An ai-based intelligent diagnosis and prescription recommendation system for historical famous doctors and famous prescriptions

Abstract

This application discloses an AI-based intelligent diagnostic and prescription recommendation system based on famous prescriptions from renowned doctors throughout history. The system includes a human-computer interaction interface, a natural language processing module, a semantic analysis module, a speech recognition module, an image processing module, a medical terminology processing module, a feature extraction and pattern analysis module, and a prescription screening module. The speech recognition module converts captured speech signals into text information. The natural language processing and semantic analysis modules perform semantic understanding on the text information and extract structured TCM syndrome feature elements. The image processing module acquires and processes facial and tongue images of patients, recognizing visual features. The medical terminology processing module connects to a standard TCM terminology database, standardizing and mapping TCM syndrome feature elements and visual features. The feature extraction and pattern analysis module uses a deep learning model to perform diagnostic logic reasoning based on the standardized mapped features and outputs the diagnostic results.

Description

Technical Field

[0001] This invention relates to medical artificial intelligence systems, and more particularly to an AI-based intelligent diagnosis and prescription recommendation system based on famous prescriptions from renowned doctors throughout history. Background Technology

[0002] In the field of traditional Chinese medicine, syndrome differentiation and treatment is one of the core principles. However, the syndrome differentiation and prescription of traditional Chinese medicine rely heavily on the personal experience of physicians. This characteristic has led to the lack of uniformity in syndrome differentiation standards and increased difficulty in inheritance, resulting in a shortage of high-quality Chinese medicine resources.

[0003] Although some TCM AI systems exist in the current technology, these systems have obvious limitations in terms of syndrome differentiation logic, prescription sources, ability to link with intelligent diagnostic and treatment equipment, and deep integration of TCM resources.

[0004] For example, Chinese patent CN121191791A discloses a method for integrated diagnosis and treatment of traditional Chinese and Western medicine based on intelligent large model. This method solves the problems of data fragmentation between traditional Chinese and Western medicine, lack of personalization, and poor intelligence and interpretability. However, it fails to deeply explore and integrate the resources of famous doctors and prescriptions from past dynasties, and it also fails to achieve the goals of standardization of TCM consultation, intelligent diagnosis, and standardized prescription.

[0005] In addition, Chinese patent CN121075680A discloses a method and system for constructing a TCM prevention and treatment system for lung cancer based on multi-source data fusion. Although it effectively solves the problems of data fragmentation, static syndrome differentiation and translation bottlenecks in traditional research, it still has shortcomings in effective linkage with intelligent diagnostic and treatment equipment and in the deep integration of TCM resources.

[0006] To address the shortcomings of the existing technologies, this application aims to provide an AI-based intelligent diagnosis and prescription recommendation system that integrates resources from renowned doctors and prescriptions throughout history.

[0007] This system can perform standardized consultations, intelligent diagnosis, and precise prescription recommendations. It can also connect to various intelligent diagnostic and treatment devices to achieve standardization, intelligentization, and normalization of TCM diagnosis and treatment, thereby improving the accuracy, consistency, and scalability of TCM diagnosis and treatment.

[0008] By deeply exploring and integrating the resources of famous doctors and prescriptions from past dynasties, this system can overcome the limitations of traditional Chinese medicine's reliance on personal experience, realize intelligent diagnosis and prescription, thereby breaking through the bottleneck of existing technology and promoting the development of TCM diagnosis and treatment technology. Summary of the Invention

[0009] This application relates to an AI-based intelligent diagnosis and prescription recommendation system based on famous doctors and prescriptions throughout history. The system includes a human-computer interaction interface, a natural language processing module, a semantic analysis module, a speech recognition module, an image processing module, a medical terminology processing module, a feature extraction and regularity analysis module, and a prescription screening module.

[0010] The speech recognition module can convert captured speech signals into text information, while the natural language processing module and semantic analysis module perform semantic understanding on the text information and extract structured TCM syndrome feature elements.

[0011] The image processing module acquires and processes images of the patient's face and tongue, and identifies visual features.

[0012] The medical terminology processing module is connected to the TCM standard terminology database, which standardizes and maps the TCM syndrome feature elements and visual features.

[0013] The feature extraction and regularity analysis module uses a deep learning model to perform dialectical logical reasoning based on the standardized mapped features and outputs the dialectical results.

[0014] The prescription screening module connects to a database of famous prescriptions from renowned doctors throughout history, matching prescriptions that match the diagnosis results and calculating drug dosages.

[0015] Preferably, the natural language processing module integrates a combined model based on bidirectional long short-term memory network and conditional random field to perform word segmentation, part-of-speech tagging, and named entity recognition on doctor-patient dialogue text, and extract symptom features.

[0016] Furthermore, dependency parsing techniques are used to establish the association between symptom words and modifiers, generating symptom vectors. These symptom vectors are then input into a pre-constructed TCM knowledge graph through a semantic analysis module, and transformed into standard TCM syndrome codes.

[0017] Preferably, the image processing module uses a mask region convolutional neural network to segment the tongue image, extract tongue features, locate key points in the facial image, and identify facial color parameters and gloss parameters.

[0018] Preferably, the system includes a four-diagnostic instrument terminal, equipped with a pressure sensor, to perform real-time acquisition of pulse signals.

[0019] The feature extraction and regularity analysis module uses wavelet transform technology to denoise the pulse signal, extract dynamic parameters, and construct a multimodal fusion dialectical matrix.

[0020] Preferably, the deep learning model in the feature extraction and regularity analysis module adopts a multi-head attention mechanism to perform feature weighting processing on the inspection data, auscultation data, inquiry data and palpation data respectively, and outputs candidate syndrome types and their probability distributions based on the syndrome differentiation results.

[0021] Preferably, the prescription screening module integrates a general Reactor model, which dynamically adjusts the initial dosage of each drug in the prescription based on patient information, and modifies the prescription according to the theory of four qi and five flavors in traditional Chinese medicine.

[0022] The Reactor model has a built-in algorithm for checking drug incompatibilities. It performs safety checks on the initially screened prescriptions, including automatic screening of the Eighteen Incompatibilities, Nineteen Antagonisms, and drugs contraindicated during pregnancy. When a drug incompatibility is found, it triggers an alternative drug search procedure.

[0023] Preferably, the system includes an interface to the HIS / pharmacies management system and a Chatbot subsystem, which transmits the generated prescription suggestions and syndrome differentiation analysis reports to external medical business systems. The Chatbot subsystem provides patients with medication guidance, dietary restrictions reminders, and follow-up appointment services.

[0024] The system in this application achieves intelligent processing of the entire process from patient information collection and syndrome differentiation analysis to prescription recommendation through deep learning models and multimodal fusion technology, thereby improving the efficiency and accuracy of TCM diagnosis and treatment. Attached Figure Description

[0025] To more clearly illustrate the embodiments of this application, the accompanying drawings used in this application will be briefly described below. Detailed Implementation

[0026] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0027] like Figure 1As shown, this application provides an AI-based intelligent diagnosis and prescription recommendation system based on famous prescriptions from renowned doctors throughout history, including: a human-computer interaction interface 31, a natural language processing module 21, a semantic analysis module 22, a speech recognition module 23, an image processing module 24, a medical terminology processing module 25, a feature extraction and regularity analysis module 26, and a prescription screening module 27; the speech recognition module 23 is configured to convert captured speech signals into text information; the natural language processing module 21 and the semantic analysis module 22 are configured to perform semantic understanding on the text information and extract structured TCM syndrome feature elements; the image processing module 24 is configured to acquire and process the patient's facial and tongue images and identify visual features; the medical terminology processing module 25 is connected to a preset TCM standard terminology database 13 and is configured to perform standardized mapping on the TCM syndrome feature elements and visual features; the feature extraction and regularity analysis module 26 is configured to use a deep learning model to perform diagnostic logic reasoning based on the standardized mapped features and output the diagnosis results; the prescription screening module 27 is connected to a database of famous prescriptions from renowned doctors throughout history 11. The connection is configured to match prescriptions that correspond to the diagnostic results and calculate drug dosages.

[0028] Preferably, such as Figure 1 As shown, the natural language processing module 21 integrates a combined model based on bidirectional long short-term memory network and conditional random field. The combined model is configured to perform word segmentation, part-of-speech tagging and named entity recognition on doctor-patient dialogue text, and extract symptom features including onset time, symptom location, nature, degree and cause.

[0029] Preferably, such as Figure 1 As shown, the natural language processing module 21 is further configured to establish the association between symptom words and modifiers using dependency parsing technology to generate symptom vectors; the semantic analysis module 22 is configured to input the symptom vectors into a pre-constructed TCM knowledge graph, and transform non-standardized language descriptions into standard TCM syndrome codes through entity links and concept mapping.

[0030] Preferably, such as Figure 1 As shown, the image processing module 24 is configured to segment the tongue image using a mask region convolutional neural network and extract tongue features including tongue color, tongue coating color, tongue coating thickness parameters, tongue shape and teeth mark features; the image processing module 24 is further configured to locate key points in the facial image and identify facial color parameters and gloss parameters.

[0031] Preferably, such as Figure 1As shown, the system further includes a four-diagnosis instrument terminal 32, which is equipped with a pressure sensor for real-time acquisition of pulse signals. The feature extraction and regularity analysis module 26 is configured to use wavelet transform technology to denoise the pulse signals, extract dynamic parameters including pulse rate, pulse force, pulse position and pulse momentum, and combine the visual features output by the image processing module 24 and the symptom features output by the natural language processing module 21 to construct a multimodal fusion diagnostic matrix.

[0032] Preferably, such as Figure 1 As shown, the deep learning model in the feature extraction and regularity analysis module 26 adopts a multi-head attention mechanism, which is configured to perform feature weighting processing on the inspection data, auscultation data, inquiry data and palpation data respectively, and output the candidate syndrome type and its probability distribution of the current patient according to the syndrome differentiation results.

[0033] Preferably, such as Figure 1 As shown, the prescription screening module 27 integrates a general Reactor model. The Reactor model is configured to dynamically adjust the initial dosage of each drug in the prescription based on the patient's age, weight, disease course parameters and seasonal factors, and to modify the prescription according to the theory of four qi and five flavors in traditional Chinese medicine.

[0034] Preferably, such as Figure 1 As shown, the Reactor model has a built-in incompatibility checking algorithm. The algorithm is configured to perform safety verification on the initially screened prescriptions by searching the Chinese medicine incompatibility matrix. The safety verification includes automatic screening of the Eighteen Incompatibilities, Nineteen Antagonisms, and drugs contraindicated during pregnancy. The Reactor model is further configured to trigger an alternative drug search procedure when incompatibility is found, and to find drugs with the same medicinal properties and no incompatibility conflicts for replacement.

[0035] Preferably, such as Figure 1 As shown, the system further includes an HIS / pharmacies management system interface 33 and a Chatbot dialogue subsystem 34; the system is configured to transmit the generated prescription suggestions and syndrome differentiation analysis reports to an external medical business system through the HIS / pharmacies management system interface 33; the Chatbot dialogue subsystem 34 is configured to use generative artificial intelligence technology to provide patients with medication guidance, dietary restrictions reminders, and follow-up appointment services.

[0036] The preferred embodiments of the present invention have been described in detail above.

[0037] It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of this invention without creative effort.

[0038] Therefore, any technical solution that can be obtained by those skilled in the art based on the concept of this invention and through logical analysis, reasoning, or limited experimentation on the basis of the prior art should be within the scope of protection defined by the claims.

Claims

1. An AI-based intelligent diagnostic and prescription recommendation system based on famous prescriptions from renowned doctors throughout history, characterized in that: The system includes: a human-computer interaction interface (31), a natural language processing module (21), a semantic analysis module (22), a speech recognition module (23), an image processing module (24), a medical terminology processing module (25), a feature extraction and regularity analysis module (26), and a prescription screening module (27); the speech recognition module (23) is configured to convert the captured speech signal into text information; the natural language processing module (21) and the semantic analysis module (22) are configured to perform semantic understanding on the text information and extract structured TCM syndrome feature elements; the image processing module... Block (24) is configured to acquire and process the patient's facial and tongue images and identify visual features; the medical terminology processing module (25) is connected to the preset TCM standard terminology library (13) and is configured to perform standardized mapping on the TCM syndrome feature elements and visual features; the feature extraction and regularity analysis module (26) is configured to use a deep learning model to perform syndrome differentiation logic reasoning based on the standardized mapped features and output the syndrome differentiation results; the prescription screening module (27) is connected to the database of famous doctors and prescriptions throughout history (11) and is configured to match prescriptions that match the syndrome differentiation results and calculate drug dosages.

2. The AI-based intelligent diagnosis and prescription recommendation system based on famous doctors and prescriptions throughout history, as described in claim 1, is characterized in that... The natural language processing module (21) integrates a combined model based on bidirectional long short-term memory network and conditional random field. The combined model is configured to perform word segmentation, part-of-speech tagging and named entity recognition on doctor-patient dialogue text, and extract symptom features including onset time, symptom location, nature, degree and cause.

3. The AI-based intelligent diagnosis and prescription recommendation system based on famous doctors and prescriptions throughout history, as described in claim 2, is characterized in that... The natural language processing module (21) is further configured to establish the association between symptom words and modifiers using dependency parsing technology to generate symptom vectors; the semantic analysis module (22) is configured to input the symptom vectors into a pre-constructed TCM knowledge graph and convert non-standardized language descriptions into standard TCM syndrome codes through entity links and concept mapping.

4. The AI-based intelligent diagnosis and prescription recommendation system based on famous doctors and prescriptions throughout history, as described in claim 1, is characterized in that... The image processing module (24) is configured to segment the tongue image using a mask region convolutional neural network and extract tongue features including tongue color, tongue coating color, tongue coating thickness parameters, tongue shape and teeth mark features; the image processing module (24) is further configured to locate key points in the facial image and identify facial color parameters and gloss parameters.

5. The AI-based intelligent diagnosis and prescription recommendation system based on famous doctors and prescriptions throughout history, as described in claim 1, is characterized in that... The system further includes a four-diagnosis instrument terminal (32), which is equipped with a pressure sensor for real-time acquisition of pulse signals; the feature extraction and regularity analysis module (26) is configured to use wavelet transform technology to denoise the pulse signals, extract dynamic parameters including pulse rate, pulse force, pulse position and pulse momentum, and combine the visual features output by the image processing module (24) and the symptom features output by the natural language processing module (21) to construct a multimodal fusion diagnostic matrix.

6. The AI-based intelligent diagnosis and prescription recommendation system based on famous doctors and prescriptions throughout history, as described in claim 5, is characterized in that... The deep learning model in the feature extraction and regularity analysis module (26) adopts a multi-head attention mechanism and is configured to perform feature weighting processing on the inspection data, auscultation data, inquiry data and palpation data respectively, and output the candidate syndrome type and its probability distribution of the current patient according to the syndrome differentiation results.

7. The AI-based intelligent diagnosis and prescription recommendation system based on famous doctors and prescriptions throughout history, as described in claim 1, is characterized in that... The prescription screening module (27) integrates a general Reactor model. The Reactor model is configured to dynamically adjust the initial dosage of each drug in the prescription based on the patient's age, weight, disease course parameters and seasonal factors, and to modify the prescription according to the theory of four qi and five flavors in traditional Chinese medicine.

8. The AI-based intelligent diagnosis and prescription recommendation system based on famous doctors and prescriptions throughout history, as described in claim 7, is characterized in that... The Reactor model has a built-in incompatibility check algorithm. The algorithm is configured to perform a safety check on the initially screened prescriptions by searching the Chinese medicine incompatibility matrix. The safety check includes automatic screening for the Eighteen Incompatibilities, Nineteen Antagonisms, and drugs contraindicated during pregnancy. The Reactor model is further configured to trigger an alternative drug search procedure when incompatibility is found, and to find drugs with the same medicinal properties and no incompatibility conflicts for replacement.

9. The AI-based intelligent diagnosis and prescription recommendation system based on famous doctors and prescriptions throughout history, as described in claim 1, is characterized in that... The system further includes an HIS / Pharmacy Management System interface (33) and a Chatbot subsystem (34); the system is configured to transmit the generated prescription suggestions and syndrome differentiation analysis reports to an external medical business system through the HIS / Pharmacy Management System interface (33); the Chatbot subsystem (34) is configured to provide patients with medication guidance, dietary restrictions reminders and follow-up appointment services using generative artificial intelligence technology.

Images