An artificial intelligence-based emergency patient triage system

Abstract

The present invention relates to the technical field of medical triage, and discloses an emergency patient triage system based on artificial intelligence, including: a data acquisition module: used to obtain the patient's identity and basic information by performing face recognition and fingerprint recognition on the patient, and receive health data uploaded by a wearable device, and transmit it to the patient's corresponding historical pathology library; a medical interaction module: used to provide an interactive platform for patients to remotely input pathological conditions based on 5G wireless communication technology; a multimodal data fusion module; a pathological feature matching module; a triage matching module; and a system integration and display module. Through the convolutional neural network model combined with pathological feature matching, based on the patient's question reply content, the condition and the judgment of the severity, a rapid judgment of the possibility of the disease, the generation of consultation content and the optimization of triage results are achieved, thereby improving the accuracy and efficiency of triage and reducing misdiagnosis and missed diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of medical triage, and in particular to an emergency patient triage system based on artificial intelligence. Background Art

[0002] With the rapid development of modern society, emergency medical services are facing increasing challenges and demands. Urban populations continue to gather, and the incidence of sudden illnesses and accidents continues to rise. As the forefront of medical treatment, the emergency room needs to receive a large number of patients every day. The use of artificial intelligence for triage provides rich technical means and possibilities for reconstructing and optimizing the emergency patient triage process in terms of data processing, information transmission, and identity recognition. Traditional triage methods mainly rely on the experience and judgment of medical staff for triage.

[0003] However, with current technology, manual triage often finds it difficult to accurately assess and quickly triage patients' conditions in a short period of time when faced with complex and diverse conditions, resulting in reduced triage efficiency and accuracy. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides an emergency patient triage system based on artificial intelligence to solve the problem that it is difficult to accurately assess and quickly triage patients' conditions in a short period of time.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: an emergency patient triage system based on artificial intelligence, comprising:

[0006] Data collection module: used to obtain the patient's identity and basic information through face recognition and fingerprint recognition, and receive health data uploaded by wearable devices and transmit it to the patient's corresponding historical pathology library;

[0007] Medical interaction module: It is used to provide an interactive platform for patients to remotely input their pathological status based on 5G wireless communication technology;

[0008] Multimodal data fusion module: used to integrate multimodal data of patients and extract pathological features, and to form a comprehensive patient feature vector by fusing features of different modalities;

[0009] Pathological feature matching module: used to match the characteristic data of pathological status through the pathological database, and compare to determine the possibility of disease, and propose single or multiple consultation contents based on possible diseases;

[0010] Triage matching module: determines the severity of the patient, outputs triage suggestions, and matches the triage information to the corresponding medical department;

[0011] System integration and display module: used to integrate with the hospital system and display the patient's triage results and pathological information to patients and medical staff through an interactive platform.

[0012] Preferably, the data acquisition module includes a biometric identification unit, a device data receiving unit, and a data storage and association unit;

[0013] The biometric identification unit is used to capture the patient's facial image through face recognition technology, and compare and match it with the facial image template pre-stored in the hospital database or import new patient information into the hospital database;

[0014] The device data receiving unit is used to collect data from the patient's wearable device and other medical monitoring devices in real time;

[0015] The data storage and association unit is used to store the patient identity and basic information obtained through the biometric identification unit, as well as various types of health data collected by the device data receiving unit, establish associations, and store them in the hospital database.

[0016] Preferably, the medical interaction module includes a wireless remote communication unit, a speech processing unit and a dialogue consultation unit;

[0017] The wireless remote communication unit is used to establish and maintain a wireless communication connection with the remote medical terminal, and receive patient expression information, including voice expression and text expression;

[0018] The speech processing unit is used to process and recognize the patient's speech, including the conversion between speech and text;

[0019] The dialogue consultation unit is used to generate relevant consultation questions based on the preset medical consultation template and disease knowledge base, according to the information initially provided by the patient, through a generative pre-trained language model.

[0020] Preferably, the multimodal data fusion module includes a data preprocessing unit, a feature extraction unit and a feature fusion unit;

[0021] The data preprocessing unit is used to perform preliminary processing on the collected raw data and convert different health data formats into a computer format;

[0022] The feature extraction unit is used to extract keywords, phrases, and semantic vector features from text data, and simultaneously extract pathological features of health data uploaded by wearable device data;

[0023] The feature fusion unit is used to fuse the features extracted from different modal data through feature layer fusion to form a patient feature vector.

[0024] Preferably, the pathological feature matching module includes a pathological database management unit, a feature extraction and comparison unit, and a consultation content generation unit;

[0025] The pathology database management unit is used to update the pathology diagnosis database in real time, and classify the collected pathology data based on the pathology diagnosis database;

[0026] The feature extraction and comparison unit is used to analyze the feature vector of the patient by learning and training the patient's diagnostic expression based on the medical history and the standard features in the pathological diagnosis database or features of other similar cases through a convolutional neural network model;

[0027] The consultation content generation unit is used to generate a triage recommendation result based on the results of the convolutional neural network model analysis and the patient's question response content in the dialogue consultation unit.

[0028] Preferably, the triage matching module includes a disease severity determination unit, an information matching and sending unit, and a feedback and adjustment unit;

[0029] The severity determination unit is used to extract pathological features from the fused patient information, including vital sign data, symptom manifestations, medical history information and current examination and test results, and evaluate their impact on the severity of the disease;

[0030] The information matching and sending unit is used to match the patient's condition and department information, and send the triage results to the medical staff of the corresponding department and the patient or family members;

[0031] The feedback and adjustment unit is used to collect feedback information about triage from medical staff, patients and their families, and adjust the triage output results of the convolutional neural network model through the feedback information.

[0032] Preferably, the system integration and display module includes a system docking unit and a visual display unit;

[0033] The system docking unit is used to perform data interaction with the hospital's internal system;

[0034] The visualization display unit is used to transmit the acquired pathological data and triage results to the inside of the hospital, and display the data to the medical staff of the corresponding department.

[0035] An artificial intelligence-based emergency patient triage method comprises the following steps:

[0036] S1. First, obtain the patient's identity information through biometric technology and collect health data from wearable devices and other medical monitoring devices;

[0037] S2. Select a 5G wireless communication remote or in-hospital medical interaction platform to receive the patient's voice or text information;

[0038] S3, preprocessing, feature extraction and fusion of multimodal data to form a patient feature vector;

[0039] S4, matching the patient's feature vector with the pathology database, determining the possibility of the disease, and generating the consultation content;

[0040] S5. Output triage suggestions based on the severity of the disease, determine the triage results based on the content of the consultation, and then match the triage results to the corresponding medical department;

[0041] S6. Integrate the triage results with the hospital's internal system and display them to medical staff through a visual interface to locate the location of the triage department, including the internal department routes of the hospital.

[0042] The present invention provides an emergency patient triage system based on artificial intelligence. It has the following beneficial effects:

[0043] 1. The present invention combines the convolutional neural network model with pathological feature matching, the condition and the severity judgment based on the patient's question response content, to achieve rapid judgment of disease possibility, generate consultation content and optimize triage results, improve triage accuracy and efficiency, and reduce misdiagnosis and missed diagnosis.

[0044] 2. The present invention integrates multimodal patient data, forms a comprehensive patient feature vector through data preprocessing, feature extraction and fusion, makes up for the limitations of single modality data, provides rich and accurate information for precise diagnosis and triage, and improves the accuracy and reliability of disease diagnosis.

[0045] 3. The present invention relies on 5G wireless communication technology to build an interactive platform for remote input of pathological status, realize patient and remote triage, thereby achieving early triage and guidance, breaking through geographical restrictions, promoting the balanced distribution of medical resources, and improving the accessibility and overall efficiency of medical services. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is an architecture diagram of an emergency patient triage system based on artificial intelligence of the present invention;

[0047] Figure 2 This is a data acquisition module architecture diagram of an emergency patient triage system based on artificial intelligence of the present invention;

[0048] Figure 3 This is a medical interaction module architecture diagram of an artificial intelligence-based emergency patient triage system of the present invention;

[0049] Figure 4This is a multimodal data fusion module architecture diagram of an artificial intelligence-based emergency patient triage system of the present invention;

[0050] Figure 5 This is a diagram of the architecture of a pathological feature matching module of an emergency patient triage system based on artificial intelligence of the present invention;

[0051] Figure 6 This is an architecture diagram of a triage matching module of an emergency patient triage system based on artificial intelligence of the present invention;

[0052] Figure 7 This is a system integration and display module architecture diagram of an artificial intelligence-based emergency patient triage system of the present invention;

[0053] Figure 8 The present invention is a flowchart of an emergency patient triage method based on artificial intelligence. DETAILED DESCRIPTION

[0054] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Please refer to the attached Figure 1 -Attached Figure 7 The embodiment of the present invention provides an emergency patient triage system based on artificial intelligence, comprising:

[0056] Data collection module: used to obtain the patient's identity and basic information through face recognition and fingerprint recognition, and receive health data uploaded by wearable devices and transmit it to the patient's corresponding historical pathology library;

[0057] Medical interaction module: It is used to provide an interactive platform for patients to remotely input their pathological status based on 5G wireless communication technology;

[0058] Multimodal data fusion module: used to integrate multimodal data of patients and extract pathological features, and to form a comprehensive patient feature vector by fusing features of different modalities;

[0059] Pathological feature matching module: used to match the characteristic data of pathological status through the pathological database, and compare to determine the possibility of disease, and propose single or multiple consultation contents based on possible diseases;

[0060] Triage matching module: determines the severity of the patient, outputs triage suggestions, and matches the triage information to the corresponding medical department;

[0061] System integration and display module: used to integrate with the hospital system and display the patient's triage results and pathological information to patients and medical staff through an interactive platform.

[0062] Specifically, the data collection module collects multimodal key information of patients to provide comprehensive data support for subsequent medical work. Through face recognition and fingerprint recognition, the patient's identity can be accurately confirmed, and basic information such as name, age, and medical history can be quickly obtained to avoid medical errors caused by identity confusion. At the same time, health data uploaded by wearable devices, such as real-time heart rate and sleep monitoring, are received to dynamically track the patient's health status and transmit it to the patient's corresponding historical pathology library in real time, so as to more comprehensively and accurately evaluate the patient's condition for subsequent diagnosis, formulate personalized diagnosis and treatment plans, and improve the quality and efficiency of medical services.

[0063] Through the medical interaction module, an interactive platform for patients to input various pathological conditions is built, allowing patients to upload symptom descriptions, examination reports and other pathological information on the way to the hospital through the platform, and timely professional guidance and preliminary diagnosis are given, and triage is carried out in advance. While improving the accessibility of medical services, it can also make medical resources more reasonably allocated, improve overall medical efficiency, and realize the efficient implementation of telemedicine services;

[0064] The multimodal data fusion module integrates the multimodal data of patients, including symptom descriptions, examination reports and other pathological information, and fuses these pathological features to form a comprehensive patient feature vector, which effectively makes up for the limitations of single-modal data, provides more complete and accurate patient condition information, improves the efficiency of triage, and helps doctors to judge the condition more accurately during subsequent treatment;

[0065] The pathology feature matching module is based on the pathology database. By matching and comparing the patient's pathology status feature data with the massive data in the database, it can quickly and accurately determine the possibility that the patient has various diseases. Based on these possible diseases, it can propose single or multiple consultation contents in a targeted manner, further explore the patient's symptom details and medical history information, and improve the accuracy of triage.

[0066] The triage matching module can quickly and accurately determine the severity of the patient's condition, output precise triage suggestions based on the severity of the condition, disease type, etc., and intelligently match the patient with the corresponding medical department to ensure that the patient is sent to the appropriate department for treatment in a timely manner, avoiding blind medical treatment for patients and irrational allocation of department resources, and significantly improving the efficiency of medical services.

[0067] Through system integration and display modules, patients are linked to medical staff and hospitals. The triage results of patients are presented intuitively, allowing patients to quickly identify the direction of subsequent treatment and reduce waiting and confusion; at the same time, detailed pathological information is displayed to medical staff, so that they can grasp the full picture of the disease at the first time, without the need for tedious searches in multiple systems, improving the efficiency of information transmission, optimizing the medical process, promoting communication between doctors and patients, providing convenience for precision medicine, and comprehensively improving the quality and efficiency of medical services.

[0068] See attached Figure 2 ,The data acquisition module includes a biometric recognition unit, a device data receiving unit, and a data storage and association unit;

[0069] The biometric unit is used to capture the patient's facial image through face recognition technology and compare it with the facial image template pre-stored in the hospital database or import new patient information into the hospital database;

[0070] The device data receiving unit is used to collect data from the patient's wearable devices and other medical monitoring devices in real time;

[0071] The data storage and association unit is used to store the patient identity and basic information obtained through the biometric unit, as well as various types of health data collected by the equipment data receiving unit, establish associations, and store them in the hospital database.

[0072] Specifically, the biometric unit can accurately confirm the patient's identity, and the facial recognition technology can be used to capture the patient's facial image. By comparing it with the facial image template stored in the hospital database, the identity of old patients can be quickly verified, and their past medical records can be easily retrieved, helping medical staff to quickly understand the condition. For new patients, their information can be imported into the database to create files for subsequent diagnosis and treatment records. Improve the accuracy and convenience of medical processes;

[0073] The device data receiving unit collects patient health data in real time and continuously collects information from wearable devices and other medical monitoring devices, such as heart rate, blood pressure, sleep data, etc., to achieve dynamic monitoring of the patient's health status, thereby intervening in disease risks in advance, and quickly triaging and dividing departments, thereby improving the timeliness and preventiveness of medical services;

[0074] Through the data storage and association unit, various types of patient data are integrated and orderly stored. The patient identity and basic information obtained by the biometric unit and the health data collected by the equipment data receiving unit are stored in the hospital database and associated with each other, providing a comprehensive basis for accurate diagnosis and treatment plan formulation, and improving the scientificity and accuracy of medical decision-making.

[0075] See attached Figure 3,The medical interaction module includes a wireless remote communication unit, a speech processing unit and a dialogue diagnosis and inquiry unit;

[0076] The wireless remote communication unit is used to establish and maintain a wireless communication connection with the remote medical terminal and receive patient expression information, including voice expression and text expression;

[0077] The speech processing unit is used to process and recognize the patient's speech, including the conversion between speech and text;

[0078] The dialogue consultation unit is used to generate relevant consultation questions based on the preset medical consultation template and disease knowledge base, according to the information initially provided by the patient, through a generative pre-trained language model.

[0079] Specifically, a stable wireless communication connection is established and maintained through a wireless remote communication unit, so that patients can be remotely triaged anytime and anywhere, and whether they describe their symptoms in detail by voice or briefly describe their discomfort in text, it promotes the balanced allocation of medical resources and improves the accessibility of medical services;

[0080] The voice processing unit converts the patient's voice into text, which is convenient for doctors to view and analyze. It can also convert the text of the consultation content into voice and convey it to the patient, meeting different communication habits, optimizing the doctor-patient communication experience, making communication more convenient and efficient, and improving the accuracy of triage.

[0081] Through the dialogue consultation unit, relying on preset medical consultation templates and a rich disease knowledge base, combined with the initial information provided by the patient, the generative pre-trained language model is used to generate targeted consultation questions, guide patients to describe their condition more comprehensively, explore key information, and improve the accuracy and efficiency of diagnosis.

[0082] See attached Figure 4 ,The multimodal data fusion module includes a data preprocessing unit, a feature extraction unit and a feature fusion unit;

[0083] The data preprocessing unit is used to perform preliminary processing on the collected raw data and convert different health data formats into computer formats;

[0084] The feature extraction unit is used to extract keywords, phrases, and semantic vector features from text data, and also to extract pathological features of health data uploaded by wearable device data;

[0085] The feature fusion unit is used to fuse the features extracted from different modal data through feature layer fusion to form a patient feature vector.

[0086] Specifically, the data preprocessing unit will collect health data from different channels, such as imaging data collected by hospital equipment, text data recorded by patients themselves, and data monitored by wearable devices, for preliminary processing to eliminate compatibility issues caused by differences in data formats, lay a solid foundation for subsequent analysis, ensure that various types of data can be processed and analyzed smoothly, and improve data processing efficiency and accuracy.

[0087] The feature extraction unit mines key information from a variety of data and extracts features such as keywords, phrases, and semantic vectors from text data to help understand the meaning behind the text, which is the key basis for accurately judging the patient's health status. It provides rich and valuable information for disease diagnosis and health assessment, and assists doctors in accurately judging the condition.

[0088] The features extracted from different modal data are integrated through the feature fusion unit to form a comprehensive and representative patient feature vector, which comprehensively reflects the patient's health status and avoids the limitations of single modality data. It enables subsequent doctors to understand the patient's condition from multiple dimensions, thereby making more accurate diagnosis and treatment decisions, significantly improving the accuracy and reliability of disease diagnosis, and providing strong support for personalized medicine.

[0089] See attached Figure 5 ,The pathology feature matching module includes a pathology database management unit, a feature extraction and comparison unit, and a consultation content generation unit;

[0090] The pathology database management unit is used to update the pathology diagnosis database in real time, and classify the collected pathology data based on the pathology diagnosis database;

[0091] The feature extraction and comparison unit is used to analyze the patient's feature vector by using a convolutional neural network model, collecting the patient's diagnostic expression based on the medical history, and learning and training with the standard features in the pathological diagnosis database or features of other similar cases;

[0092] The consultation content generation unit is used to generate triage recommendation results based on the results of the convolutional neural network model analysis and the patient's question responses in the dialogue consultation unit.

[0093] Specifically, the pathology diagnosis database is updated in real time through the pathology database management unit to ensure that the data in the database is always up-to-date and accurate. Based on this continuously updated database, the collected pathology data is scientifically classified, such as by disease type, site of onset, severity, etc., to provide more accurate triage data for triage and improve the accuracy and efficiency of medical work.

[0094] The pathology database management unit collects the patient's diagnostic expressions based on the medical history, and conducts learning and training with the standard features in the pathology diagnosis database or features of other similar cases, so that the model has the ability to accurately identify features, find similarities and differences with known cases, and reduce the occurrence of triage errors;

[0095] The consultation content generation unit uses the analysis results of the convolutional neural network model and the patient's responses in the dialogue consultation unit to comprehensively consider the patient's condition characteristics, symptoms and other factors, and generates specific triage recommendation results to quickly clarify the patient's treatment direction.

[0096] See attached Figure 6 ,The triage matching module includes a disease severity judgment unit, an ,information matching and sending unit, and a feedback and adjustment unit;

[0097] The severity judgment unit is used to extract pathological features from the fused patient information, including vital sign data, symptom manifestations, medical history information and current examination and test results, and evaluate their impact on the severity of the disease;

[0098] The information matching and sending unit is used to match the patient's condition and department information, and send the triage results to the medical staff of the corresponding department and the patient or family members;

[0099] The feedback and adjustment unit is used to collect feedback information about triage from medical staff, patients and their families, and adjust the triage output results of the convolutional neural network model through the feedback information.

[0100] Specifically, the severity judgment unit is used to evaluate the impact of pathological characteristics on the severity of the disease, accurately classify the patient's disease level, and provide a key basis for subsequent triage, thereby arranging the patient's triage matching order and improving timeliness;

[0101] The information matching and sending unit matches the patient's condition according to the severity and type of the disease, combined with the professional fields, bed conditions, medical staff configuration and other information of each department of the hospital, to ensure that the patient can be sent to the most suitable department. At the same time, the triage results are sent to the medical staff of the corresponding department in a timely manner, so that they can prepare for the reception in advance, and also inform the patient or family members so that they can know where to go for treatment;

[0102] The feedback and adjustment unit collects feedback from all parties on the triage work, thereby adjusting the triage output results of the convolutional neural network model, optimizing the disease judgment criteria and triage rules, and continuously improving the triage matching module to make it more suitable for actual medical scenarios, continuously improving the accuracy and efficiency of triage, and providing patients with more efficient medical services.

[0103] See attached Figure 7,The system integration and display module includes a system docking unit and a visual display unit;

[0104] The system docking unit is used to exchange data with the hospital's internal system;

[0105] The visualization display unit is used to transmit the acquired pathological data and triage results to the hospital and display the data to the medical staff in the corresponding department.

[0106] Specifically, the system docking unit is responsible for building a data interaction bridge with the hospital's internal systems, achieving seamless docking with various hospital systems, such as the electronic medical record system, laboratory inspection system, and registration system, so that patient information can flow freely among different systems and improve the overall operational efficiency of the hospital;

[0107] By presenting key information such as pathological data and triage results obtained from the system docking unit to medical staff in the corresponding department, medical staff can quickly obtain key information without having to screen through large amounts of data, saving time, improving the efficiency of diagnosis and treatment, and assisting medical staff in making more accurate medical decisions.

[0108] See attached Figure 8 , an artificial intelligence-based emergency patient triage method, comprising the following steps:

[0109] S1. First, obtain the patient's identity information through biometric technology and collect health data from wearable devices and other medical monitoring devices;

[0110] S2. Select a 5G wireless communication remote or in-hospital medical interaction platform to receive the patient's voice or text information;

[0111] S3, preprocessing, feature extraction and fusion of multimodal data to form a patient feature vector;

[0112] S4, matching the patient's feature vector with the pathology database, determining the possibility of the disease, and generating the consultation content;

[0113] S5. Output triage suggestions based on the severity of the disease, determine the triage results based on the content of the consultation, and then match the triage results to the corresponding medical department;

[0114] S6. Integrate the triage results with the hospital's internal system and display them to medical staff through a visual interface to locate the location of the triage department, including the internal department routes of the hospital.

[0115] Specifically, biometric technology and equipment are used to collect patient information, 5G communication is used to receive patient statements, multimodal data is processed and fused to form feature vectors, which are matched with the pathology database to determine the disease and generate consultation content. Triage suggestions are output based on the condition to determine the results and match the department. Finally, it is integrated with the hospital system and displayed visually to medical staff. It can integrate multi-source information and use artificial intelligence technology to assist in triage, so as to quickly and accurately determine the patient's condition, optimize the triage process, improve emergency efficiency, help medical staff quickly locate the triage department, gain precious treatment time for patients, and improve the efficiency of medical resource utilization.

[0116] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An emergency patient triage system based on artificial intelligence, characterized in that: include: Data collection module: used to obtain the patient's identity and basic information through face recognition and fingerprint recognition, and receive health data uploaded by wearable devices and transmit it to the patient's corresponding historical pathology library; Medical interaction module: It is used to provide an interactive platform for patients to remotely input their pathological status based on 5G wireless communication technology; Multimodal data fusion module: used to integrate multimodal data of patients and extract pathological features, and to form a comprehensive patient feature vector by fusing features of different modalities; Pathological feature matching module: used to match the characteristic data of pathological status through the pathological database, and compare to determine the possibility of disease, and propose single or multiple consultation contents based on possible diseases; Triage matching module: determines the severity of the patient, outputs triage suggestions, and matches the triage information to the corresponding medical department; System integration and display module: used to integrate with the hospital system and display the patient's triage results and pathological information to patients and medical staff through an interactive platform.

2. The artificial intelligence-based emergency patient triage system according to claim 1, characterized in that: The data acquisition module includes a biometric recognition unit, a device data receiving unit, and a data storage and association unit; The biometric identification unit is used to capture the patient's facial image through face recognition technology, and compare and match it with the facial image template pre-stored in the hospital database or import new patient information into the hospital database; The device data receiving unit is used to collect data from the patient's wearable device and other medical monitoring devices in real time; The data storage and association unit is used to store the patient identity and basic information obtained through the biometric identification unit, as well as various types of health data collected by the device data receiving unit, establish associations, and store them in the hospital database.

3. The artificial intelligence-based emergency patient triage system according to claim 1, characterized in that: The medical interaction module includes a wireless remote communication unit, a voice processing unit and a dialogue inquiry unit; The wireless remote communication unit is used to establish and maintain a wireless communication connection with the remote medical terminal, and receive patient expression information, including voice expression and text expression; The speech processing unit is used to process and recognize the patient's speech, including the conversion between speech and text; The dialogue consultation unit is used to generate relevant consultation questions based on the preset medical consultation template and disease knowledge base, according to the information initially provided by the patient, through a generative pre-trained language model.

4. The artificial intelligence-based emergency patient triage system according to claim 1, characterized in that: The multimodal data fusion module includes a data preprocessing unit, a feature extraction unit and a feature fusion unit; The data preprocessing unit is used to perform preliminary processing on the collected raw data and convert different health data formats into a computer format; The feature extraction unit is used to extract keywords, phrases, and semantic vector features from text data, and simultaneously extract pathological features of health data uploaded by wearable device data; The feature fusion unit is used to fuse the features extracted from different modal data through feature layer fusion to form a patient feature vector.

5. The artificial intelligence-based emergency patient triage system according to claim 1, characterized in that: The pathological feature matching module includes a pathological database management unit, a feature extraction and comparison unit, and a consultation content generation unit; The pathology database management unit is used to update the pathology diagnosis database in real time, and classify the collected pathology data based on the pathology diagnosis database; The feature extraction and comparison unit is used to analyze the feature vector of the patient by learning and training the patient's diagnostic expression based on the medical history collection and the standard features in the pathological diagnosis database or features of other similar cases through a convolutional neural network model; The consultation content generation unit is used to generate a triage recommendation result based on the results of the convolutional neural network model analysis and the patient's question response content in the dialogue consultation unit.

6. The artificial intelligence-based emergency patient triage system according to claim 1, characterized in that: The triage matching module includes a disease severity judgment unit, an information matching and sending unit, and a feedback and adjustment unit; The severity judgment unit is used to extract pathological features from the fused patient information, including vital sign data, symptom manifestations, medical history information and current examination and test results, and evaluate their impact on the severity of the disease; The information matching and sending unit is used to match the patient's condition and department information, and send the triage results to the medical staff of the corresponding department and the patient or family members; The feedback and adjustment unit is used to collect feedback information about triage from medical staff, patients and their families, and adjust the triage output results of the convolutional neural network model through the feedback information.

7. The artificial intelligence-based emergency patient triage system according to claim 1, characterized in that: The system integration and display module includes a system docking unit and a visual display unit; The system docking unit is used to perform data interaction with the hospital's internal system; The visualization display unit is used to transmit the acquired pathological data and triage results to the inside of the hospital, and display the data to the medical staff of the corresponding department.

8. An artificial intelligence-based emergency patient triage method, characterized in that: An artificial intelligence-based emergency patient triage system for use in any one of claims 1 to 7, comprising the following steps: S1. First, obtain the patient's identity information through biometric technology and collect health data from wearable devices and other medical monitoring devices; S2. Select a 5G wireless communication remote or in-hospital medical interaction platform to receive the patient's voice or text information; S3, preprocessing, feature extraction and fusion of multimodal data to form a patient feature vector; S4, matching the patient's feature vector with the pathology database, determining the possibility of the disease, and generating the consultation content; S5. Output triage suggestions based on the severity of the disease, determine the triage results based on the content of the consultation, and then match the triage results to the corresponding medical department; S6. Integrate the triage results with the hospital's internal system and display them to medical staff through a visual interface to locate the location of the triage department, including the internal department routes of the hospital.

Images