Intelligent treatment method, system and equipment for AI virtual human accompanying diagnosis and medium

Abstract

The invention discloses an intelligent doctor seeing method, system and device for AI virtual person accompanying diagnosis and a medium. The method comprises the steps that an intelligent hospital guide demand instruction is received, a virtual consulting room is constructed, and an AI virtual person is generated; after it is judged that the user enters the virtual consulting room, user input information is obtained in real time; determining a user object type and a user hospital guide application scene according to the user input information; the abnormal emotion of the user is recognized, and a preset emotion pacifying strategy is matched, so that the AI virtual human interacts with the user according to the matched emotion pacifying strategy; according to the real-time interaction information of the user and the historical intelligent hospital guide information of the user, matching a preset deep learning model in combination with a hospital guide application scene of the user, and outputting AI virtual human interaction reaction information and a preliminary diagnosis result; and determining a registration department according to the preliminary diagnosis result, generating registration appointment recommendation and completing registration appointment. According to the invention, the user experience in the intelligent treatment process of AI virtual human accompanying diagnosis can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence medical care, and specifically to an intelligent medical consultation method, system, equipment and medium with AI virtual human accompanying the patient. Background Art

[0002] With the development of Internet technology, online medical consultation has become increasingly popular, and more and more patients choose to conduct preliminary consultations through online platforms. This online consultation method not only improves the convenience and accessibility of medical services, but also provides patients with preliminary medical advice. However, most platforms only provide text or video communication, lack intelligent consultation and consultation planning services, making it difficult for patients to obtain comprehensive and professional guidance before consultation.

[0003] In order to solve the above problems, the following methods are usually used in the existing technology: First, patients judge their own condition and choose the department and doctor. Second, the customer service staff on the platform assists patients in judging their condition and choosing the appropriate department and doctor. Third, some platforms provide some simple medical process planning services, but these services are only based on basic information and common symptoms, and lack personalized and intelligent recommendations. They cannot achieve intelligent interaction for different consultation objects and consultation needs, complete preliminary diagnosis, and generate intelligent recommendations.

[0004] There are also some intelligent accompanying systems in the existing technology to assist users in completing intelligent medical consultations, but these systems mainly focus on improving the efficiency of medical services before, during and after the consultation by virtual human accompanying patients, and ignore the improvement of user medical experience such as emotional recognition and comfort of the patient during the consultation, and different needs of different patients and different application scenarios for medical consultation processes and intelligent recommendations. Summary of the invention

[0005] In order to improve the user experience during the intelligent medical consultation process accompanied by AI virtual people, the present application provides an intelligent medical consultation method, system, device and medium accompanied by AI virtual people.

[0006] In a first aspect, the present application provides an intelligent medical consultation method using an AI virtual person as an accompaniment, comprising: Receive intelligent guidance demand instructions sent by user terminals, build a virtual clinic and generate AI virtual people; After determining that the user enters the virtual clinic based on the user terminal operation, the user input information is obtained in real time; the user input information includes pre-entered user portrait information, user selected intelligent diagnosis scene information, user real-time interaction information and user historical intelligent diagnosis information; Determine the user object type according to the acquired user portrait information and / or the user's historical intelligent diagnosis information; the user object type includes: child type, adult type, elderly type and interaction disorder group type; Determine the user's medical guidance application scenario based on the acquired user's historical intelligent medical guidance information and the intelligent medical guidance scenario information selected by the user; the user medical guidance application scenario includes: general consultation, emergency consultation, general follow-up consultation and emergency follow-up consultation; Using deep learning algorithms, the user's abnormal emotions are identified based on the user's real-time interaction information. The abnormal conditions include anxiety, tension, and sadness. Based on the identified abnormal emotions, a preset emotion soothing strategy is matched with the determined user object type, so that the AI ​​virtual person interacts with the user according to the matched emotion soothing strategy. Among them, each user object type is matched with each abnormal emotion by a preset emotion soothing strategy. Matching a preset deep learning model according to the determined user guidance application scenario; outputting AI virtual human interaction reaction information and preliminary diagnosis results in combination with the matched deep learning model according to the user's real-time interaction information and the user's historical intelligent guidance information; wherein, the number of AI virtual human interaction reaction information output by different preset deep learning models and the output time of preliminary diagnosis results are different; Determine the registration department based on the preliminary diagnosis results, and generate a registration appointment recommendation based on the registration department; receive the registration appointment request information sent by the user terminal based on the registration appointment recommendation, connect to the hospital information system of the selected registration department and complete the registration appointment.

[0007] By adopting the above solution, the user object type and guidance application scenario are determined based on user portrait information and historical intelligent guidance information, providing a basis for the subsequent intelligent accompanying of AI virtual people; deep learning algorithms are used to identify abnormal user emotions and match emotional soothing strategies that match user types, effectively soothing users' emotions during the medical treatment process and improving user experience; intelligent matching of deep learning models that meet application scenarios, adaptively simplifying medical process planning and timely output of preliminary diagnosis results, improving user experience.

[0008] Preferably, determining the registration department according to the preliminary diagnosis result and generating a registration appointment recommendation according to the registration department includes: Obtain the geographical location information of the corresponding user terminal based on the last interaction information between the user and the AI ​​virtual person received before the output of the preliminary diagnosis result; obtain the current number of appointments in the hospital with the determined registration department based on the preliminary diagnosis result; Combined with the determined user object type and the user guidance application scenario, the acquired geographical location information of the user terminal and the current number of appointments in the hospital with the determined registration department, a registration appointment recommendation sequence satisfying the first rule is generated according to the user's historical intelligent guidance information and the determined registration department; The first rule includes: the hospital where the registration department corresponding to the ordinary follow-up and emergency follow-up in the user guidance application scenario is located is the same as the hospital where the original registration department is located, and the current number of appointments for the hospital with a determined registration department is less than the maximum appointment threshold for the corresponding hospital matched by the user guidance application scenario, and the hospital with a determined registration department has a smaller current number of appointments, and the hospital with the smaller the number of appointments, the more priority it is recommended, and the hospital where the registration department is located is closer to the geographical location information of the user terminal, and the hospital where the registration department is located is a user object type exclusive hospital, and priority recommendation is given to the hospital; wherein, different user guidance application scenarios are set with corresponding matching maximum appointment thresholds; the priority of priority recommendation generated based on the geographical location information of the user terminal, the priority of priority recommendation order generated based on the user object type, and the priority of priority recommendation generated based on the number of appointments are determined according to the priority level pre-set by the user, and the larger the priority level, the higher the corresponding priority recommendation order.

[0009] By adopting the above solution, when intelligently generating registration appointment recommendations, the geographical location information of the user terminal, the number of appointments in the current hospital, the user object type, and the user guidance application scenario are comprehensively considered, so as to generate a more accurate registration appointment recommendation order, ensure that patients receive more reasonable medical arrangements, reduce waiting time, and improve medical efficiency and user experience.

[0010] Preferably, it also includes: The generated preliminary diagnosis report is sent synchronously when the hospital information system of the appointment department is connected; when the preliminary diagnosis report request instruction is received from the doctor's terminal of the appointment department in the hospital information system, the preliminary diagnosis report is sent to the doctor's terminal of the appointment department; the preliminary diagnosis report is generated by integrating the preliminary diagnosis result with the key information extracted from the user's real-time interaction information converted into text content and the AI ​​virtual human interaction response information by using natural language technology; Receive the real doctor's recommended examination items uploaded by the doctor's terminal of the appointment department, determine the examination items required for the confirmation of the preliminary diagnosis result according to the medical knowledge graph, compare the determined examination items with the real doctor's recommended examination items for similarity, and if the similarity is lower than the first preset similarity, generate a prompt message for reconfirmation of the real doctor's recommended examination items and display the examination items required for the confirmation of the preliminary diagnosis result, and when receiving the reconfirmation of the real doctor's recommended examination items, use the real doctor's recommended examination items as the final recommended examination items; After receiving the results of various examination items of the user uploaded by the hospital information system, the secondary diagnosis result is output according to the results of various examination items of the user, the real-time interaction information of the user and the historical intelligent guidance information of the user, combined with the deep learning algorithm; the real doctor's diagnosis result uploaded by the terminal of the doctor in the registration department of the appointment is received, and the similarity is compared with the generated secondary diagnosis result. If the similarity is lower than the second preset similarity, the real doctor's diagnosis result re-confirmation information is generated and the preliminary diagnosis result is displayed on the terminal of the doctor in the registration department. When the real doctor's diagnosis result confirmation information is received, the real doctor's diagnosis information is used as the final diagnosis result.

[0011] By adopting the above scheme, the basic information in the user's preliminary diagnosis report is informed to the doctor so that the doctor can understand the patient's condition in advance; at the same time, by comparing the similarity between the preliminary diagnosis report and the examination items recommended by the doctor, a confirmation message is generated to prompt the doctor when the similarity is low, thereby ensuring the accuracy of the examination items; after the user completes all the examination items, a secondary diagnosis result is generated again in combination with the examination results and compared for similarity with the doctor's diagnosis result. When the similarity is low, a confirmation message is generated to prompt the doctor, thereby ensuring the accuracy of the final diagnosis result.

[0012] Preferably, it also includes: When receiving the first route planning request instruction sent by the user terminal, the location information of each department of the hospital and the current density of people in each area of ​​the hospital are obtained, and the determined user object type and the user guidance application scenario are combined, and a random algorithm is used to generate a first planned route for the user to reach a single fixed target location in the hospital where the appointment registration department is located, which satisfies the first constraint condition; the first constraint condition includes: for the elderly type or the child type, the density of people in the area where the planned route passes is less than the first preset density threshold and the distance to the single fixed target location in the hospital where the appointment registration department is located is the shortest; for the adult type, the density of people in the area where the planned route passes is less than the second preset density threshold and the distance to the single fixed target location in the hospital where the appointment registration department is located is the shortest; the first preset density threshold is less than the second preset density threshold; When receiving the second route planning request instruction sent by the user terminal, according to the final recommended inspection items, the location information of each inspection item in each area of ​​the hospital, the current density of people in each area of ​​the hospital and the current waiting time for each inspection item are obtained, and combined with the determined user object type and the user guidance application scenario, a random algorithm is used to generate a second planned route for users to complete all inspection items that meets the second constraint condition; the second constraint condition includes: for the elderly type or the child type, the density of people in the area where the planned route passes is less than the first preset density threshold, and the waiting time required to complete all inspection items is the shortest, and the distance to complete all inspection items is the shortest; for the adult type, the density of people in the area where the planned route passes is less than the second preset density threshold, and the waiting time required to complete all inspection items is the shortest, and the distance to complete all inspection items is the shortest.

[0013] By adopting the above solution, the first planned route and the second planned route are generated according to the user's object type and guidance application scenario, combined with the real-time crowd density of the hospital, so that the first planned route ensures that the user can avoid crowded areas, ensuring the user's convenience and safety; the second planned route ensures that all the user's examination items are completed in the shortest time, and at the same time has the shortest route distance and avoids high-crowded areas, thereby improving the user's medical efficiency and user experience.

[0014] Preferably, it also includes: When the preliminary diagnosis report is sent to the terminal of the doctor in the appointment department, the preliminary diagnosis result in the preliminary diagnosis report is in a pending unlocking state within a preset time period. When the preliminary diagnosis result acquisition request sent by the terminal of the doctor in the appointment department is received or after the preset time period, the preliminary diagnosis result unlocking password in the preliminary diagnosis report is sent to the terminal of the doctor in the appointment department.

[0015] By adopting the above solution, the preliminary diagnosis results are set to be in a pending unlocking state within a preset time period, and are unlocked after receiving a request or exceeding the preset time period reserved for the doctor's self-reflection. This ensures that the preliminary diagnosis results will not be immediately exposed to unauthorized personnel, thereby enhancing information security and patient privacy protection, and reducing the doctor's dependence on the intelligently generated preliminary diagnosis results, thereby preventing the doctor from misdiagnosing.

[0016] Preferably, it also includes: Receive the medical advice information uploaded by the doctor's terminal in the appointment department, and generate the user's personalized health management information based on the medical advice information and user portrait information using a deep learning algorithm; the medical advice information includes: dietary taboo information, medication information and rehabilitation training information; the user's personalized health management information includes: user's personalized recipe recommendations, user medication reminder information, and rehabilitation training reminder information.

[0017] By adopting the above solution and taking the user's habits into consideration, personalized health management based on doctor's advice information and user portrait information can be achieved, thereby improving the scientificity and personalization of the user's diet, medication and rehabilitation training after the doctor's visit, and improving user experience satisfaction.

[0018] Preferably, it also includes: Receive the drug extraction completion information sent by the hospital information system of the appointment department, generate drug inspection information and display it on the user terminal; Receive the drug image and drug inspection request instruction uploaded by the user terminal, use image recognition technology to identify the drug information, compare the identified drug information with the drug information in the drug administration information to see if they are the same, if there is a difference, generate a drug error prompt message and send it to the hospital information system or user terminal where the appointment department is located; if the identified drug information includes prescription drug information, in addition to comparing the prescription drug information in the identified drug information with the prescription drug information in the drug administration information to see if they are the same, it is also necessary to find out whether there is a doctor's signature information in the received medical order information, if there is no doctor's signature information, then also generate a drug error prompt message and send it to the hospital information system where the appointment department is located and the doctor's terminal in the appointment department.

[0019] By adopting the above solution, users are reminded to check the medicines they extract based on the doctor's order information to ensure that the medicines they receive are accurate. At the same time, the prescription medicines are confirmed to be those that the doctor has determined need to be taken, thereby avoiding safety hazards and improving user experience.

[0020] In the second aspect, the present application provides an intelligent medical consultation system with AI virtual human accompanying, comprising: The virtual clinic building module is used to receive the intelligent clinic demand instructions sent by the user terminal, build a virtual clinic and generate an AI virtual person; The medical guidance information acquisition module is used to obtain the user input information in real time after judging that the user enters the virtual clinic based on the user terminal operation; the user input information includes the pre-entered user portrait information, the user selected intelligent medical guidance scene information, the user real-time interaction information and the user's historical intelligent medical guidance information; The diagnosis guidance influencing factor judgment module is used to determine the user object type based on the acquired user portrait information and / or the user's historical intelligent diagnosis guidance information; the user object type includes: child type, adult type, elderly type and interaction disorder group type; determine the user diagnosis guidance application scenario based on the acquired user's historical intelligent diagnosis guidance information and the user's selected intelligent diagnosis guidance scenario information; the user diagnosis guidance application scenario includes: general consultation, emergency consultation, general follow-up consultation and emergency follow-up consultation; The guidance and interactive emotion soothing module is used to use a deep learning algorithm to complete the recognition of abnormal user emotions based on the user's real-time interaction information, and the abnormal conditions include anxiety, tension, and sadness; based on the recognition of abnormal emotions and the determination of user object types, a preset emotion soothing strategy is matched, so that the AI ​​virtual person interacts with the user according to the matched emotion soothing strategy; wherein each user object type is matched with each abnormal emotion by a preset emotion soothing strategy; The guidance diagnosis and recommendation module is used to match the preset deep learning model according to the determined user guidance application scenario; according to the user's real-time interaction information and the user's historical intelligent guidance information, combined with the matched deep learning model, the AI ​​virtual human interaction reaction information and the preliminary diagnosis results are output; among which, the number of AI virtual human interaction reaction information output by different preset deep learning models and the output time of the preliminary diagnosis results are different; The guidance registration appointment operation module is used to determine the registration department based on the preliminary diagnosis results, and generate registration appointment recommendations based on the registration department; receive the registration appointment request information sent by the user terminal based on the registration appointment recommendation, connect to the hospital information system of the selected registration department and complete the registration appointment.

[0021] By adopting the above solution, the construction and application of the intelligent medical treatment system is realized, which can provide personalized medical services according to the specific needs of different users and improve the user experience in the intelligent medical treatment process accompanied by AI virtual people.

[0022] In a third aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method as described above.

[0023] In a fourth aspect, the present application provides a computer device, comprising a memory, a processor, and a program stored and executable on the memory, wherein the program implements the steps of the above method when executed by the processor.

[0024] In summary, this application has the following beneficial effects: 1. Obtain a variety of user information, including user portrait information, intelligent medical guidance scenario information, real-time interaction information and historical intelligent medical guidance information to determine the user object type and its application scenario, and then match the corresponding emotion soothing strategy to the abnormal emotion of the identified user according to the user object type, match the corresponding deep learning model according to the application scenario, and output the interactive response information and preliminary diagnosis results of the AI ​​virtual person in a timely manner, and recommend the appropriate registration department and appointment registration service based on the preliminary diagnosis results, thereby improving the user's experience during the medical treatment process; 2. When users visit the recommended registration department for medical treatment, the system will recommend planned routes based on the needs of different users to meet the needs of different users for planned routes in different application scenarios, generate planned routes with low density and short time consumption, and improve user experience. When users visit the recommended registration department for medical treatment, the system will compare the examination items corresponding to the preliminary diagnosis results with the examination items recommended by the real doctor, and the preliminary diagnosis results with the diagnosis results given by the real doctor, remind the doctor of possible errors, and ensure the accuracy of the diagnosis results. 3. When the consultation is completed and the medication is collected, the user is instructed to check the medication to ensure the accuracy of the medication extraction. After the consultation is completed, personalized health management information is generated based on the doctor's order information and user portrait, so as to achieve scientific and personalized management of the user's diet, medication and rehabilitation training after the consultation, and improve user experience satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a flowchart of the intelligent medical consultation method with AI virtual person accompanying the patient in a specific embodiment; Figure 2 It is a structural diagram of the intelligent medical consultation system with AI virtual human accompanying the patient as described in the specific embodiment. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0027] like Figure 1 As shown, the embodiment of the present application discloses an intelligent medical consultation method with AI virtual person accompanying the patient, and the specific steps include: S1. Receive intelligent medical guidance instructions, build a virtual clinic and generate an AI virtual person.

[0028] Specifically, users can log in to the AI ​​virtual person accompanying diagnosis APP and pre-enter the user's portrait information, such as: basic personal information such as age, gender, ID card, habits, historical illness, sensitive drugs and other disease diagnosis information, as well as blood pressure, weight and other physical health information; operate in the AI ​​virtual person accompanying diagnosis APP to generate intelligent guidance demand instructions, and the corresponding server of the AI ​​virtual person accompanying diagnosis APP receives the intelligent guidance demand instructions sent by the user terminal. After receiving the intelligent guidance demand instructions sent by the user terminal, it builds a virtual clinic and generates an AI virtual person to be displayed on the AI ​​virtual person accompanying diagnosis APP interface. Among them, the AI ​​virtual person generated by the corresponding server of the AI ​​virtual person accompanying diagnosis APP can generate an AI virtual person image that meets the user's preferences based on the user portrait information.

[0029] S2. After determining that the user enters the virtual clinic based on the user terminal operation, obtain the user input information in real time.

[0030] Specifically, the server corresponding to the AI ​​virtual human accompanying medical consultation APP will determine whether the user enters the virtual clinic through the user terminal operation on the APP interface, and after determining that the user enters the virtual clinic based on the user terminal operation, obtain the user input information in real time.

[0031] Among them, the AI ​​virtual human accompanying medical consultation APP displays the intelligent medical guidance scene mode information. Users can select the intelligent medical guidance scene on the AI ​​virtual human accompanying medical consultation APP interface according to their own needs. After judging that the user enters the virtual clinic based on the user terminal operation, the user input information such as the pre-entered user portrait information and the user-selected intelligent medical guidance scene information can be obtained in real time; the corresponding server of the AI ​​virtual human accompanying medical consultation APP will store the user's historical intelligent medical guidance information of the corresponding user. After judging that the user enters the virtual clinic based on the user terminal operation, the user's historical intelligent medical guidance information can be queried and called in real time; after the user enters the virtual clinic through the user terminal operation, the real-time interactive information will be uploaded, including: input text information, video information, voice information and other types of interactive information to obtain real-time user interactive information.

[0032] S3. Determine the user object type and user guidance application scenario based on the user input information.

[0033] Specifically, considering the differences in the needs of different user object types for interacting with AI virtual humans or for AI virtual humans to give preliminary diagnoses in different guidance application scenarios, the current user object type and user guidance application scenario are prioritized.

[0034] Among them, considering that users of different age groups have different needs for interacting with AI virtual people, the user object types are mainly divided into the following types, including: children type, adult type, elderly type and people with interaction disorders (such as people with voice disorders or hearing disorders, etc.); the user object type currently logging into the APP to perform intelligent diagnosis needs operations is determined based on the acquired user portrait information and / or the age information and physical health information in the user's historical intelligent diagnosis information; Taking into account the significant differences in the length of time it takes for the AI ​​virtual human to give a preliminary diagnosis or the simplicity of interaction with the AI ​​virtual human in the scenarios of general consultation and emergency consultation, and normal consultation and follow-up consultation, the user guidance application scenarios are mainly divided into the following types, including: general consultation, emergency consultation, general follow-up and emergency follow-up; the user guidance application scenario is determined based on the user's historical intelligent guidance information obtained and the intelligent guidance scene information selected by the user, whether it is the first consultation, and the selected intelligent guidance scene mode (general or emergency).

[0035] S4. Determine the user object type and the user guidance application scenario based on the user input information, and match the preset emotion soothing strategy and deep learning model, so that the AI ​​virtual person interacts with the user according to the matched emotion soothing strategy and deep learning model, and generates a preliminary diagnosis result.

[0036] Specifically, after the user enters interactive information on the APP, the AI ​​virtual person will interact and react according to the interactive information entered, thereby completing the preliminary diagnosis of whether the user is ill or not. During the diagnosis process, the user may have some abnormal emotions such as anxiety, tension, sadness, etc. These abnormal emotions will affect the user's description of the feeling of illness. In order to output accurate diagnosis results as much as possible, identify the user's abnormal emotions and adopt emotional comfort strategies to alleviate the user's abnormal emotions, improve the emotional intelligence and personalization level of the AI ​​virtual person's accompanying service, and thus achieve a more intimate and efficient accompanying experience in different scenarios and populations.

[0037] Among them, the corresponding server of the AI ​​virtual human accompanying medical treatment APP has a built-in emotion recognition model built and trained based on a deep learning algorithm. The input of the emotion recognition model is the user's real-time interaction information, including: user text information, user facial and body video information or user semantic information and other multimodal real-time interaction information. The output of the model is whether the user has abnormal emotions and the type of abnormal emotions. The abnormal conditions include anxiety, tension, and sadness. The emotion recognition model is used to realize the recognition of user abnormal emotions based on user real-time interaction information.

[0038] Based on the identification of abnormal emotions and the determination of user object types, a preset emotion soothing strategy is matched, so that the AI ​​virtual person interacts with the user according to the matched emotion soothing strategy; wherein each user object type matches a preset emotion soothing strategy with each abnormal emotion, and these preset emotion soothing strategies can be based on the emotion soothing strategies set by experts for different user object types and different types of abnormal emotions. For example: for the child type, the first emotion soothing strategy to the third emotion soothing strategy are preset in the order of anxiety, tension, and sadness, such as: for the nervous child type, the user presets the second emotion soothing strategy, and the second emotion soothing strategy includes: playing animations, music and other elements that children like, distracting their attention, and linking their nervous emotions. The AI ​​virtual person plays animations, music and other elements that children like according to the matched second emotion soothing strategy to interact with the user; another example: for the elderly type, the fourth emotion soothing strategy to the sixth emotion soothing strategy are preset in the order of anxiety, tension, and sadness, such as: for the elderly type with sad emotions, the sixth emotion soothing strategy is preset, and the sixth emotion soothing strategy includes: sharing the illness with them The AI ​​virtual person generates successful treatment cases of the user's disease or outputs encouraging words to relieve the user's sadness, and interacts with the user according to the matched sixth emotion soothing strategy; for example, for adults, the seventh to ninth emotion soothing strategies are preset in the order of anxiety, tension, and sadness, such as: for adults with anxious emotions, the seventh emotion soothing strategy is preset, and the seventh emotion soothing strategy includes: using calm language to guide the user to explain the symptoms, and the AI ​​virtual person interacts with the user according to the matched seventh emotion soothing strategy; for example, for people with interaction disorders, the tenth to twelfth emotion soothing strategies are preset in the order of anxiety, tension, and sadness.

[0039] Taking into account that users have different requirements for the output time of preliminary diagnosis results and the simplicity of interaction in different application scenarios, the preset deep learning models are matched according to the determined user guidance application scenarios; among which, the number of AI virtual human interaction response information and the output time of preliminary diagnosis results output by different preset deep learning models are different, including: the first deep learning model matching ordinary consultations, the second deep learning model matching emergency consultations, the third deep learning model matching ordinary follow-up visits, and the fourth deep learning model matching emergency follow-up visits; the input of each deep learning model is the user's real-time interaction information and the user's historical intelligent guidance information, and the model output is AI virtual human interaction response information or preliminary diagnosis results. The number of AI virtual human interaction response information (i.e., the rounds of the AI ​​interaction process, reflecting the simplicity of AI interaction) and the output time of preliminary diagnosis results (reflecting diagnostic efficiency) of each deep learning model gradually decreases with the first deep learning model, the second deep learning model, the third deep learning model, and the fourth deep learning model.

[0040] Based on the user's real-time interaction information and historical intelligent guidance information, the matched deep learning model is combined to output AI virtual human interaction response information and preliminary diagnosis results.

[0041] S5. Determine the registration department according to the preliminary diagnosis results, and generate a registration appointment recommendation according to the registration department; receive the registration appointment request information sent by the user terminal according to the registration appointment recommendation, connect to the hospital information system of the selected registration department and complete the registration appointment.

[0042] Specifically, the registration department corresponding to the preliminary diagnosis result is determined based on the registration department information corresponding to the diagnosis results of different diseases in the medical knowledge graph. For example, for the hand injury input by the user, the registration department is determined to be general surgery. Subsequently, the registration appointment recommendation is generated based on the registration department considering factors such as location and registration appointment volume. The specific steps include: The geographical location information of the corresponding user terminal is obtained based on the last interaction information between the user and the AI ​​virtual person received before the output of the preliminary diagnosis result, that is, the geographical location is determined based on the transmission path and time of information sending; According to the preliminary diagnosis results, the current number of appointments in the hospital with the determined registration department is obtained, that is, the hospital information system of the hospital with the determined registration department is connected to query the current number of appointments in the department, such as: the current number of appointments in the general surgery department of the First People's Hospital is 10, and the current number of appointments in the general surgery department of the Second People's Hospital is 4.

[0043] Based on the determined user object type and user guidance application scenario, the acquired geographical location information of the user terminal and the current number of appointments in the hospital with a determined registration department, a recommended order of registration appointments that meets the first rule is generated based on the user's historical intelligent guidance information and the determined registration department and displayed on the APP interface for user selection.

[0044] Among them, the first rule includes: satisfying that the hospital where the registration department corresponding to the ordinary follow-up visit and the emergency follow-up visit in the user's guidance application scenario is located is the same as the hospital where the original registration department is located, that is, once it is determined that the current user guidance application scenario is a follow-up visit, the hospital and department corresponding to the registration appointment are determined according to the hospital where the original registration department is located in the user's historical intelligent guidance information; And it satisfies that the current number of appointments for the hospital with a determined registration department is less than the maximum appointment threshold for the corresponding hospital matched by the user guidance application scenario, that is, different user guidance application scenarios are set with corresponding maximum appointment thresholds for the corresponding hospitals, and the threshold is determined according to the data with the maximum average value of the number of appointments accepted historically under different user guidance application scenarios of the hospital with the determined registration department. If the maximum appointment threshold is exceeded, there is a risk of registration failure; specifically, it is recorded as: the first maximum appointment threshold for ordinary consultation matching, the second maximum appointment threshold for emergency consultation, the third maximum appointment threshold for ordinary follow-up matching, and the fourth maximum appointment threshold for emergency follow-up matching. Since the consultation time for follow-up is shorter than that for ordinary consultations, and the number of appointment numbers for emergency consultations is more restricted than that for ordinary consultations, the second maximum appointment threshold, the first maximum appointment threshold, the fourth maximum appointment threshold, and the third maximum appointment threshold for a specific hospital generally increase gradually; The hospital with the confirmed registration department with the less number of current appointments is recommended with priority, such as: the General Surgery Department of the Second People's Hospital is preferred over the General Surgery Department of the First People's Hospital; and the hospital where the registration department is located has the closest distance to the geographical location information of the user terminal, so it is recommended with priority, such as: the geographical location of the user terminal is closer to the General Surgery Department of the First People's Hospital, so the General Surgery Department of the First People's Hospital is recommended with priority over the General Surgery Department of the Second People's Hospital; And if the hospital where the registration department is located is a dedicated hospital for the user type, priority will be given to recommendation. For example, if there is a dedicated hospital for children, then the dedicated hospital for children will be recommended for children. Among them, since there may be contradictions in the priority recommendations based on different factors, the priority of the priority recommendation generated based on the geographical location information of the user terminal, the priority of the priority recommendation order generated based on the user object type, and the priority of the priority recommendation generated based on the number of appointments are determined according to the priority level pre-set by the user (that is, the login information pre-entered by the user). The larger the priority level, the higher the order of the corresponding priority recommendation, that is, the corresponding priority recommendation with a larger priority level is used as the standard. For example, the user pre-sets the priority of the priority recommendation generated based on the number of appointments to be the highest, that is, the user attaches the most importance to the number of appointments, then the General Surgery Department of the Second People's Hospital is higher in the priority recommendation order than the General Surgery Department of the First People's Hospital. In addition, the priority level of the priority recommendation generated based on different factors can be determined according to the user's historical selection. The factor with the most selection times has a higher priority level for generating the priority recommendation.

[0045] The registration appointment request information sent by the user terminal according to the registration appointment recommendation is received, the hospital information system of the selected registration department is connected and the registration appointment is completed, so that the user can directly see the doctor.

[0046] In summary, by adopting the method described in the above embodiment, it is possible to improve the user experience in the process of intelligent medical consultation accompanied by AI virtual people, and meet the needs of different types of users in different application scenarios.

[0047] In a specific embodiment, in the process of intelligent medical consultation in which the user is accompanied by an AI virtual person, in order to further improve the efficiency and accuracy of medical consultation and enhance the user's experience of medical consultation service, the method further includes: The generated preliminary diagnosis report is sent synchronously when the hospital information system of the appointment department is connected; when the preliminary diagnosis report request instruction is received from the doctor's terminal of the appointment department in the hospital information system, the preliminary diagnosis report is sent to the doctor's terminal of the appointment department, which is equivalent to informing the doctor of the patient's basic condition in advance, so that the doctor can have a general understanding of the situation in advance and realize timely diagnosis and output of recommended examination items.

[0048] Among them, the preliminary diagnosis report is to use natural language technology or deep learning technology to convert the user's real-time interaction information and AI virtual human interaction response information into text content to extract key information, such as key information related to the description of the disease; and integrate the key information with the preliminary diagnosis results to generate the report.

[0049] Receive the real doctor's recommended examination items uploaded by the doctor's terminal of the appointment registration department, determine the examination items required for the confirmation of the preliminary diagnosis result according to the medical knowledge graph, compare the determined examination items with the examination items recommended by the real doctor for similarity, if the similarity is lower than the first preset similarity, it indicates that the examination items recommended by the real doctor may have omitted items, and accordingly generate a prompt message for reconfirmation of the real doctor's recommended examination items and display the examination items required for the confirmation of the preliminary diagnosis result, so that the real doctor can judge whether additional adjustment and modification are needed based on the examination items required for the confirmation of the preliminary diagnosis result; when receiving the reconfirmation of the examination items recommended by the real doctor, it indicates that the real doctor has determined that the examination items are correct, and the examination items recommended by the real doctor are used as the final recommended examination items, otherwise, the modified examination items recommended by the real doctor are used as the final recommended examination items; Taking into account that combining the examination results with the user interaction information can analyze and obtain more accurate diagnosis results, after receiving the results of various examination items of the user uploaded by the hospital information system, the secondary diagnosis result is output according to the results of each examination item of the user, the real-time interaction information of the user and the historical intelligent guidance information of the user, combined with the deep learning algorithm; the real doctor's diagnosis result uploaded by the terminal of the doctor sitting in the appointment department is received, and the similarity is compared with the generated secondary diagnosis result. If the similarity is lower than the second preset similarity, it indicates that the real doctor's diagnosis may be wrong, and the real doctor's diagnosis result reconfirmation information is generated and the preliminary diagnosis result is displayed on the terminal of the doctor sitting in the registration department, so that the real doctor can reconsider his own diagnosis result and judge whether the diagnosis result needs to be modified; when the real doctor's diagnosis result confirmation information is received, the real doctor's diagnosis information is used as the final diagnosis result, and the modified real doctor's diagnosis information is used as the final diagnosis result.

[0050] In a specific embodiment, in order to further consider the different needs of different users and application scenarios for route planning during medical consultation and provide a planned route that meets the needs of users as much as possible, the method further includes: After the user completes the registration appointment, he will go to the hospital where the registration appointment is located. After arriving at the hospital, in order to facilitate the user to quickly reach the registration department, a planned route generation function is set in the APP to facilitate the user to move according to the planned route; among them, the planned route generation includes two levels of route planning, one is the planned route planning to reach a single fixed target location, such as: reaching the registration department or picking up the pharmacy, corresponding to the input of a single fixed target location operation to generate a first route planning request instruction; the other is a planned route that needs to continuously reach the locations of multiple inspection items according to the inspection items, and the corresponding operation of obtaining the final recommended inspection items generates a second route planning request instruction.

[0051] In this embodiment, taking arrival at the registration department as an example, considering the dense flow of people and the distance that the user needs to move, when receiving the first route planning request instruction sent by the user terminal, a single fixed target location in the hospital where the reserved registration department is located (such as the location information of each department) and the current density of people in each area of ​​the hospital are obtained, and combined with the determined user object type and the user guidance application scenario, a random algorithm is used to generate a first planned route for users who meet the first constraint condition to reach a single fixed target location in the hospital where the reserved registration department is located (such as the reserved registration department); due to different types of users having different abilities to move safely in areas with different degrees of crowd density, adults are more likely to move safely in areas with different crowd densities than the elderly, children and people with interaction disabilities. It is able to achieve safe movement in areas with higher crowd density. The first constraint condition includes: for the elderly type, the child type, or the type of people with interaction disabilities, the crowd density in the area where the planned route is passed is less than the first preset crowd density threshold and the distance to a single fixed target location in the hospital where the appointment department is located (such as the appointment department) is the shortest; for the adult type, the crowd density in the area where the planned route is passed is less than the second preset crowd density threshold and the distance to a single fixed target location in the hospital where the appointment department is located (such as the appointment department) is the shortest; the first preset crowd density threshold is less than the second preset crowd density threshold, and the specific value can be determined according to expert settings.

[0052] When receiving the second route planning request instruction sent by the user terminal, according to the final recommended inspection items, the location information of each inspection item in each area of ​​the hospital, the current density of people in each area of ​​the hospital and the current waiting time for each inspection item are obtained, and combined with the determined user object type and the user guidance application scenario, a random algorithm is used to generate a second planned route for users to complete all inspection items that meets the second constraint condition; the second constraint condition includes: for the elderly type or the child type, the density of people in the area where the planned route passes is less than the first preset density threshold, and the waiting time required to complete all inspection items is the shortest, and the distance to complete all inspection items is the shortest; for the adult type, the density of people in the area where the planned route passes is less than the second preset density threshold, and the waiting time required to complete all inspection items is the shortest, and the distance to complete all inspection items is the shortest.

[0053] In a specific embodiment, in order to protect the user's personal security information and reduce the doctor's reliance on AI intelligence, the method further includes: When the preliminary diagnosis report is sent to the doctor's terminal of the appointment department, the preliminary diagnosis result in the preliminary diagnosis report is in a waiting unlocking state within a preset time period, that is, the preliminary diagnosis result in the preliminary diagnosis report is encrypted and hidden, and the unlocking password needs to be entered to be displayed; the preset time period can be set manually, and can be adaptively set according to the practice time of different doctors, such as: if the practice time of the doctor is set to exceed the first preset practice time (10 years), the first preset time period (2 minutes) is set accordingly, otherwise the second preset time period (5 minutes) is set accordingly; Taking into account the existence of difficult problems, the attending doctor needs to rely on the preliminary diagnosis results output by the AI ​​virtual person. When the preliminary diagnosis result acquisition request sent by the attending doctor's terminal of the appointment department is received or after the preset time period, the preliminary diagnosis result unlocking password in the preliminary diagnosis report will be sent to the attending doctor's terminal of the appointment department.

[0054] In a specific embodiment, in the intelligent medical consultation process accompanied by an AI virtual person, in addition to providing some auxiliary functions before and during the consultation to improve the user experience, the AI ​​virtual person can also provide personalized management information to assist the user in recovery during the recovery process after the consultation and the period between the consultation and the follow-up visit. The method also includes: Receive the medical advice information uploaded by the doctor's terminal in the appointment department, and generate the user's personalized health management information based on the medical advice information and user portrait information using a deep learning algorithm; the medical advice information includes: dietary taboo information, medication information, rehabilitation training information, etc.; the user's personalized health management information includes: user's personalized recipe recommendations, user medication reminder information, rehabilitation training reminder information, etc.

[0055] In a specific embodiment, considering that there may be information that affects user safety, such as drug errors, during the process of picking up medicine after the user has finished the medical consultation, in order to further ensure the safety of the user's medical consultation, the method further includes: The staff uploads the drug extraction completion information to the drug collection terminal in the hospital information system of the appointment registration department, and the APP receives the drug extraction completion information sent by the hospital information system of the appointment registration department, generates drug inspection information and displays it on the user terminal; Receive the drug image and drug inspection request instruction uploaded by the user terminal, use image recognition technology to identify drug information, compare the identified drug information with the drug information in the drug administration information to see if they are the same, and if there is a difference, generate a drug error prompt message and send it to the drug collection terminal or user terminal in the hospital information system where the appointment registration department is located, so that the user can remind the staff who picks up the medicine or directly remind the staff who picks up the medicine; If the identified drug information includes prescription drug information, in addition to comparing whether the prescription drug information in the identified drug information is the same as the prescription drug information in the drug administration information, it is also necessary to find out whether there is a doctor's signature information in the received medical order information. If there is no doctor's signature information, a drug error prompt message is also generated and sent to the drug collection terminal in the hospital information system where the appointment registration department is located and the doctor's terminal in the appointment registration department to remind the drug collection staff to confirm whether the drug collection voucher is correct or to remind the doctor whether there is an error.

[0056] like Figure 2 As shown, the embodiment of the present application discloses an intelligent medical consultation system with AI virtual person accompanying the patient, including: A virtual consultation room construction module 101 is used to receive an intelligent consultation demand instruction sent by a user terminal, construct a virtual consultation room and generate an AI virtual person; The medical consultation information acquisition module 102 is used to obtain user input information in real time after determining that the user enters the virtual consultation room based on the user terminal operation; the user input information includes pre-entered user portrait information, user selected intelligent medical consultation scene information, user real-time interaction information and user historical intelligent medical consultation information; The diagnosis guidance influencing factor judgment module 103 is used to determine the user object type according to the acquired user portrait information and / or the user's historical intelligent diagnosis guidance information; the user object type includes: child type, adult type, elderly type and interaction barrier group type; determine the user diagnosis guidance application scenario according to the acquired user's historical intelligent diagnosis guidance information and the intelligent diagnosis guidance scenario information selected by the user; the user diagnosis guidance application scenario includes: general consultation, emergency consultation, general follow-up consultation and emergency follow-up consultation; The guidance and interactive emotion soothing module 104 is used to use a deep learning algorithm to complete the recognition of abnormal emotions of users according to real-time interaction information of users, and the abnormal conditions include anxiety, tension, and sadness; based on the recognition of abnormal emotions and the determination of user object types, a preset emotion soothing strategy is matched, so that the AI ​​virtual person interacts with the user according to the matched emotion soothing strategy; wherein each user object type is matched with each abnormal emotion by a preset emotion soothing strategy; The guidance diagnosis and recommendation module 105 is used to match the preset deep learning model according to the determined user guidance application scenario; according to the user's real-time interaction information and the user's historical intelligent guidance information, combined with the matched deep learning model, output the AI ​​virtual human interaction reaction information and the preliminary diagnosis result; wherein, the number of AI virtual human interaction reaction information output by different preset deep learning models and the output time of the preliminary diagnosis result are different; The guidance registration appointment operation module 106 is used to determine the registration department according to the preliminary diagnosis results, and generate a registration appointment recommendation based on the registration department; receive the registration appointment request information sent by the user terminal according to the registration appointment recommendation, connect to the hospital information system of the selected registration department and complete the registration appointment.

[0057] The system further comprises: The medical guidance route planning module 107 is used to, upon receiving the first route planning request instruction sent by the user terminal, obtain a single fixed target location in the hospital where the appointment registration department is located and the density of people in each area of ​​the current hospital, and combine the determined user object type and the user medical guidance application scenario to use a random algorithm to generate a first planned route for the user who meets the first constraint condition to reach the single fixed target location in the hospital where the appointment registration department is located; the first constraint condition includes: for the elderly type or the child type, the density of people in the area where the planned route passes is less than a first preset density threshold and the distance to the single fixed target location in the hospital where the appointment registration department is located is the shortest; for the adult type, the density of people in the area where the planned route passes is less than a second preset density threshold and the distance to the single fixed target location in the hospital where the appointment registration department is located is the shortest; the first preset density threshold is less than the first preset density threshold 2. Preset crowd density threshold; when receiving the second route planning request instruction sent by the user terminal, according to the final recommended inspection items, obtain the location information of each inspection item in each area of ​​the hospital, the current crowd density in each area of ​​the hospital and the current waiting time for each inspection item, combine the determined user object type and the user guidance application scenario, and use a random algorithm to generate a second planned route for users to complete all inspection items that meets the second constraint condition; the second constraint condition includes: for the elderly type or the child type, the crowd density in the area where the planned route passes is less than the first preset crowd density threshold, and the waiting time required to complete all inspection items is the shortest, and the distance to complete all inspection items is the shortest; for the adult type, the crowd density in the area where the planned route passes is less than the second preset crowd density threshold, and the waiting time required to complete all inspection items is the shortest, and the distance to complete all inspection items is the shortest.

[0058] The real diagnosis guidance result acquisition module 108 is used to synchronously send the generated preliminary diagnosis report when connecting to the hospital information system of the selected appointment department; when receiving the preliminary diagnosis report request instruction sent by the doctor terminal of the appointment department in the hospital information system where the appointment department is located, the preliminary diagnosis report is sent to the doctor terminal of the appointment department; the preliminary diagnosis report is generated by integrating the key information extracted from the user real-time interaction information converted into text content and the AI ​​virtual person interaction response information by natural language technology, combined with the preliminary diagnosis result; receive the real doctor's recommended examination items uploaded by the doctor terminal of the appointment department, determine the examination items required for the confirmation of the preliminary diagnosis result according to the medical knowledge graph, compare the determined examination items with the real doctor's recommended examination items for similarity, if the similarity is lower than the first preset similarity, generate a prompt message for reconfirmation of the real doctor's recommended examination items and display the examination items required for the confirmation of the preliminary diagnosis result, when receiving the reconfirmation of the real doctor's recommended examination items, The recommended examination items are used as the final recommended examination items; upon receiving the results of various examination items uploaded by the user from the hospital information system, the secondary diagnosis results are output according to the results of various examination items of the user, the real-time interaction information of the user and the historical intelligent guidance information of the user, combined with the deep learning algorithm; the real doctor's diagnosis results uploaded by the terminal of the doctor in the appointment department are received, and the similarity is compared with the generated secondary diagnosis results. If the similarity is lower than the second preset similarity, the real doctor's diagnosis result reconfirmation information is generated and the preliminary diagnosis result is displayed on the terminal of the doctor in the appointment department. When the real doctor's diagnosis result confirmation information is received, the real doctor's diagnosis information is used as the final diagnosis result; it is also used for sending the preliminary diagnosis report to the terminal of the doctor in the appointment department, and the preliminary diagnosis result in the preliminary diagnosis report is in a pending unlocking state within a preset time period. When the preliminary diagnosis result acquisition request sent by the terminal of the doctor in the appointment department is received or after the preset time period has expired, the preliminary diagnosis result unlocking password in the preliminary diagnosis report is sent to the terminal of the doctor in the appointment department.

[0059] The guidance drug inspection module 109 is used to receive the drug extraction completion information sent by the hospital information system of the appointment registration department, generate drug inspection information and display it on the user terminal; receive the drug image and drug inspection request instruction uploaded by the user terminal, use image recognition technology to identify the drug information, compare the identified drug information with the drug information in the drug administration information to see if they are the same, if there is a difference, generate drug error prompt information and send it to the hospital information system or user terminal of the appointment registration department; if the identified drug information includes prescription drug information, in addition to comparing the prescription drug information in the identified drug information with the prescription drug information in the drug administration information to see if they are the same, it is also necessary to find out whether there is a doctor's signature information in the received medical order information. If there is no doctor's signature information, then a drug error prompt information is also generated and sent to the hospital information system of the appointment registration department and the doctor's terminal of the appointment registration department.

[0060] The guidance health management module 110 is used to receive the medical advice information uploaded by the doctor's terminal in the appointment department, and generate user personalized health management information based on the medical advice information and user portrait information using a deep learning algorithm; the medical advice information includes: dietary taboo information, medication information and rehabilitation training information; the user personalized health management information includes: user personalized recipe recommendations, user medication reminder information, and rehabilitation training reminder information.

[0061] The embodiment of the present application also discloses a computer-readable storage medium.

[0062] Specifically, the computer-readable storage medium stores a computer program that can be loaded by a processor and execute an intelligent medical treatment method such as the above-mentioned AI virtual human accompanying doctor. The computer-readable storage medium includes, for example: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0063] The embodiment of the present application also discloses a computer device.

[0064] Specifically, the computer device includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and execute the above-mentioned AI virtual human accompanying medical intelligent medical method.

[0065] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in this specification (including the abstract and drawings), unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

Claims

1. An intelligent medical consultation method using AI virtual human as an accompaniment, characterized in that: include: Receive intelligent guidance demand instructions sent by user terminals, build a virtual clinic and generate AI virtual people; After determining that the user enters the virtual clinic based on the user terminal operation, the user input information is obtained in real time; the user input information includes pre-entered user portrait information, user selected intelligent diagnosis scene information, user real-time interaction information and user historical intelligent diagnosis information; Determine the user object type based on the acquired user portrait information and / or the user's historical intelligent diagnosis information; The user object types include: children type, adult type, elderly type and interaction barrier group type; determine the user guidance application scenario based on the acquired user historical intelligent guidance information and the intelligent guidance scenario information selected by the user; the user guidance application scenarios include: general consultation, emergency consultation, general follow-up and emergency follow-up; use deep learning algorithms to complete the user's abnormal emotion recognition based on the user's real-time interaction information, and the abnormal conditions include anxiety, tension, and sadness; match the preset emotion soothing strategy based on the recognition of abnormal emotions and the determined user object type, so that the AI ​​virtual person interacts with the user according to the matched emotion soothing strategy; wherein each user object type is matched with each Abnormal emotions are matched with a preset emotion soothing strategy; a preset deep learning model is matched according to a determined user guidance application scenario; based on the user's real-time interaction information and the user's historical intelligent guidance information, the matched deep learning model is combined to output the AI ​​virtual human interaction response information and the preliminary diagnosis results; among which, the number of AI virtual human interaction response information output by different preset deep learning models is different from the output time of the preliminary diagnosis results; the registration department is determined according to the preliminary diagnosis results, and a registration appointment recommendation is generated according to the registration department; the registration appointment request information sent by the user terminal according to the registration appointment recommendation is received, the hospital information system of the selected registration department is connected, and the registration appointment is completed.

2. The intelligent medical treatment method with AI virtual person accompanying the patient according to claim 1 is characterized in that: Determining the registration department according to the preliminary diagnosis result and generating a registration appointment recommendation according to the registration department includes: Obtain the geographical location information of the corresponding user terminal based on the last interaction information between the user and the AI ​​virtual person received before the output of the preliminary diagnosis result; obtain the current number of appointments in the hospital with the determined registration department based on the preliminary diagnosis result; In combination with the determined user object type and the user guidance application scenario, the acquired geographical location information of the user terminal and the current number of appointments in the hospital with the determined registration department, a registration appointment recommendation order that meets the first rule is generated according to the user's historical intelligent guidance information and the determined registration department; the first rule includes: the hospital where the registration department corresponding to the general follow-up and emergency follow-up in the user guidance application scenario is located is the same as the hospital where the original registration department is located, and the current number of appointments in the hospital with the determined registration department is less than the maximum number of appointments in the corresponding hospital matching the user guidance application scenario, and the current number of appointments in the hospital with the determined registration department is less than the maximum number of appointments in the corresponding hospital matching the user guidance application scenario. The hospital with fewer appointments will be recommended with higher priority; the hospital with a smaller distance between the geographical location information of the registration department and the geographical location information of the user terminal will be recommended with higher priority; and the hospital where the registration department is located belongs to the exclusive hospital of the user object type will be recommended with higher priority; among them, different user guidance application scenarios are set with corresponding matching maximum appointment number thresholds; the priority of the priority recommendation generated based on the geographical location information of the user terminal, the priority of the priority recommendation order generated based on the user object type, and the priority of the priority recommendation generated based on the number of appointments are determined according to the priority level pre-set by the user, and the larger the priority level, the higher the corresponding priority recommendation order.

3. The intelligent medical treatment method with AI virtual person accompanying the patient according to claim 1 is characterized in that: Also includes: The generated preliminary diagnosis report is sent synchronously when the hospital information system of the appointment department is connected; when the preliminary diagnosis report request instruction is received from the terminal of the doctor sitting in the appointment department in the hospital information system where the appointment department is located, the preliminary diagnosis report is sent to the terminal of the doctor sitting in the appointment department; the preliminary diagnosis report is generated by integrating the key information extracted from the real-time user interaction information converted into text content and the AI ​​virtual person interaction response information using natural language technology, and combined with the preliminary diagnosis result; the real doctor's recommended examination items uploaded by the doctor sitting in the appointment department are received, and the examination items required for the confirmation of the preliminary diagnosis result are determined according to the medical knowledge map, and the determined examination items are compared with the real doctor's recommended examination items for similarity. If the similarity is lower than the first preset similarity, the real The doctor's recommended examination items are prompted for reconfirmation and the examination items required for confirming the preliminary diagnosis results are displayed. When the real doctor's recommended examination items are reconfirmed, the real doctor's recommended examination items are used as the final recommended examination items; when the user's examination item results uploaded by the hospital information system are received, the secondary diagnosis results are output based on the user's examination item results, the user's real-time interaction information and the user's historical intelligent guidance information, combined with the deep learning algorithm. The real doctor's diagnosis results uploaded by the terminal of the doctor in the registration department of the appointment are received, and the similarity is compared with the generated secondary diagnosis results. If the similarity is lower than the second preset similarity, the real doctor's diagnosis result reconfirmation information is generated and the preliminary diagnosis results are displayed on the terminal of the doctor in the registration department. When the real doctor's diagnosis result confirmation information is received, the real doctor's diagnosis information is used as the final diagnosis result.

4. The intelligent medical treatment method with AI virtual person accompanying the patient according to claim 3 is characterized in that: Also includes: When receiving the first route planning request instruction sent by the user terminal, a single fixed target location in the hospital where the appointment department is located and the current density of people in each area of ​​the hospital are obtained, and a first planned route for the user to reach the single fixed target location in the hospital where the appointment department is located is generated by using a random algorithm in combination with the determined user object type and the user guidance application scenario, which satisfies the first constraint condition; The first constraint condition includes: for the elderly or children, the crowd density in the area where the planned route passes is less than the first preset crowd density threshold and the distance to a single fixed target location in the hospital where the appointment registration department is located is the shortest; for the adult type, the crowd density in the area where the planned route passes is less than the second preset crowd density threshold and the distance to a single fixed target location in the hospital where the appointment registration department is located is the shortest; the first preset crowd density threshold is less than the second preset crowd density threshold; When receiving the second route planning request instruction sent by the user terminal, according to the final recommended inspection items, the location information of each inspection item in each area of ​​the hospital, the current density of people in each area of ​​the hospital and the current waiting time for each inspection item are obtained, and combined with the determined user object type and the user guidance application scenario, a random algorithm is used to generate a second planned route for users to complete all inspection items that meets the second constraint condition; the second constraint condition includes: for the elderly type or the child type, the density of people in the area where the planned route passes is less than the first preset density threshold, and the waiting time required to complete all inspection items is the shortest, and the distance to complete all inspection items is the shortest; for the adult type, the density of people in the area where the planned route passes is less than the second preset density threshold, and the waiting time required to complete all inspection items is the shortest, and the distance to complete all inspection items is the shortest.

5. The intelligent medical treatment method with AI virtual person accompanying the patient according to claim 1 is characterized in that: Also includes: When the preliminary diagnosis report is sent to the terminal of the doctor in the appointment department, the preliminary diagnosis result in the preliminary diagnosis report is in a pending unlocking state within a preset time period. When the preliminary diagnosis result acquisition request sent by the terminal of the doctor in the appointment department is received or after the preset time period, the preliminary diagnosis result unlocking password in the preliminary diagnosis report is sent to the terminal of the doctor in the appointment department.

6. The intelligent medical treatment method with AI virtual person accompanying the patient according to claim 1 is characterized in that: Also includes: Receive the medical advice information uploaded by the doctor in the appointment department, and generate personalized health management information for the user using deep learning algorithms based on the medical advice information and user profile information; The doctor's advice information includes: dietary taboo information, medication information and rehabilitation training information; the user's personalized health management information includes: user's personalized recipe recommendation, user medication reminder information, and rehabilitation training reminder information.

7. The intelligent medical treatment method with AI virtual person accompanying the patient according to claim 6 is characterized in that: Also includes: Receive the drug extraction completion information sent by the hospital information system of the appointment department, generate drug inspection information and display it on the user terminal; Receive the drug image and drug inspection request instruction uploaded by the user terminal, use image recognition technology to identify the drug information, compare the identified drug information with the drug information in the drug administration information to see if they are the same, if there is a difference, generate a drug error prompt message and send it to the hospital information system or user terminal where the appointment department is located; if the identified drug information includes prescription drug information, in addition to comparing the prescription drug information in the identified drug information with the prescription drug information in the drug administration information to see if they are the same, it is also necessary to find out whether there is a doctor's signature information in the received medical order information, if there is no doctor's signature information, then also generate a drug error prompt message and send it to the hospital information system where the appointment department is located and the doctor's terminal in the appointment department.

8. An intelligent medical consultation system with AI virtual person accompanying the patient, characterized in that: include: The virtual clinic building module is used to receive the intelligent clinic demand instructions sent by the user terminal, build a virtual clinic and generate an AI virtual person; The medical guidance information acquisition module is used to obtain the user input information in real time after judging that the user enters the virtual clinic based on the user terminal operation; the user input information includes the pre-entered user portrait information, the user selected intelligent medical guidance scene information, the user real-time interaction information and the user's historical intelligent medical guidance information; A diagnosis guidance influencing factor judgment module is used to determine the user object type based on the acquired user portrait information and / or the user's historical intelligent diagnosis guidance information; The user object types include: children, adults, the elderly, and people with interaction disorders; the user guidance application scenario is determined based on the acquired user historical intelligent guidance information and the intelligent guidance scenario information selected by the user; the user guidance application scenarios include: general consultation, emergency consultation, general follow-up consultation, and emergency follow-up consultation; The guidance and interactive emotion soothing module is used to use a deep learning algorithm to complete the recognition of abnormal user emotions based on the user's real-time interaction information, and the abnormal conditions include anxiety, tension, and sadness; based on the recognition of abnormal emotions and the determination of user object types, a preset emotion soothing strategy is matched, so that the AI ​​virtual person interacts with the user according to the matched emotion soothing strategy; wherein each user object type is matched with each abnormal emotion by a preset emotion soothing strategy; The guidance diagnosis and recommendation module is used to match the preset deep learning model according to the determined user guidance application scenario; according to the user's real-time interaction information and the user's historical intelligent guidance information, combined with the matched deep learning model, the AI ​​virtual human interaction reaction information and the preliminary diagnosis results are output; among which, the number of AI virtual human interaction reaction information output by different preset deep learning models and the output time of the preliminary diagnosis results are different; The guidance registration appointment operation module is used to determine the registration department based on the preliminary diagnosis results, and generate registration appointment recommendations based on the registration department; receive the registration appointment request information sent by the user terminal based on the registration appointment recommendation, connect to the hospital information system of the selected registration department and complete the registration appointment.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. A computer device, characterized in that: The computer device comprises a memory, a processor and a program stored and executable on the memory, and the program implements the steps of the method according to any one of claims 1 to 7 when executed by the processor.

Images