is used for simulating a pulse diagnosis medical scene between a doctor and a patient, and realizing traditional Chinese medicine clinical thinking training and examination of four-diagnosis combined reference and syndrome differentiation treatment in a man-machine online question and answer mode. The traditional Chinese medicine AI inquiry is combined with an artificial intelligence language large model, the traditional Chinese medicine clinical inquiry scene is highly restored, and students can freely train observation, smell and interrogation skills, four-diagnosis combined reference and syndrome differentiation decision-making thinking.

Description

Technical Field

[0001] The present invention relates to the technical field of medical education, and particularly relates to a digitalized traditional Chinese medicine pulse diagnosis system based on VR and AI technologies. Background Art

[0002] Traditional Chinese medicine diagnostics is a bridge between basic traditional Chinese medicine disciplines and various clinical departments, and is a core course in the professional curriculum system of traditional Chinese medicine. Among them, interrogation, as one of the four diagnostic methods of "inspection, auscultation and olfaction, interrogation, and palpation", was called by Zhang Jingyue in the Ming Dynasty as "the key to diagnosing diseases and the top priority in clinical practice". It has important clinical significance and extremely high clinical value. It is one of the most commonly used diagnostic methods for doctors to understand the disease conditions of patients, and has an important position of "the key to examining the pathogenesis" and "the shortcut to saving lives".

[0003] In actual teaching, due to problems such as the difficulty of training clinical interrogation skills in theoretical teaching and the extremely scarce opportunities for practical training teaching, it has become one of the urgent problems to be solved in the teaching of traditional Chinese medicine diagnostics that it is difficult to cultivate students' interrogation ability. Moreover, in traditional pulse diagnosis training, there are no intuitive and quantitative graphs and indicators, and there is a lack of clinically typical pulse cases. It is difficult for students to experience the finger feeling characteristics of various clinical pulses by imagination alone, and it is difficult to understand the finger feeling differences of different pulses, which affects the learning effect. Summary of the Invention

[0004] The present invention provides a digitalized traditional Chinese medicine pulse diagnosis system based on VR and AI technologies, which solves the technical problems such as the lack of clinically typical pulse cases in traditional pulse diagnosis training and the difficulty for students to experience the finger feeling characteristics of various clinical pulses by imagination alone.

[0005] The present invention can be realized through the following technical solutions:

[0006] A digitalized traditional Chinese medicine pulse diagnosis system based on VR and AI technologies, including a basic knowledge learning module, a machine analysis learning module, and an interrogation training module.

[0007] The basic knowledge learning module demonstrates the origin of pulse diagnosis, the formation mechanism of pulse diagrams / pulses, pulse-taking methods, and the eight elements of pulses in a way combining animation, voice, and / or graphics and text.

[0008] The machine analysis learning module demonstrates the working principle of the pulse detector, the pulse signal acquisition and analysis process in a way combining animation, voice, and / or graphics and text.

[0009] The interrogation training module is used to simulate the pulse diagnosis medical scenario between doctors and patients, and realizes the training and assessment of the traditional Chinese medicine clinical thinking of combining the four diagnostic methods and syndrome differentiation and treatment through an online human-computer Q&A method.

[0010] Furthermore, the interrogation training module constructs a case management database and virtual digital patients, integrates multiple speech recognition APIs or text recognition technologies such as Azure and AliYun to analyze the questions asked by the trainer playing the role of a doctor, and uses an intelligent large model for analysis and combines the information of virtual digital patients to generate voice or text answers in real time, realizing natural language interaction, completing clinical information collection. Then, the trainer writes corresponding cases based on the collected clinical information, and then uses the traditional Chinese medicine syndrome-symptom association model to identify and analyze the case content and give an evaluation result.

[0011] Furthermore, the case management database provides a standardized template, supports the entry of symptoms, signs, and syndrome types, and is associated with the traditional Chinese medicine disease and syndrome standards;

[0012] The virtual digital patient is based on a digital standard human model and constructs a case virtual patient using 3D virtual simulation technology. It can select a character model by gender and age and supports online selection of voice models;

[0013] The traditional Chinese medicine syndrome-symptom association model is constructed by analyzing the association path between symptom combinations and syndrome types through a Neo4j knowledge graph;

[0014] The intelligent large model is constructed by integrating large language models such as DeepSeek, Doubao, and ZhiPu.

[0015] Furthermore, the interrogation training module includes a learning mode, an advanced mode, and an exam mode. In the learning mode, under the auxiliary guidance of the AI robot learning companion, the trainer playing the role of a doctor and the virtual digital patient complete the clinical information collection of the virtual digital patient through natural language-style questions and answers. Then, the trainer writes a case and gives an evaluation result;

[0016] In the advanced mode, without the auxiliary guidance of the AI robot learning companion, the trainer playing the role of a doctor and the virtual digital patient complete the clinical information collection of the virtual digital patient through natural language-style questions and answers. Then, the trainer writes a case and gives an evaluation result;

[0017] The exam mode is used for online assessment of the trainer.

[0018] Furthermore, the basic knowledge learning module includes an origin module, a formation mechanism module, a pulse-taking module, and an eight-element module,

[0019] The origin module uses animation to demonstrate the origin of pulse diagnosis;

[0020] The formation mechanism module includes a pulse formation mechanism module and a pulse graph formation mechanism module. The pulse formation mechanism module uses a page-turning mode to respectively display the pulse formation mechanism, the operation of qi and blood in the human body, and the cooperation with internal organs; the pulse graph formation mechanism module uses pictures and texts to demonstrate the concept and formation mechanism of the pulse graph.

[0021] The pulse-taking module uses animations, audio, and pictures and texts to demonstrate the pulse-taking position and pulse-taking finger methods.

[0022] The eight-element module uses animations combined with pictures and texts to demonstrate the eight elements of the pulse condition.

[0023] Furthermore, the pulse formation mechanism module uses animations of the heart beating and blood flowing in blood vessels, combined with audio, pictures, and texts to demonstrate the pulse formation mechanism; uses animations of blood flowing in the blood vessels of the whole human body, combined with audio, pictures, and texts to demonstrate the operation of qi and blood in the human body; uses animations of the heart beating in cooperation with relevant internal organs, combined with audio and texts to demonstrate the internal organs that cooperate with the operation of qi and blood in the human body.

[0024] Furthermore, the machine analysis and learning module includes a pulse meter module and an analysis module. The pulse meter module uses a combination of voice and pictures and texts to demonstrate the various components of the pulse sensor, their functions, the assembly process, and the working principle, and uses a combination of voice and pictures and texts to demonstrate the various components of the pulse transducer, their functions, and the working principle.

[0025] The analysis module uses a combination of voice and pictures and texts to demonstrate the entire acquisition process of the pulse graph, demonstrates the entire analysis process of the acquired pulse graph, and uses a combination of pictures and texts to demonstrate the recognition results of the conventional pulse graph.

[0026] Furthermore, the entire analysis process of the acquired pulse graph includes noise reduction, drift removal, analysis, time-domain analysis / hemodynamic analysis, and a pulse graph analysis report.

[0027] The beneficial technical effects of the present invention are as follows:

[0028] 1. For the first time, virtual-real integration simulates a real clinical scenario, integrating advanced digital training teaching into the original case-based teaching. Compared with the past teaching method of dividing students into groups of two to respectively simulate doctors and patients, it is more realistic, can better improve students' learning enthusiasm and learning efficiency. At the same time, using this system for training teaching has a high degree of flexibility, providing students with an opportunity for self-practice that is no longer restricted by case sources, the number of people, etc., and supporting various teaching application scenarios such as online learning, advanced training, and skills examinations.

[0029] 2. The traditional Chinese medicine pulse diagnosis training system of the present invention can be said to be a "TCM AI teacher and intelligent learning partner by your side", which helps students learn independently without being restricted by time and space, giving full play to the main role of students. Teachers can also compile and build their own cases to create personalized teaching cases, enhancing the flexibility and practicality of teaching. At the same time, the content developed and designed by the system is based on the teaching syllabus of traditional Chinese medicine diagnostics and the examination syllabus of practical skills for the national medical licensing examination for traditional Chinese medicine practitioners. Senior experts are invited to participate in case writing to ensure the reliability and correctness of case sources.

[0030] In addition to medical history collection, the key content of interrogation lies in the humanistic care for patients. Only when the doctor has a kind attitude and always embodies humanistic care can there be good communication with patients. Teachers can add assessment points on humanistic care and empathic communication in the system, and the system will automatically score and provide feedback.

[0031] 3. The TCM interrogation AI training system developed by combining the traditional Chinese medicine pulse diagnosis training system of the present invention with artificial intelligence large language model technology is the first TCM education practice product based on AI technology, promoting the innovation of the training method of traditional Chinese medicine diagnosis methods and realizing the transformation to digital and scenario-based experiential learning. The real-time interactive virtual digital patients constructed by it can conduct training on the skills of inspection, auscultation and olfaction, interrogation, and palpation, as well as dialectical decision-making thinking, adding digital "wings" to the cultivation of TCM talents. At the same time, the system has an ever-expanding disease syndrome knowledge base, covering rich TCM disease syndrome information, which can improve classroom efficiency, stimulate students' learning enthusiasm, and inject new impetus into TCM digital education.

[0032] Through investigation, it is found that the use of the traditional Chinese medicine pulse diagnosis training system of the present invention can double the classroom efficiency, improve the teaching efficiency; increase the student participation rate by 40%, upgrading the learning experience; and increase the clinical dialectical accuracy rate by 25%, making it the industry benchmark for the revolutionary application of AI in TCM medical scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is the overall structural block diagram of the present invention;

[0034] Figure 2 is the demonstration diagram of the origin module of the present invention;

[0035] Figure 3 is a partial demonstration diagram of the pulse formation mechanism module of the present invention;

[0036] Figure 4 is a partial demonstration diagram of the pulse graph formation mechanism module of the present invention;

[0037] Figure 5 is a partial demonstration diagram of the pulse diagnosis position of the present invention;

[0038] Figure 6 is a partial demonstration diagram of the finger method for feeling the pulse of the present invention;

[0039] Figure 7 It is a partial demonstration diagram of the eight elements of the pulse condition of the present invention;

[0040] Figure 8 It is a partial demonstration diagram of the pulse detector module of the present invention;

[0041] Figure 9 It is a partial demonstration diagram of the analysis module of the present invention;

[0042] Figure 10 It is a schematic diagram of three modes of the inquiry training module of the present invention;

[0043] Figure 11 It is a demonstration diagram of the virtual scene of the four diagnostic methods of the virtual digital patient and the trainer of the present invention;

[0044] Figure 12 It is a demonstration diagram of writing a medical record of the present invention;

[0045] Figure 13 It is a partial demonstration diagram of the virtual digital patient of the present invention;

[0046] Figure 14 It is a demonstration diagram of the intelligent dialectics of the inquiry training module of the present invention according to the medical record;

[0047] Figure 15 It is a bar chart drawn by the inquiry training module of the present invention based on the growth record summarized from the students' nearly 20 inquiries. Detailed implementation manners

[0048] The following details the specific implementation manners of the present invention in conjunction with the accompanying drawings and preferred embodiments.

[0049] As Figure 1 shown, the present invention provides a digital system for traditional Chinese medicine pulse diagnosis based on VR and AI technologies, mainly including a basic knowledge learning module, a machine analysis learning module, and an inquiry training module. It uses 3D VR modeling to present the pulse-taking method, the formation mechanism of the pulse condition / pulse diagram, and the tactile characteristics such as the eight elements of the pulse condition in a vivid, intuitive, visual, and digital form. It uses 3D modeling to highly simulate the key components of the pulse detector, visually presents the internal components and working processes, and under the guidance of the system, conducts pulse condition collection and pulse diagram analysis, trains the students' standardized operation skills of pulse diagnosis, deepens the students' understanding of each feature of pulse diagnosis, improves the students' clinical practice skills of pulse diagnosis, cultivates the students' thinking of objective research on traditional Chinese medicine pulse diagnosis and innovative ability; at the same time, combined with the traditional Chinese medicine AI inquiry of the artificial intelligence language large model, it highly restores the traditional Chinese medicine clinical inquiry scene, enables the students to learn immersively, and the students can freely conduct training on the skills of inspection, auscultation and olfaction, inquiry, palpation, comprehensive analysis of the four diagnostic methods, and dialectical decision-making thinking. The students can fully feel the rhythm and pressure of actual diagnosis and treatment.

[0050] Specifically as follows:

[0051] I. Basic Knowledge Learning Module: It demonstrates the origin of pulse diagnosis, the formation mechanism of pulse patterns / images, pulse-taking methods, and the eight elements of pulse patterns in a combined way of animation, voice, and / or graphics and text, enabling students to understand the origin of pulse diagnosis, intuitively and quantitatively understand the connotations of the eight elements of pulse patterns from the shallower to the deeper level, experience and master the main tactile characteristics and clinical significance of common pulse patterns, be familiar with the formation mechanism of pulse patterns, and help students master basic pulse-taking methods and pulse pattern judgment skills through the 3D pulse diagnosis virtual simulation training platform and pulse pattern simulation devices, thus consolidating the theoretical knowledge of traditional Chinese medicine pulse diagnosis.

[0052] This basic knowledge learning module includes an origin module, a formation mechanism module, a pulse-taking module, and an eight-elements module. The origin module demonstrates the origin of pulse diagnosis in an animated way, such as Figure 2 shown, vividly showing the definition of pulse diagnosis, pulse-taking methods, and briefly describing the development process of pulse diagnosis and the main works.

[0053] The formation mechanism module includes a pulse pattern formation mechanism module and a pulse image formation mechanism module. The pulse pattern formation mechanism module uses a page-turning mode to respectively show the formation mechanism of the pulse, the movement of qi and blood in the human body, and the cooperation of internal organs, such as Figure 3 shown. It highlights that the forward wave generated by the heart contraction and propagating along the artery and the reflected wave reflected from the periphery are superimposed to form the pulse wave, and reflects the roles of qi and blood, blood vessels, and the five internal organs. Specifically, it shows three pages of content, namely, using animations of heart beating and blood flowing in blood vessels, combined with audio, graphics, and text to demonstrate the formation mechanism of the pulse; using animations of blood flowing in the blood vessels throughout the human body, combined with audio, graphics, and text to demonstrate the movement of qi and blood in the human body; using animations of heart beating in cooperation with relevant internal organs, combined with audio and text to demonstrate the internal organs that cooperate with the movement of qi and blood in the human body.

[0054] The pulse image formation mechanism module demonstrates the concept and formation mechanism of pulse images in a graphics and text way, showing the concept and formation mechanism of pulse images in text form, such as Figure 4 shown. It is the trajectory of blood vessel pulsation, mainly integrating the content such as the ejection activity of the heart and various information carried by the pulse along the blood vessel tree during propagation, and attaching the waveform of a cycle of pulse wave, enabling students to more intuitively understand pulse images.

[0055] The pulse-taking module demonstrates pulse-taking positions and pulse-taking finger methods in an animated, audio, and graphics and text way, such as Figure 5 shown. For example, the pulse-taking position uses technologies such as virtual simulation to present the consulting room, the positions of virtual patients, and the pulse-taking actions of virtual doctors in a three-dimensional and panoramic manner, and successively demonstrates the operation key points with icons, such as placing the arm flat at the same horizontal level as the heart, and the pulse-taking finger forming a 45° angle with the skin of the wrist, etc.; as Figure 6As shown in the figure, the finger methods for feeling the pulse are further detailed into positioning, finger placement, and finger movement. Taking the "finger movement" part as an example, the finger methods such as lifting method, seeking method, single pressing, and overall pressing are demonstrated on the 3D model hand in sequence, accompanied by audio and subtitle explanations. During the use process, students can click the mouse to watch the pulse-taking steps in sequence. They need to select the correct operation options regarding "technical points" that appear on the screen to continue. Students can also freely choose different finger methods to repeatedly observe the operation details.

[0056] This eight-element module demonstrates the eight elements of the pulse condition in the form of animation combined with pictures and texts, such as Figure 7 As shown, it uses a 3D model of forearm anatomy to present the mechanical and histomorphological differences of different pulse signals, and highly restores different pulse elements that are difficult to observe in the real world, such as the hemodynamics and histology changes at the cun-kou area, in the form of three-dimensional frequency animations. At the same time, it is accompanied by audio, subtitles, and traditional pulse diagram illustrations for interpretation. Clicking on the eight-element buttons of "pulse position", "rhythm", "pulse length", "pulse width", "pulse force", "evenness", "fluency", and "tension" can play the three-dimensional animation change videos showing each pulse element under different finger method operations, which helps students intuitively and vividly understand the connotations of the eight elements of the pulse condition. In order to strengthen the connection between pulse diagnosis experiment learning and clinical actual work, emphasizing both theory and practice, the learning content of "clinical significance of common pulse characteristics" is added to explain the clinical significance, physiological and pathological states represented by the pulse condition, etc., and guide students to combine pulse characteristics with clinical practice.

[0057] II. Machine analysis learning module: It demonstrates the working principle of the pulse detector, the process of pulse signal acquisition and analysis in the form of animation, voice, and / or combination of pictures and texts. Mainly based on the research results of pulse diagnosis, using 3D modeling and virtual reality technologies, it can visually present the composition, working principle, and the process and technology of pulse diagram acquisition and analysis of the pulse detector.

[0058] This machine analysis learning module includes a pulse detector module and an analysis module. The pulse detector module demonstrates the various components of the pulse sensor, their functions, assembly process, and working principle in the form of voice and combination of pictures and texts, and also demonstrates the various components of the pulse transducer, their functions, and working principle in the form of voice and combination of pictures and texts, such as Figure 8 As shown, it uses 3D modeling to highly simulate the key components of the pulse sensor and the pulse transducer, and visually presents each component and its working process. During the experiment, students can click to identify each component and complete the assembly of the pulse sensor through human-computer interaction, and conduct pulse signal transmission under the guidance of the system, which can better understand the working principle of the pulse sensor and the pulse transducer for pulse acquisition and amplification.

[0059] As Figure 9 shown, this analysis module demonstrates the entire process of pulse diagram acquisition, the entire analysis process of the acquired pulse diagram, and demonstrates the recognition results of conventional pulse diagrams in the form of combination of pictures and texts.

[0060] Students can connect the pulse transducer and the pulse sensor by dragging with the mouse, start the pulse acquisition and analysis software system, select virtual patients, conduct pulse acquisition, and under the guidance of the system, perform noise reduction, drift removal, cycle segmentation, and select different pulse graph analysis methods to extract features from the pulse graph, such as time-domain analysis method / hemodynamic analysis method, and finally generate a complete pulse graph analysis report. Through the training of the content of this module, students' research ideas on the modernization of traditional Chinese medicine pulse diagnosis and the innovation of traditional Chinese medicine engineering can be inspired, providing references for participating in scientific research innovation projects, the "Challenge Cup" extracurricular academic and technological works competition for college students, and the national "Internet +" college students' innovation and entrepreneurship competition, etc.

[0061] III. Inquiry Training Module: Simulate the medical scenario of pulse diagnosis between doctors and patients, and realize the training and assessment of the clinical thinking of traditional Chinese medicine of combining the four diagnostic methods and syndrome differentiation and treatment in the form of online human-computer Q&A.

[0062] Based on the large language model of artificial intelligence, this module constructs a simulation teaching scenario by using virtual digital patients and virtual medical scenarios, designs teaching content according to the teaching syllabus of traditional Chinese medicine diagnostics and the examination syllabus of practical skills for traditional Chinese medicine practicing physicians, supports online traditional Chinese medicine inquiry and syndrome differentiation training, and trains students' clinical thinking abilities of traditional Chinese medicine such as doctor-patient communication, combining inquiry and differentiation, combining the four diagnostic methods, and syndrome differentiation and treatment. As Figure 10 shown, the system includes teaching application scenarios such as learning mode, advanced training mode, and examination mode, supports various communication methods such as real-time voice, and can give learners information feedback based on the natural language context through AI virtual patients.

[0063] This module adopts a layered microservice architecture, constructs the asynchronous core service of the backend based on the Python FastAPI framework, realizes the separation of the front and backend through RESTful API, the front end realizes a dynamic interactive interface with the Vue.js, HTML5, and CSS3 frameworks, and its data layer selects the MySQL 8.2 version to store the case library, patient information, and inquiry records, and realizes the millisecond-level retrieval of symptom keywords through vector transformation and retrieval; the knowledge layer constructs a traditional Chinese medicine syndrome-symptom association model through the Neo4j knowledge graph to support dynamic syndrome differentiation reasoning; the AI service layer integrates the large language models DeepSeek, Doubao, and ZhiPu to realize natural language interaction and intelligent syndrome differentiation.

[0064] Specifically, the core function of this inquiry training module is to use natural language interaction to complete clinical information collection and give an evaluation result through intelligent dialectics;

[0065] For the first function, by constructing a case management database and virtualizing digital patients, integrating multiple speech recognition APIs or text recognition technologies of Azure and AliYun to parse the questions asked by the trainers playing doctors, and using an intelligent large model for analysis and combining the information of virtualized digital patients to generate real-time voice or text answers, natural language interaction is achieved to complete clinical information collection. The intelligent large model is constructed by integrating large language models DeepSeek, Doubao, and ZhiPu; the case management database is provided with a standardized template by the background, supporting the entry of symptoms (such as chief complaints, accompanying symptoms), signs (such as tongue manifestations, pulse conditions, etc.), and syndrome types, and associating with traditional Chinese medicine disease and syndrome standards, and score assignment and verification can be realized through a rule engine;

[0066] The virtualized digital patient is based on a digital standard human model and uses 3D virtual simulation technology to construct a case virtual patient, which can select a character model by gender and age, and supports online selection of voice models to generate differentiated cases.

[0067] For the second function, the trainer writes the corresponding case according to the collected clinical information, and then uses the traditional Chinese medicine syndrome-symptom association model to identify and analyze the case content and give an evaluation result. The traditional Chinese medicine syndrome-symptom association model is constructed by analyzing the association path between symptom combinations and syndrome types through a Neo4j knowledge graph;

[0068] To implement a stepped training plan for students' interrogation skills and evaluate the training results of students to understand their mastery, this module includes three modules: learning mode, advanced mode, and exam mode. The learning module is a learning mode based on virtualized digital patients, which can experience the intelligent teaching of the AI robot learning companion "Xiaoyou" and is suitable for beginners; the advanced module can freely select cases for training, set a time limit during training, and provide detailed structured evaluations to help students quickly improve their skills; the exam module can only be entered after the teacher issues an exam task and is used for teachers to master the learning situation of students.

[0069] Specifically, in the learning mode, under the auxiliary guidance of the AI robot learning companion, the trainer playing the doctor and the virtualized digital patient complete the clinical information collection of the virtualized digital patient through natural language-style questions and answers. Then the trainer writes a case and gives an evaluation result.

[0070] 1. Input and recording of interrogation information

[0071] At the bottom of the learning interface is the interrogation information input module. Clicking on the avatar of the virtualized digital patient can view the patient's basic information. By clicking the "Press and hold to speak" button, you can select the press-and-hold-to-speak and text input modes. As Figure 11 shown, the questions asked by the learner and the answers of the virtualized digital patient will be displayed in the "Interrogation Record" on the left side of the interface.

[0072] Click on the avatar of the AI robot learning companion "Xiaoyou" at the top of the "Medical Consultation Record" to switch the conversation with the robot learning companion "Xiaoyou". Learners can ask questions to the robot learning companion "Xiaoyou", and at the same time, the robot learning companion "Xiaoyou" will provide guiding assistance. Click on the patient avatar at the top of the "Medical Consultation Record" to return to the consultation with the patient.

[0073] 2. Four Diagnostic Information Collection and Medical Record Writing

[0074] As Figure 12 shown, clicking on the function buttons on the right side of the interface can respectively obtain the information of observing the head and face and tongue of the patient for inspection, the information of smelling odors and listening to sounds of the patient for auscultation, and the information of pulse types and pulse shape diagrams obtained by palpating the patient. Finally, click on the medical record icon at the bottom of the right function bar to write the medical record. It is necessary to input the patient's chief complaint, current medical history, past medical history, personal life history, family history, and the situation of the four traditional Chinese medicine diagnoses, and independently make a diagnosis of traditional Chinese medicine disease names and traditional Chinese medicine syndromes, and finally establish the treatment principles, methods, and prescription medications.

[0075] The learning mode uses virtual patients + medical intelligent large language models, allowing learners to communicate with virtual patients in natural language, providing a consultation experience that highly restores the real clinical consultation scenario. The AI robot learning companion has teaching guidance and auxiliary functions, and will provide operation assistance when the learner first enters the page to help beginners quickly adapt to and understand the system interface and functions. Learners can seek help from the robot learning companion "Xiaoyou" in the learning mode. The robot learning companion "Xiaoyou" can locate the current progress of the learner's medical history collection and quickly give guiding assistance feedback. The teaching guidance function of the robot learning companion "Xiaoyou" is reflected in the analysis of the current medical history collection content, the analysis, push of missed medical history content, the guidance and push of the content to continue collecting, and the analysis and guidance push of consultation skills such as humanistic care.

[0076] Regarding humanistic care, its core concept is that doctors need to identify the emotional needs of patients during communication and establish a trust relationship through different levels of responses (from neglect to deep empathy). Therefore, based on the trainer's collection of the four diagnostic information of virtual digital patients, the intelligent large model is used for identification and analysis to obtain empathy keywords, and then combined with the Empathic Communication Coding System (ECCS) empathy communication encoder to provide empathy push for virtual digital patients according to the ECCS dimension, and make judgments and analyses on the trainer's responses. It can detect the trainer's emotional attention to the patient, make a detailed analysis of whether the trainer provides suggestions and whether the suggestions are reasonable, and give evaluations.

[0077] ECCS (Empathic Communication Coding System) is a system for evaluating empathic ability in doctor-patient communication. It is a tool for assessing the level of empathic communication between doctors and patients in a medical scenario. Its core lies in quantifying the quality of doctors' responses to patients' emotions, progress, and challenges through a grading system. Its core concept is that doctors need to identify patients' emotional needs during communication and establish a trusting relationship through different levels of responses (from ignoring to deep empathy).

[0078] ECCS classifies empathic responses into 7 levels (0 - 6), ranging from completely ignoring the empathic opportunity of patients to sharing emotions or experiences with patients, as follows:

[0079] · Level 0: Denial / Disconfirmation, where doctors completely ignore the empathic opportunity of patients.

[0080] · Level 1: Perfunctory Confirmation, where doctors only give simple responses such as "um", "I see", etc.

[0081] · Level 2: Implicit Confirmation, where doctors implicitly confirm the secondary issues in the empathic opportunity.

[0082] · Level 3: Confirmation, where doctors confirm the main issues in the empathic opportunity.

[0083] · Level 4: Confirmation and Probing, where doctors confirm the main issues in the empathic opportunity and further ask questions or make comments.

[0084] · Level 5: Confirming Emotions, where doctors confirm the emotions expressed by patients.

[0085] · Level 6: Sharing Emotions or Experiences, where doctors explicitly express sharing emotions or similar experiences with patients.

[0086] In the advanced mode, without the assistance and guidance of the AI robot learning companion, the trainer playing the doctor and the virtual digital patient complete the clinical information collection of the virtual digital patient through natural language Q&A. Then the trainer writes the case and gives the evaluation results.

[0087] In the advanced mode, as Figure 13 shown, the system provides richer cases for students to choose by themselves to deal with complex and changeable clinical patients. Since there is no longer an AI robot learning companion, students receive less intelligent assistance compared to the learning module. More importantly, students need to give full play to their subjective initiative to independently complete the collection of medical history information and comprehensively judge the syndrome and disease name based on the four diagnostic information. Therefore, it can be regarded as an early simulation of the exam mode, which is convenient for students to self-check their learning situation, identify deficiencies, and also increases the assessment of the efficiency of medical history taking, which needs to be completed within 15 minutes, truly restoring the limited consultation time in clinical practice.

[0088] Similar to the learning mode, after students complete the medical record writing and diagnosis analysis, the system supports automatic medical record generation and automatic scoring and review. Students can see the score of the scoring points in the feedback review. The system conducts intelligent analysis of learners from five aspects: medical consultation content review, diagnosis analysis, traditional Chinese medicine treatment, medical consultation skills, and humanistic care. Figure 14 As shown, the review of consultation content includes consultation records, missed questions, and unscored questions; diagnosis analysis includes disease name diagnosis and syndrome diagnosis; TCM treatment includes treatment principles and methods, prescription names, and drug composition; consultation skills include organization and focus, consultation sequence, and communication skills. The system will display detailed scoring criteria so as to give scores to each module in a transparent, fair and impartial manner and make a comprehensive evaluation. Students can use this to understand their mistakes and shortcomings, and reflect and improve themselves.

[0089] like Figure 15 As shown in the figure, the system will summarize the growth records of students in the past 20 consultations, and use bar charts to visually present the strengths and weaknesses of students in various aspects of their abilities, and make formative evaluations. The suggestions made by students in the "Course Evaluation" will be reflected in the next update of the system, forming a good cycle of two-way feedback.

[0090] The examination mode is used to conduct online assessment of trainees.

[0091] It is only used when teachers issue exam tasks, and students complete the exam within the specified time to ensure the rigor and practicality of the assessment. The assessment results are combined with the scores of the interview robot and the teachers to ensure the objectivity and comprehensiveness of the evaluation.

[0092] Although specific embodiments of the present invention are described above, those skilled in the art should understand that these are merely examples and that various changes or modifications may be made to these embodiments without departing from the principles and essence of the present invention. Therefore, the scope of protection of the present invention is limited by the appended claims.

Claims

1. A digitalized traditional Chinese medicine pulse diagnosis system based on VR and AI technologies, characterized in that: It includes a basic knowledge learning module, a machine analysis learning module, and an interrogation training module. The basic knowledge learning module demonstrates the origin of pulse diagnosis, the formation mechanism of pulse diagrams / pulse conditions, pulse-taking methods, and the eight elements of pulse conditions in a way that combines animation, voice, and / or graphics and text. The machine analysis learning module demonstrates the working principle of a pulse condition detector, as well as the process of pulse condition signal acquisition and analysis, in a way that combines animation, voice, and / or graphics and text. The interrogation training module is used to simulate the medical scenario of pulse diagnosis between doctors and patients, and realizes the training and assessment of the traditional Chinese medicine clinical thinking of combining the four diagnostic methods and syndrome differentiation and treatment through an online human-machine Q&A method.

2. The digital Chinese medicine pulse diagnosis system based on VR and AI technologies according to claim 1, wherein: The interrogation training module analyzes the questions asked by the trainer playing the doctor by integrating multiple speech recognition APIs or text recognition technologies such as Azure and AliYun through building a case management database and virtual digital patients, and uses an intelligent large model for analysis and combines the information of virtual digital patients to generate voice or text answers in real time to achieve natural language interaction, complete clinical information collection. Then, the trainer writes the corresponding case according to the collected clinical information, and then uses the traditional Chinese medicine syndrome-symptom association model to identify and analyze the case content and give an evaluation result.

3. The digitalized traditional Chinese medicine pulse diagnosis system based on VR and AI technologies according to claim 2, characterized in that: The case management database provides a standardized template, supports the entry of symptoms, signs, and syndrome types, and is associated with the traditional Chinese medicine disease and syndrome standards. Based on a digital standard human model, the virtual digital patient constructs a virtual patient for cases using 3D virtual simulation technology, can select a character model by gender and age, and supports online selection of voice models. The traditional Chinese medicine syndrome-symptom association model is constructed by analyzing the association path between symptom combinations and syndrome types through a Neo4j knowledge graph. The intelligent large model is constructed by integrating large language models such as DeepSeek, Doubao, and ZhiPu.

4. The digital Chinese medicine pulse diagnosis system based on VR and AI technologies according to claim 3, characterized in that: The interrogation training module includes a learning mode, an advanced mode, and an exam mode. In the learning mode, with the assistance and guidance of an AI robot learning companion, the trainer playing the doctor and the virtual digital patient complete the clinical information collection of the virtual digital patient through natural language-style Q&A. Then, the trainer writes a case and gives an evaluation result. In the advanced mode, without the assistance and guidance of an AI robot learning companion, the trainer playing the doctor and the virtual digital patient complete the clinical information collection of the virtual digital patient through natural language-style Q&A. Then, the trainer writes a case and gives an evaluation result. The exam mode is used to conduct an online assessment of the trainer.

5. The digitalized traditional Chinese medicine pulse diagnosis system based on VR and AI technologies according to claim 1, characterized in that: The basic knowledge learning module includes an origin module, a formation mechanism module, a pulse-taking module, and an eight-elements module. The origin module demonstrates the origin of pulse diagnosis in an animated way. The formation mechanism module includes a pulse condition formation mechanism module and a pulse diagram formation mechanism module. The pulse condition formation mechanism module uses a page-turning mode to respectively show the formation mechanism of the pulse, the operation of qi and blood in the human body, and the cooperation of internal organs. The pulse diagram formation mechanism module demonstrates the concept and formation mechanism of the pulse diagram in a way that combines graphics and text. The pulse-taking module demonstrates the pulse-taking position and pulse-taking finger methods in a way that combines animation, audio, and graphics and text. The eight-elements module demonstrates the eight elements of pulse conditions in a way that combines animation and graphics and text.

6. The digitalized traditional Chinese medicine pulse diagnosis system based on VR and AI technologies according to claim 5, characterized in that: The pulse formation mechanism module uses animations of the heart beating and blood flowing in blood vessels, combined with audio, graphics, and text to demonstrate the pulse formation mechanism; uses animations of blood flowing in the blood vessels of the entire human body, combined with audio, graphics, and text to demonstrate the circulation of qi and blood in the human body; uses animations of the heart beating in coordination with relevant organs, combined with audio and text to demonstrate the organs that coordinate the circulation of qi and blood in the human body.

7. The digitalized traditional Chinese medicine pulse diagnosis system based on VR and AI technologies according to claim 1, characterized in that: The machine analysis and learning module includes a pulse detector module and an analysis module. The pulse detector module uses a combination of voice, graphics, and text to demonstrate the components of the pulse sensor, their functions, the assembly process, and the working principle, and uses a combination of voice, graphics, and text to demonstrate the components of the pulse transducer, their functions, and the working principle. The analysis module uses a combination of voice, graphics, and text to demonstrate the entire process of collecting the pulse diagram, demonstrates the entire analysis process of the collected pulse diagram, and uses a combination of graphics and text to demonstrate the recognition results of the conventional pulse diagram.

8. The digital Chinese medicine pulse diagnosis system based on VR and AI technologies according to claim 7, characterized in that: The entire analysis process of the collected pulse diagram includes noise reduction, drift removal, analysis, time domain analysis / hemodynamic analysis, and a pulse diagram analysis report.

Images