AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine

Abstract

The present invention relates to the field of AI diagnosis and treatment technology for identification and formulation optimization of Chinese medicinal materials, and specifically discloses an AI-assisted diagnosis and treatment system for intelligent identification and formulation optimization of Chinese medicinal materials, including: a symptom similarity acquisition module, a basic dosage confirmation module, a Chinese medicinal material quality identification module, a Chinese medicinal material formulation optimization module, a formulation dynamic adjustment module, and a database. The present invention achieves quantitative matching of symptom vectors through unique-hot encoding and cosine similarity, determines the type based on statistics of high-frequency medicinal materials in historical cases, utilizes a multimodal data fusion algorithm to evaluate the quality of medicinal materials, and dynamically adjusts the dosage based on quality deviation and physiological parameters. This breaks through the subjectivity and empirical limitations of traditional Chinese medicine diagnosis and treatment, realizes the objectivity of symptom differentiation, intelligent identification of medicinal materials, and dynamic optimization of dosage, and constructs a full-process closed loop of "data collection-analysis-prescription adjustment", significantly improving the accuracy and efficacy of Chinese medicine diagnosis and treatment, and is suitable for intelligent diagnosis and treatment scenarios in Chinese medicine halls.

Description

Technical Field

[0001] The present invention relates to the field of AI diagnosis and treatment technology for traditional Chinese medicine identification and formula optimization, and specifically to an AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine. Background Art

[0002] Although the traditional Chinese medicine diagnosis and treatment system has the unique advantages of holistic view and syndrome differentiation and treatment, it faces significant challenges in modern application. Therefore, in order to break through the subjective and empirical limitations of traditional Chinese medicine diagnosis and treatment, and improve the standardization level of Chinese medicinal materials quality control and the individualized accuracy of prescription compatibility, it is urgent to build an intelligent and data-based AI-assisted diagnosis and treatment system.

[0003] The existing technology still has the following problems: 1. It only relies on the physician's experience or fixed dosage template to determine the dosage of Chinese medicinal materials required by the patient, without combining dynamic adjustment with the quality of the medicinal materials. There is a lack of an automated response mechanism to the differences in the quality of the medicinal materials, which can easily lead to unstable therapeutic effects.

[0004] 2. The quality assessment of Chinese medicinal materials mostly relies on a single detection method (such as only appearance inspection or chemical composition analysis), and multimodal data fusion (dual analysis of color and smell) analysis is not proposed. This makes it impossible to comprehensively assess the quality of Chinese medicinal materials, which may lead to the missed detection of inferior medicinal materials.

[0005] 3. Most prescription recommendations are static, and it is impossible to optimize the dosage for the next course of treatment based on the patient's real-time physiological parameters and changes in the quality of medicinal materials after taking the medicine, thus failing to achieve a closed-loop process of "treatment-feedback-optimization". Summary of the Invention

[0006] In view of this, in order to solve the problems raised in the above background technology, an AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine is proposed.

[0007] The purpose of the present invention can be achieved through the following technical solutions: The present invention provides an AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine, including: a symptom similarity acquisition module, which extracts the types of diseases and various symptoms input by the current patient into the AI ​​diagnosis and treatment platform of the target Chinese medicine store, and extracts various symptoms corresponding to historical patients with similar diseases in the historical library of the AI ​​diagnosis and treatment platform, and obtains the symptom vector similarity between the current patient and each historical patient based on this.

[0008] The basic dosage confirmation module extracts the Chinese herbal medicine usage information of each historical patient and confirms the type of Chinese herbal medicine and the basic dosage of each type of Chinese herbal medicine required by the current patient based on the similarity of symptom vectors.

[0009] The Chinese medicinal materials quality identification module starts the automatic medicinal materials grabbing machine of the target Chinese medicine store and uses it to collect quality information of various Chinese medicinal materials, and uses a multimodal data fusion algorithm to obtain the quality qualification of various Chinese medicinal materials.

[0010] The Chinese herbal medicine formula optimization module adjusts the initial dosage of various Chinese herbal medicines required by the current patient based on the basic dosage and quality qualification of various Chinese herbal medicines.

[0011] The dynamic prescription adjustment module collects the real-time physiological parameters of the current patient at the end of the current medication course, and collects the quality information of various Chinese medicinal materials in the target Chinese medicine store after the end of the current medication course, and optimizes the dosage of various Chinese medicinal materials required for the current patient in the next medication course accordingly.

[0012] Specifically, the specific process of obtaining the similarity of the symptom vectors of the current patient and each historical patient is: summarizing all historical patients with similar diseases in the history library of the AI ​​diagnosis and treatment platform and all corresponding symptoms of the current patient to form a symptom set.

[0013] Based on the symptom set, the symptom vector corresponding to the disease of the current patient and the symptom vector corresponding to each historical patient are generated by using one-hot encoding.

[0014] The cosine similarity calculation is performed based on the symptom vector corresponding to the disease of the current patient and the symptom vector corresponding to each historical patient to obtain the similarity between the symptom vector of the current patient and each historical patient.

[0015] Specifically, the specific process of confirming the type of Chinese medicinal materials that the current patient needs to take is: comparing the symptom vector similarity between the current patient and each historical patient with the symptom vector similarity of the set reference; if the symptom vector similarity between the current patient and a historical patient is greater than or equal to the symptom vector similarity of the set reference, then the historical patient is recorded as a highly similar patient, thereby screening out the highly similar patients corresponding to the current patient.

[0016] The types of Chinese medicinal materials taken by each highly similar patient and the initial dosage of each type of Chinese medicinal material are extracted from the Chinese medicinal materials taking information of each historical patient.

[0017] The number of times each type of Chinese medicinal material appears is counted, and the ratio between the number of times each type of Chinese medicinal material appears and the number of highly similar patients is used as the frequency of occurrence of each type of Chinese medicinal material.

[0018] The frequency of occurrence of each type of Chinese medicinal materials is compared with the frequency of occurrence of the set reference medicinal materials. If the frequency of occurrence of a certain type of Chinese medicinal materials is greater than or equal to the frequency of occurrence of the set reference medicinal materials, then this type of Chinese medicinal materials will be recorded as the target Chinese medicinal materials, thereby obtaining the target Chinese medicinal materials type, and using it as the type of Chinese medicinal materials that the current patient needs to take.

[0019] Specifically, the specific method of confirming the basic dosage of each type of Chinese medicinal materials that the current patient needs to take is: extracting the initial dosage of each type of Chinese medicinal materials that the current patient needs to take from the initial dosage of each type of Chinese medicinal materials taken by each highly similar patient, and calculating the average thereof, and using the calculation result as the basic dosage of each type of Chinese medicinal materials that the current patient needs to take.

[0020] Specifically, the specific process of using the multimodal data fusion algorithm to obtain the quality qualification of various Chinese medicinal materials is: performing color processing based on the grayscale image in the quality information of various Chinese medicinal materials to obtain the color qualification of various Chinese medicinal materials.

[0021] Based on the characteristic values ​​of the odor component spectrum in the quality information of various Chinese medicinal materials and the characteristic standard values ​​of the standard component spectrum of various Chinese medicinal materials stored in the database, a coupling analysis is performed to obtain the odor qualification of various Chinese medicinal materials.

[0022] The color qualification and smell qualification of each type of Chinese medicinal materials are weightedly calculated and summed with the corresponding weights to obtain the quality qualification of each type of Chinese medicinal materials.

[0023] Specifically, the specific method of obtaining the color qualification of various Chinese medicinal materials is: extracting the grayscale value corresponding to each grayscale area from the grayscale image of each Chinese medicinal material, and comparing it with the grayscale value interval corresponding to the standard conditions of each Chinese medicinal material stored in the database.

[0024] If the grayscale value corresponding to a grayscale area of ​​a certain type of Chinese medicinal materials is not within the grayscale value interval corresponding to its standard condition, the grayscale area of ​​this type of Chinese medicinal materials is recorded as an abnormal area, and the number of abnormal areas of each type of Chinese medicinal materials is counted.

[0025] The number of grayscale regions corresponding to each type of Chinese medicinal material is counted from the grayscale images of each type of Chinese medicinal material, and the ratio between the number of abnormal regions of each type of Chinese medicinal material and the number of grayscale regions is recorded as the abnormal region ratio of each type of Chinese medicinal material.

[0026] The difference between the proportion of abnormal areas of the set reference and the proportion of abnormal areas of each type of Chinese medicinal materials is compared with the proportion of abnormal areas of the set reference, and the ratio result is used as the color qualification of each type of Chinese medicinal materials.

[0027] Specifically, the specific process of obtaining the odor qualification of each type of Chinese medicinal material is: combining each characteristic value in the odor component spectrum of each type of Chinese medicinal material into a multidimensional vector as the multidimensional characteristic vector of each type of Chinese medicinal material.

[0028] The characteristic standard values ​​in the standard component spectra of various Chinese medicinal materials stored in the database are combined into a multidimensional vector as the standard characteristic vector of each Chinese medicinal material.

[0029] Obtain the Euclidean distance between the multidimensional feature vector and the standard feature vector of each type of Chinese medicinal material.

[0030] The smell qualification of each type of Chinese medicinal material is obtained by calculating the Euclidean distance between the multidimensional feature vector and the standard feature vector of each type of Chinese medicinal material.

[0031] Specifically, the specific process of adjusting the initial dosage of each type of Chinese medicinal material that the current patient needs to take is: comparing the quality qualification of each type of Chinese medicinal material with the quality qualification of the set reference; if the quality qualification of a certain type of Chinese medicinal material is greater than the quality qualification of the set reference, then this type of Chinese medicinal material is recorded as a high-quality medicinal material; if the quality qualification of a certain type of Chinese medicinal material is less than the quality qualification of the set reference, then this type of Chinese medicinal material is recorded as a low-quality medicinal material; if the quality qualification of a certain type of Chinese medicinal material is equal to the quality qualification of the set reference, then this type of Chinese medicinal material is recorded as a standard quality medicinal material.

[0032] If the Chinese medicinal materials that the patient currently needs to take are of standard quality, the basic dosage of this type of Chinese medicinal materials will remain unchanged.

[0033] If the Chinese medicinal materials that the current patient needs to take are high-quality medicinal materials, the difference between the quality qualification of this type of Chinese medicinal materials and the set reference quality qualification is obtained, and the product of the difference and the required adjustment dose corresponding to the unit quality qualification deviation of this type of Chinese medicinal materials stored in the database is used as the required reduced dose of this type of Chinese medicinal materials, and the difference between the basic dose and the required reduced dose is used as the initial dose of this type of Chinese medicinal materials.

[0034] If the Chinese medicinal materials that the current patient needs to take are of low quality, the difference between the set reference quality qualification and the quality qualification of this type of Chinese medicinal materials is obtained, and the product of the difference and the required adjustment dose corresponding to the unit quality qualification deviation of this type of Chinese medicinal materials stored in the database is used as the required increased dose of this type of Chinese medicinal materials, and the sum of the basic dose and the required increased dose is used as the initial dose of this type of Chinese medicinal materials.

[0035] In summary, the initial dosage of various Chinese medicinal materials that the current patient needs to take is obtained.

[0036] Specifically, the specific process of optimizing the dosage of various Chinese medicinal materials required for the current patient in the next course of medication is: based on the pulse rate and tongue coating image of the current patient's real-time physiological parameters corresponding to the end of the current course of medication, a fusion processing is performed to obtain the physiological health index of the current patient at the end of the current course of medication.

[0037] Based on the quality information of various Chinese medicinal materials in the target Chinese medicine hall after the end of the current medication course, the quality qualification of various Chinese medicinal materials after the end of the current medication course is analyzed in the same way as the analysis method of the quality qualification of various Chinese medicinal materials.

[0038] Based on the physiological health index corresponding to the current patient at the end of the current medication course and the quality qualification of various Chinese medicinal materials after the end of the current medication course, an optimization analysis is performed to obtain the dosage of various Chinese medicinal materials required for the current patient in the next medication course.

[0039] Specifically, the specific process of obtaining the physiological health index corresponding to the end of the current medication course of the current patient is: performing pulse abnormality analysis based on the pulse rate corresponding to the end of the current medication course of the current patient, and obtaining the pulse abnormality degree corresponding to the end of the current medication course of the current patient.

[0040] Based on the tongue coating picture corresponding to the end of the current medication course of the current patient, the tongue coating color abnormality analysis is performed to obtain the tongue coating color abnormality degree corresponding to the end of the current medication course of the current patient.

[0041] A comprehensive analysis is performed on the abnormality of the pulse and the abnormality of the tongue coating color corresponding to the end of the current medication course of the current patient to obtain the physiological health index corresponding to the end of the current medication course of the current patient.

[0042] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention generates symptom vectors through one-hot encoding, and calculates the symptom matching degree between the current patient and historical cases by combining cosine similarity, thereby converting the TCM "differentiation" into a quantifiable numerical analysis and reducing subjective errors.

[0043] (2) The present invention adopts a multimodal data fusion algorithm, analyzes the color qualification through grayscale images, and calculates the odor qualification based on the Euclidean distance of the odor component spectrum, thereby realizing a digital evaluation of the color and odor of Chinese medicinal materials, improving the accuracy of the quality judgment of Chinese medicinal materials, ensuring the quality of the medicine, and at the same time guaranteeing the stability of the therapeutic effect.

[0044] (3) The present invention adjusts the initial dose based on the quality qualification. For high-quality medicinal materials, the dose is reduced according to the deviation ratio, and for low-quality medicinal materials, the dose is increased according to the deviation ratio, thereby avoiding the risk of insufficient efficacy or toxicity caused by fluctuations in the medicinal material composition.

[0045] (4) The present invention dynamically adjusts prescriptions by combining physiological parameters, comprehensively calculating the physiological health index through the abnormality of pulse rate and tongue coating color, and optimizing the dosage of the next course of treatment while taking into account the real-time quality of medicinal materials and patient efficacy feedback, forming a "monitoring-adjustment" closed loop, thereby improving the efficiency of the course of treatment and realizing dynamic adjustment of the Chinese medicinal material formula. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 This is a schematic diagram of the system module structure connection of the present invention.

[0048] Figure 2 Schematic diagram of the quality analysis process of the Chinese medicinal materials of the present invention.

[0049] Figure 3 This is a pulse rate fluctuation diagram of a normal person under normal circumstances according to the present invention. DETAILED DESCRIPTION

[0050] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0051] See also Figure 1 As shown, the present invention provides an AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine, including: a symptom similarity acquisition module, a basic dosage confirmation module, a traditional Chinese medicine quality identification module, a traditional Chinese medicine formula optimization module and a formula dynamic adjustment module.

[0052] It should be noted that the present invention also includes a database for storing the grayscale value intervals corresponding to various types of Chinese medicinal materials under standard conditions, the characteristic standard values ​​in the standard component spectra of various types of Chinese medicinal materials, the required adjustment doses corresponding to the unit quality qualification deviations of various types of Chinese medicinal materials, the pulse rate intervals of normal people under normal conditions, and the RGB values ​​in the tongue coating pictures.

[0053] The symptom similarity acquisition module is connected to the basic dose confirmation module, the basic dose confirmation module and the Chinese medicinal material quality identification module are both connected to the Chinese medicinal material formula optimization module, the Chinese medicinal material quality identification module and the Chinese medicinal material formula optimization module are both connected to the formula dynamic adjustment module, and the Chinese medicinal material quality identification module, the Chinese medicinal material formula optimization module and the formula dynamic adjustment module are all connected to the database.

[0054] The symptom similarity acquisition module extracts the types of diseases and symptoms input by the current patient into the AI ​​diagnosis and treatment platform of the target Chinese medicine clinic, and extracts the various symptoms corresponding to each historical patient with the same disease in the historical library of the AI ​​diagnosis and treatment platform, thereby obtaining the symptom vector similarity between the current patient and each historical patient.

[0055] In a specific embodiment of the present invention, the types of diseases include but are not limited to colds, stomachaches and high blood pressure. The symptoms corresponding to the colds include but are not limited to fever, clear runny nose, sneezing, headache and pale red tongue. The symptoms corresponding to the stomachache include but are not limited to weakness in the limbs, constipation, white tongue coating and stringy pulse. The symptoms corresponding to high blood pressure include but are not limited to dizziness, tinnitus, bitter taste in the mouth and yellow tongue coating.

[0056] In a specific embodiment of the present invention, the specific process of obtaining the similarity of the symptom vectors of the current patient and each historical patient is: summarizing all historical patients with similar diseases in the historical library of the AI ​​diagnosis and treatment platform and all corresponding symptoms of the current patient to form a symptom set.

[0057] Based on the symptom set, the symptom vector corresponding to the disease of the current patient and the symptom vector corresponding to each historical patient are generated by using one-hot encoding.

[0058] It should be noted that the specific method of generating the symptom vector corresponding to the disease suffered by the current patient and the symptom vector corresponding to each historical patient based on the symptom set by using unique-hot encoding is: each type of symptom corresponding to the disease suffered by the current patient is compared one by one with each type of symptom in the symptom set. If the current patient has a certain type of symptom, the corresponding dimension value is 1; if the current patient does not have a certain type of symptom, the corresponding dimension value is 0, thereby obtaining the symptom vector corresponding to the disease suffered by the current patient.

[0059] Compare each type of symptom corresponding to each historical patient with each type of symptom in the symptom set one by one. If a historical patient has a certain type of symptom, the corresponding dimension value is 1. If a historical patient does not have a certain type of symptom, the corresponding dimension value is 0. In this way, the symptom vector corresponding to each historical patient is obtained.

[0060] In a specific embodiment of the present invention, if the type of disease is a cold, the symptom set is defined as {fever, runny nose, sneezing, headache, pale red tongue}. If a historical patient A has symptoms of "fever and runny nose", then its vector A = (1, 1, 0, 0, 0); if the current patient B has symptoms of "fever, sneezing, headache", then its vector B = (1, 0, 1, 1, 0).

[0061] The cosine similarity calculation is performed based on the symptom vector corresponding to the disease of the current patient and the symptom vector corresponding to each historical patient to obtain the similarity between the symptom vector of the current patient and each historical patient.

[0062] It should be noted that the specific method of obtaining the similarity between the symptom vector of the current patient and each historical patient is: multiply the dimension values ​​in the symptom vector corresponding to the disease suffered by the current patient and the symptom vector corresponding to each historical patient, and then add them together, and use the addition result as the dot product between the symptom vector of the current patient and each historical patient.

[0063] The square root of the sum of the squares of the dimensional values ​​in the symptom vector corresponding to the disease suffered by the current patient is taken as the modulus of the symptom vector of the current patient.

[0064] The square root of the sum of the squares of the dimensional values ​​in the symptom vector of each historical patient is taken as the modulus of the symptom vector of each historical patient.

[0065] The dot product between the symptom vectors of the current patient and each historical patient is divided by the product between the modulus of the symptom vector of the current patient and the modulus of the symptom vector of each historical patient, and the division result is used as the similarity between the symptom vectors of the current patient and each historical patient.

[0066] The embodiment of the present invention generates a symptom vector through one-hot encoding, and calculates the symptom matching degree between the current patient and historical cases in combination with cosine similarity, thereby converting the TCM "syndrome differentiation" into a quantifiable numerical analysis and reducing subjective errors.

[0067] The basic dosage confirmation module extracts the Chinese herbal medicine taking information of each historical patient and confirms the type of Chinese herbal medicine and the basic dosage of each type of Chinese herbal medicine required by the current patient based on the symptom vector similarity.

[0068] It should be noted that the Chinese medicinal materials usage information of each historical patient is directly extracted from the AI ​​diagnosis and treatment platform of the target Chinese medicine clinic.

[0069] In a specific embodiment of the present invention, the specific process of confirming the type of Chinese medicinal materials that the current patient needs to take is: comparing the symptom vector similarity between the current patient and each historical patient with the symptom vector similarity of the set reference; if the symptom vector similarity between the current patient and a historical patient is greater than or equal to the symptom vector similarity of the set reference, the historical patient is recorded as a highly similar patient, thereby screening out the highly similar patients corresponding to the current patient.

[0070] The types of Chinese medicinal materials taken by each highly similar patient and the initial dosage of each type of Chinese medicinal material are extracted from the Chinese medicinal materials taking information of each historical patient.

[0071] The number of times each type of Chinese medicinal material appears is counted, and the ratio between the number of times each type of Chinese medicinal material appears and the number of highly similar patients is used as the frequency of occurrence of each type of Chinese medicinal material.

[0072] The frequency of occurrence of each type of Chinese medicinal materials is compared with the frequency of occurrence of the set reference medicinal materials. If the frequency of occurrence of a certain type of Chinese medicinal materials is greater than or equal to the frequency of occurrence of the set reference medicinal materials, then this type of Chinese medicinal materials will be recorded as the target Chinese medicinal materials, thereby obtaining the target Chinese medicinal materials type, and using it as the type of Chinese medicinal materials that the current patient needs to take.

[0073] In a specific embodiment of the present invention, the specific method of confirming the basic dosage of each type of Chinese medicinal material that the current patient needs to take is: extracting the initial dosage of each type of Chinese medicinal material that the current patient needs to take from the initial dosage of each type of Chinese medicinal material taken by each highly similar patient, and calculating the average thereof, and using the calculation result as the basic dosage of each type of Chinese medicinal material that the current patient needs to take.

[0074] The Chinese medicinal material quality identification module starts the medicinal material automatic grabbing machine of the target Chinese medicine store and collects the quality information of various Chinese medicinal materials through it, and adopts a multimodal data fusion algorithm to obtain the quality qualification of various Chinese medicinal materials.

[0075] See also Figure 2 As shown, in a specific embodiment of the present invention, the specific process of using the multimodal data fusion algorithm to obtain the quality qualification of various types of Chinese medicinal materials is: color processing is performed based on the grayscale image in the quality information of various types of Chinese medicinal materials to obtain the color qualification of various types of Chinese medicinal materials.

[0076] It should be noted that the grayscale images of the various Chinese medicinal materials are acquired by a high-definition camera installed on an automatic grasping machine, and the acquired images are imported into professional image processing software for grayscale processing to obtain grayscale images of the various Chinese medicinal materials.

[0077] In a specific embodiment of the present invention, the specific method of obtaining the color qualification of various types of Chinese medicinal materials is: extracting the grayscale value corresponding to each grayscale area from the grayscale image of each type of Chinese medicinal material, and comparing it with the grayscale value interval corresponding to the standard conditions of each type of Chinese medicinal material stored in the database.

[0078] If the grayscale value corresponding to a grayscale area of ​​a certain type of Chinese medicinal materials is not within the grayscale value interval corresponding to its standard condition, the grayscale area of ​​this type of Chinese medicinal materials is recorded as an abnormal area, and the number of abnormal areas of each type of Chinese medicinal materials is counted.

[0079] The number of grayscale regions corresponding to each type of Chinese medicinal material is counted from the grayscale images of each type of Chinese medicinal material, and the ratio between the number of abnormal regions of each type of Chinese medicinal material and the number of grayscale regions is recorded as the abnormal region ratio of each type of Chinese medicinal material.

[0080] The difference between the proportion of abnormal areas of the set reference and the proportion of abnormal areas of each type of Chinese medicinal materials is compared with the proportion of abnormal areas of the set reference, and the ratio result is used as the color qualification of each type of Chinese medicinal materials.

[0081] Based on the characteristic values ​​of the odor component spectrum in the quality information of various Chinese medicinal materials and the characteristic standard values ​​of the standard component spectrum of various Chinese medicinal materials stored in the database, a coupling analysis is performed to obtain the odor qualification of various Chinese medicinal materials.

[0082] It should be noted that the odor component spectra of the various types of Chinese medicinal materials are collected in the following manner: 1) Data acquisition: The electronic nose sensor is usually composed of multiple sensor arrays with different sensitivities to different odor components. When the odor molecules of the Chinese medicinal materials come into contact with the sensor array of the electronic nose placed on the automatic grasping machine, each sensor will generate a corresponding electrical signal change; 2) Signal preprocessing: Use a digital filter (such as a low-pass filter) to remove high-frequency noise to make the signal smoother, and then unify the signal values ​​of each sensor to a specific range (such as between 0 and 1) to eliminate the influence of sensitivity differences between different sensors. Finally, automatically extract parameters that can represent the odor characteristics from the preprocessed signal (common eigenvalues ​​include the peak value, mean, and standard deviation of the signal); 3) Pattern generation: The extracted eigenvalues ​​are combined into a multidimensional vector, which is the odor component spectrum.

[0083] In a specific embodiment of the present invention, the specific process of obtaining the odor qualification of each type of Chinese medicinal material is: combining each characteristic value in the odor component spectrum of each type of Chinese medicinal material into a multidimensional vector as the multidimensional characteristic vector of each type of Chinese medicinal material.

[0084] The characteristic standard values ​​in the standard component spectra of various Chinese medicinal materials stored in the database are combined into a multidimensional vector as the standard characteristic vector of each Chinese medicinal material.

[0085] Obtain the Euclidean distance between the multidimensional feature vector and the standard feature vector of each type of Chinese medicinal material.

[0086] It should be noted that the specific method of obtaining the Euclidean distance between the multidimensional feature vectors and the standard feature vectors of various types of Chinese medicinal materials is: taking the square root of the sum of the differences between the corresponding eigenvalues ​​in the multidimensional feature vectors and the standard feature vectors of various types of Chinese medicinal materials as the Euclidean distance between the multidimensional feature vectors and the standard feature vectors of various types of Chinese medicinal materials.

[0087] The smell qualification of each type of Chinese medicinal material is obtained by calculating the Euclidean distance between the multidimensional feature vector and the standard feature vector of each type of Chinese medicinal material.

[0088] It should be noted that the specific method of obtaining the odor qualification of various Chinese medicinal materials is: obtaining the odor qualification of various Chinese medicinal materials by dividing 1 by 1 and adding the Euclidean distance.

[0089] It should be further explained that the smaller the Euclidean distance, the closer the two vectors are and the higher the matching degree, which means the greater the odor qualification. Therefore, the Euclidean distance is inversely proportional to the odor qualification. Therefore, by dividing 1 by 1 plus the Euclidean distance, we can get the inverse proportional relationship between the Euclidean distance and the odor qualification. In addition, by adding 1 to the Euclidean distance is to avoid the denominator being 0, so that the fraction makes sense.

[0090] The color qualification and smell qualification of each type of Chinese medicinal materials are weightedly calculated and summed with the corresponding weights to obtain the quality qualification of each type of Chinese medicinal materials.

[0091] In a specific embodiment of the present invention, when calculating the quality qualification of various types of Chinese medicinal materials, the corresponding proportions of color qualification and smell qualification are equally important, so the corresponding proportions of color qualification and smell qualification are weighted 0.5 and 0.5 respectively.

[0092] The embodiment of the present invention adopts a multimodal data fusion algorithm, analyzes the color qualification through grayscale image analysis, and calculates the odor qualification based on the Euclidean distance of the odor component spectrum, thereby realizing digital evaluation of the color and odor of Chinese medicinal materials, improving the accuracy of the quality judgment of Chinese medicinal materials, ensuring the quality of the medicine, and at the same time guaranteeing the stability of the therapeutic effect.

[0093] The Chinese herbal medicine formula optimization module adjusts the initial dosage of various Chinese herbal medicines that the current patient needs to take based on the basic dosage and quality qualification of various Chinese herbal medicines.

[0094] In a specific embodiment of the present invention, the specific process of adjusting the initial dose of each type of Chinese medicinal material that the current patient needs to take is: comparing the quality qualification of each type of Chinese medicinal material with the quality qualification of the set reference; if the quality qualification of a certain type of Chinese medicinal material is greater than the quality qualification of the set reference, then this type of Chinese medicinal material is recorded as a high-quality medicinal material; if the quality qualification of a certain type of Chinese medicinal material is less than the quality qualification of the set reference, then this type of Chinese medicinal material is recorded as a low-quality medicinal material; if the quality qualification of a certain type of Chinese medicinal material is equal to the quality qualification of the set reference, then this type of Chinese medicinal material is recorded as a standard quality medicinal material.

[0095] If the Chinese medicinal materials that the patient currently needs to take are of standard quality, the basic dosage of this type of Chinese medicinal materials will remain unchanged.

[0096] If the Chinese medicinal materials that the current patient needs to take are high-quality medicinal materials, the difference between the quality qualification of this type of Chinese medicinal materials and the set reference quality qualification is obtained, and the product of the difference and the required adjustment dose corresponding to the unit quality qualification deviation of this type of Chinese medicinal materials stored in the database is used as the required reduced dose of this type of Chinese medicinal materials, and the difference between the basic dose and the required reduced dose is used as the initial dose of this type of Chinese medicinal materials.

[0097] It should be noted that the specific method for obtaining the required dose adjustment corresponding to the unit quality deviation is: 1) Data collection: Recording the actual dose adjustments made due to quality deviation during clinical use of Chinese medicinal materials to form a historical data set of "quality deviation-dose adjustment"; 2) Establishing a mathematical model: Using linear regression statistical methods, analyze the relationship between quality deviation and dose adjustment. For example, let the quality deviation be the independent variable x and the dose adjustment be the dependent variable y. Using the least squares method, fit y = kx + b (k is the dose adjustment coefficient for unit quality deviation, b is the intercept; if there is no quality deviation, the dose adjustment is 0 and b can be ignored). Determine the value of k through regression analysis, and thus determine the required dose adjustment corresponding to the unit quality deviation.

[0098] If the Chinese medicinal materials that the current patient needs to take are of low quality, the difference between the set reference quality qualification and the quality qualification of this type of Chinese medicinal materials is obtained, and the product of the difference and the required adjustment dose corresponding to the unit quality qualification deviation of this type of Chinese medicinal materials stored in the database is used as the required increased dose of this type of Chinese medicinal materials, and the sum of the basic dose and the required increased dose is used as the initial dose of this type of Chinese medicinal materials.

[0099] In summary, the initial dosage of various Chinese medicinal materials that the current patient needs to take is obtained.

[0100] In the embodiment of the present invention, the initial dose is adjusted based on the quality qualification. The dose of high-quality medicinal materials is reduced according to the deviation ratio, and the dose of low-quality medicinal materials is increased according to the deviation ratio, thereby avoiding the risk of insufficient efficacy or toxicity due to fluctuations in the medicinal material components.

[0101] The dynamic prescription adjustment module collects the real-time physiological parameters of the current patient at the end of the current medication course, and collects the quality information of various Chinese medicinal materials in the target Chinese medicine store after the end of the current medication course, and optimizes the dosage of various Chinese medicinal materials required for the current patient in the next medication course accordingly.

[0102] In a specific embodiment of the present invention, the specific process of optimizing the dosage of various Chinese medicinal materials required for the current patient in the next course of medication is: based on the pulse rate and tongue coating image of the current patient's real-time physiological parameters corresponding to the end of the current medication course, a fusion processing is performed to obtain the physiological health index of the current patient at the end of the current medication course.

[0103] It should be noted that the pulse rate of the current patient at the end of the current medication course is collected by a pulse diagnosis device worn at the end of the current medication course, and the tongue coating image is collected by a high-definition camera.

[0104] In a specific embodiment of the present invention, the specific process of obtaining the physiological health index corresponding to the current patient at the end of the current medication course is: performing pulse abnormality analysis based on the pulse rate corresponding to the current patient at the end of the current medication course, and obtaining the pulse abnormality degree corresponding to the current patient at the end of the current medication course.

[0105] See also Figure 3 As shown, it should be noted that the specific process of obtaining the pulse abnormality degree of the current patient at the end of the current medication course is: comparing the pulse rate of the current patient at the end of the current medication course with the pulse rate range of normal people under normal circumstances stored in the database; if the pulse rate of the current patient at the end of the current medication course is within the pulse rate range of normal people under normal circumstances, then the pulse abnormality degree of the current patient at the end of the current medication course is recorded as 0; if the pulse rate of the current patient at the end of the current medication course is not within the pulse rate range of normal people under normal circumstances, then the pulse abnormality degree of the current patient at the end of the current medication course is recorded as 1, thereby obtaining the pulse abnormality degree of the current patient at the end of the current medication course, which takes a value of 0 or 1.

[0106] In a specific embodiment of the present invention, the pulse rate of a normal person under normal circumstances is in the range of 60 to 100 beats per minute.

[0107] Based on the tongue coating picture corresponding to the end of the current medication course of the current patient, the tongue coating color abnormality analysis is performed to obtain the tongue coating color abnormality degree corresponding to the end of the current medication course of the current patient.

[0108] It should be noted that the specific method of obtaining the abnormality of the tongue coating color of the current patient at the end of the current medication course is: extracting the RGB values ​​from the tongue coating picture of the current patient at the end of the current medication course, and marking them as 、 and .

[0109] Extract the RGB values ​​of the tongue coating images of normal people under normal conditions from the database and mark them as 、 and .

[0110] Calculate the abnormality of the tongue coating color of the current patient at the end of the current medication course , .

[0111] It should be further explained that the design idea of ​​the above tongue coating color abnormality calculation formula is to use the RGB color model to convert the tongue coating color into a calculable value ( 、 and ), compared with the normal tongue coating RGB standard value ( 、 and ) comparison, quantify the degree of color abnormality of the current tongue coating, divide the difference of each color channel by 256, normalize the difference range to (0-1), eliminate the dimension difference of RGB channels, and calculate the distance between the current tongue coating and the normal tongue coating in the three-dimensional color space. The larger the distance, the higher the color abnormality.

[0112] A comprehensive analysis is performed on the abnormality of the pulse and the abnormality of the tongue coating color corresponding to the end of the current medication course of the current patient to obtain the physiological health index corresponding to the end of the current medication course of the current patient.

[0113] It should be noted that the specific method of obtaining the physiological health index corresponding to the current patient at the end of the current medication course is: weighted calculation of the abnormal pulse degree and abnormal tongue coating color degree corresponding to the current patient at the end of the current medication course with the corresponding weights and summing them up to obtain the physiological abnormality degree corresponding to the current patient at the end of the current medication course, and substituting the physiological abnormality degree as a power into The physiological health index of the current patient at the end of the current medication course is obtained from the exponential function with as the base.

[0114] It is also necessary to further explain that the physiological abnormality is brought into the The significance of the exponential function with a base of is to ensure that the degree of physiological abnormality is negatively correlated with the physiological health index, and at the same time ensure that the value of the physiological health index is always greater than 0.

[0115] In a specific embodiment of the present invention, when calculating the physiological abnormality corresponding to the current patient at the end of the current medication course, the corresponding proportions of pulse abnormality and tongue coating color abnormality are equally important, so the corresponding proportions of pulse abnormality and tongue coating color abnormality have weights of 0.5 and 0.5 respectively.

[0116] Based on the quality information of various Chinese medicinal materials in the target Chinese medicine hall after the end of the current medication course, the quality qualification of various Chinese medicinal materials after the end of the current medication course is analyzed in the same way as the analysis method of the quality qualification of various Chinese medicinal materials.

[0117] It should be noted that the method for collecting quality information of the various types of Chinese medicinal materials after the current medication course is the same as the quality information collection method mentioned above, and will not be repeated here.

[0118] Based on the physiological health index corresponding to the current patient at the end of the current medication course and the quality qualification of various Chinese medicinal materials after the end of the current medication course, an optimization analysis is performed to obtain the dosage of various Chinese medicinal materials required for the current patient in the next medication course.

[0119] It should be noted that the specific formula for obtaining the dosage of various Chinese medicinal materials required for the current patient in the next course of medication is:

[0120] ,in, Indicates the first dose of the drug that the current patient needs to take in the next course of treatment. Dosage of Chinese herbal medicines Indicates the initial dosage of various Chinese medicinal materials that the patient needs to take. and They represent the physiological health index of the current patient at the end of the current medication course and the physiological health index of the set reference. and Respectively represent the quality qualification of various Chinese medicinal materials after the current course of treatment and the quality qualification of the set reference, Indicates the required adjustment dose corresponding to the unit quality qualification deviation of various Chinese medicinal materials stored in the database. Indicates the number of various Chinese medicinal materials. .

[0121] The embodiment of the present invention dynamically adjusts prescriptions based on physiological parameters, comprehensively calculates the physiological health index through the abnormality of pulse rate and tongue coating color, and optimizes the dosage of the next course of treatment while considering the real-time quality of medicinal materials and patient efficacy feedback, forming a "monitoring-adjustment" closed loop, thereby improving the efficiency of the treatment course and realizing dynamic adjustment of the Chinese medicinal material formula.

[0122] It should be noted that the formulas described above, through the principle of dimensional consistency and mathematical standardization (e.g., normalization, dimensionless parameter conversion, or unified unit system), can translate physical quantities of different attributes into unitless standard values ​​or homogeneous, superimposable parameters. This eliminates the interference of different dimensions on the operational logic, ensuring that the formulas retain the distribution characteristics of the original data while maintaining mathematical rationality and adaptability to objective laws. The above are merely exemplary embodiments of the present invention and are not intended to limit the scope of the invention.

[0123] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they should all fall within the scope of protection of the present invention.

Claims

1. AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine, characterized by: include: The symptom similarity acquisition module extracts the disease types and symptoms that the current patient has entered into the target Chinese medicine clinic's AI diagnosis and treatment platform, and extracts the corresponding symptoms of each historical patient with the same disease in the AI ​​diagnosis and treatment platform's historical database. Based on this, the symptom vector similarity between the current patient and each historical patient is obtained; The basic dosage confirmation module extracts the TCM usage information of each historical patient and confirms the type of TCM and the basic dosage of each TCM required by the current patient based on the similarity of symptom vectors; The Chinese herbal medicine quality identification module activates the target Chinese herbal medicine store's automatic herbal medicine grabbing machine and uses it to collect quality information of various Chinese herbal medicines. It also uses a multimodal data fusion algorithm to obtain the quality qualification of various Chinese herbal medicines. The Chinese herbal medicine formula optimization module adjusts the initial dosage of various Chinese herbal medicines required by the patient based on the basic dosage and quality of each type of Chinese herbal medicine; The dynamic prescription adjustment module collects the real-time physiological parameters of the current patient at the end of the current medication course, and collects the quality information of various Chinese medicinal materials in the target Chinese medicine store after the end of the current medication course, and optimizes the dosage of various Chinese medicinal materials required for the current patient in the next medication course accordingly.

2. The AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine according to claim 1 is characterized by: The specific process of obtaining the similarity between the symptom vectors of the current patient and each historical patient is as follows: Summarize all historical patients with similar diseases in the AI ​​diagnosis and treatment platform's history database and all corresponding symptoms of current patients to form a symptom set; Generate the symptom vector corresponding to the disease of the current patient and the symptom vector corresponding to each historical patient by using one-hot encoding based on the symptom set; The cosine similarity calculation is performed based on the symptom vector corresponding to the disease of the current patient and the symptom vector corresponding to each historical patient to obtain the similarity between the symptom vector of the current patient and each historical patient.

3. The AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine according to claim 1 is characterized by: The specific process of confirming the type of Chinese medicinal materials that the current patient needs to take is: Compare the symptom vector similarity between the current patient and each historical patient with the symptom vector similarity of the set reference. If the symptom vector similarity between the current patient and a certain historical patient is greater than or equal to the symptom vector similarity of the set reference, then the historical patient is recorded as a highly similar patient, thereby screening out the highly similar patients corresponding to the current patient; Extract the types of Chinese herbal medicines taken by each highly similar patient and the initial dosage of each type of Chinese herbal medicine from the Chinese herbal medicine taking information of each historical patient; Count the number of times each type of Chinese herbal medicine appears, and use the ratio between the number of times each type of Chinese herbal medicine appears and the number of highly similar patients as the frequency of occurrence of each type of Chinese herbal medicine; The frequency of occurrence of each type of Chinese medicinal materials is compared with the frequency of occurrence of the set reference medicinal materials. If the frequency of occurrence of a certain type of Chinese medicinal materials is greater than or equal to the frequency of occurrence of the set reference medicinal materials, then this type of Chinese medicinal materials will be recorded as the target Chinese medicinal materials, thereby obtaining the target Chinese medicinal materials type, and using it as the type of Chinese medicinal materials that the current patient needs to take.

4. The AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine according to claim 3 is characterized by: The specific method for confirming the basic dosage of each type of Chinese medicinal materials that the current patient needs to take is: extracting the initial dosage of each type of Chinese medicinal materials that the current patient needs to take from the initial dosage of each type of Chinese medicinal materials taken by each highly similar patient, and calculating the average thereof, and using the calculation result as the basic dosage of each type of Chinese medicinal materials that the current patient needs to take.

5. The AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine according to claim 1 is characterized by: The specific process of using the multimodal data fusion algorithm to obtain the quality of various Chinese medicinal materials is as follows: Perform color processing based on the grayscale images in the quality information of various Chinese medicinal materials to obtain the color qualification of various Chinese medicinal materials; Based on the characteristic values ​​of the odor component spectrum in the quality information of various Chinese medicinal materials and the characteristic standard values ​​of the standard component spectrum of various Chinese medicinal materials stored in the database, the odor qualification of various Chinese medicinal materials is obtained; The color qualification and smell qualification of each type of Chinese medicinal materials are weightedly calculated and summed with the corresponding weights to obtain the quality qualification of each type of Chinese medicinal materials.

6. The AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine according to claim 5 is characterized by: The specific method for obtaining the color qualification of various Chinese medicinal materials is: Extracting the grayscale value corresponding to each grayscale area from the grayscale image of each type of Chinese medicinal material, and comparing it with the grayscale value interval corresponding to the standard conditions of each type of Chinese medicinal material stored in the database; If the grayscale value corresponding to a grayscale area of ​​a certain type of Chinese medicinal material is not within the grayscale value interval corresponding to its standard condition, then the grayscale area of ​​this type of Chinese medicinal material is recorded as an abnormal area, and the number of abnormal areas of each type of Chinese medicinal material is counted; Count the number of grayscale regions corresponding to each type of Chinese medicinal material from the grayscale images of each type of Chinese medicinal material, and record the ratio between the number of abnormal regions of each type of Chinese medicinal material and the number of grayscale regions as the abnormal region ratio of each type of Chinese medicinal material; The difference between the proportion of abnormal areas of the set reference and the proportion of abnormal areas of each type of Chinese medicinal materials is compared with the proportion of abnormal areas of the set reference, and the ratio result is used as the color qualification of each type of Chinese medicinal materials.

7. The AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine according to claim 5 is characterized by: The specific process of obtaining the odor qualification of various Chinese medicinal materials is as follows: Combining the characteristic values ​​in the odor component spectra of various Chinese medicinal materials into a multidimensional vector as the multidimensional characteristic vector of various Chinese medicinal materials; Based on the standard component spectra of various Chinese medicinal materials stored in the database, each characteristic standard value is combined into a multidimensional vector as the standard characteristic vector of each Chinese medicinal material; Obtain the Euclidean distance between the multidimensional feature vector and the standard feature vector of each type of Chinese medicinal material; The smell qualification of each type of Chinese medicinal material is obtained by calculating the Euclidean distance between the multidimensional feature vector and the standard feature vector of each type of Chinese medicinal material.

8. The AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine according to claim 1 is characterized by: The specific process of adjusting the initial dosage of various Chinese medicinal materials required by the current patient is as follows: Compare the quality qualification of each type of Chinese medicinal materials with the quality qualification of the set reference. If the quality qualification of a certain type of Chinese medicinal materials is greater than the quality qualification of the set reference, then this type of Chinese medicinal materials will be recorded as high-quality medicinal materials. If the quality qualification of a certain type of Chinese medicinal materials is less than the quality qualification of the set reference, then this type of Chinese medicinal materials will be recorded as low-quality medicinal materials. If the quality qualification of a certain type of Chinese medicinal materials is equal to the quality qualification of the set reference, then this type of Chinese medicinal materials will be recorded as standard quality medicinal materials. If the patient currently needs to take a certain type of Chinese herbal medicine of standard quality, the basic dosage of this type of Chinese herbal medicine will remain unchanged; If the Chinese herbal medicine that the patient needs to take is of high quality, the difference between the quality qualification of the Chinese herbal medicine and the set reference quality qualification is obtained, and the product of the difference and the required adjustment dose corresponding to the unit quality qualification deviation of the Chinese herbal medicine stored in the database is used as the required reduced dose of the Chinese herbal medicine, and the difference between the basic dose and the required reduced dose is used as the initial dose of the Chinese herbal medicine; If the Chinese herbal medicine that the patient needs to take is of low quality, the difference between the set reference quality qualification and the quality qualification of the Chinese herbal medicine of this type is obtained, and the product of the difference and the required adjustment dose corresponding to the unit quality qualification deviation of the Chinese herbal medicine of this type stored in the database is used as the required increased dose of the Chinese herbal medicine of this type, and the sum of the basic dose and the required increased dose is used as the initial dose of the Chinese herbal medicine of this type; In summary, the initial dosage of various Chinese medicinal materials that the current patient needs to take is obtained.

9. The AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine according to claim 5 is characterized by: The specific process of optimizing the dosage of various Chinese medicinal materials required for the current patient in the next course of medication is as follows: Based on the pulse rate and tongue coating image of the current patient's real-time physiological parameters corresponding to the end of the current medication course, a fusion process is performed to obtain the current patient's physiological health index corresponding to the end of the current medication course; Based on the quality information of various Chinese medicinal materials in the target Chinese medicine hall after the current course of treatment, the quality qualification of various Chinese medicinal materials after the current course of treatment is analyzed in the same way as the quality qualification of various Chinese medicinal materials; Based on the physiological health index corresponding to the current patient at the end of the current medication course and the quality qualification of various Chinese medicinal materials after the end of the current medication course, an optimization analysis is performed to obtain the dosage of various Chinese medicinal materials required for the current patient in the next medication course.

10. The AI-assisted diagnosis and treatment system for intelligent identification and formula optimization of traditional Chinese medicine according to claim 9 is characterized by: The specific process of obtaining the physiological health index corresponding to the current patient at the end of the current medication course is as follows: Perform pulse abnormality analysis based on the pulse rate of the current patient at the end of the current medication course to obtain the pulse abnormality degree of the current patient at the end of the current medication course; Based on the tongue coating picture corresponding to the end of the current medication course of the current patient, the abnormality of the tongue coating color is analyzed to obtain the abnormality degree of the tongue coating color corresponding to the end of the current medication course of the current patient; A comprehensive analysis is performed on the abnormality of the pulse and the abnormality of the tongue coating color corresponding to the end of the current medication course of the current patient to obtain the physiological health index corresponding to the end of the current medication course of the current patient.

Images