Ai tongue vein auxiliary system and method for treating ankylosing spondylitis in combination of traditional chinese and western medicine

Abstract

The application relates to the technical field of medical detection, in particular to an AI tongue pulse auxiliary system and method for treating ankylosing spondylitis by combining traditional Chinese medicine with western medicine, wherein the system comprises: a tongue pulse collection module that collects tongue image and inch, Guan and chi three-part pulse signals of a patient; a tongue pulse preprocessing module that pre-processes the tongue image to obtain first combined data; a western medicine data module that standardizes western medicine diagnosis and treatment data of the patient to obtain second combined data; an AI analysis module that inputs the two types of combined data based on a preset AI model and outputs a TCM syndrome type and a western medicine disease activity score of the patient; and a traditional Chinese medicine and western medicine decision fusion module that matches and outputs a corresponding combined traditional Chinese medicine and western medicine treatment scheme according to the syndrome type and the score. The application realizes objective quantification of tongue pulse characteristics and deep fusion of western medicine diagnosis and treatment data by AI technology, relieves the predicament of a shortage of high-quality TCM doctor resources and a long training cycle, and meets the diagnosis and treatment needs of primary medical institutions and a large number of ankylosing spondylitis patients.

Description

Technical Field

[0001] This invention relates to the field of medical testing technology, specifically to an AI-assisted tongue and pulse system and method for treating ankylosing spondylitis using a combination of traditional Chinese and Western medicine. Background Technology

[0002] Ankylosing spondylitis is a chronic, progressive inflammatory rheumatic immune disease that primarily affects the sacroiliac joints, spine, and peripheral joints. In severe cases, it can lead to spinal rigidity, deformity, and even loss of working ability, significantly impacting patients' quality of life. Currently, the clinical approach often employs an integrated traditional Chinese and Western medicine treatment model. This model fully leverages the advantages of Western medicine in anti-inflammatory and analgesic effects and disease progression control, combined with the characteristics of traditional Chinese medicine's syndrome differentiation and holistic approach, achieving a comprehensive therapeutic effect that addresses both the symptoms and the root cause. This approach has become one of the mainstream directions in the diagnosis and treatment of ankylosing spondylitis.

[0003] In the TCM diagnostic system, tongue diagnosis and pulse diagnosis are the core basis for syndrome differentiation. By observing the tongue appearance (tongue color, coating color, tongue shape, etc.) and palpating the pulse (pulse rate, pulse amplitude, pulse rhythm, etc.), it is possible to accurately determine whether the patient has common syndromes of ankylosing spondylitis such as "kidney deficiency and depletion of the Du meridian" and "spleen deficiency and dampness". This provides support for the formulation of personalized TCM treatment plans.

[0004] However, the traditional tongue and pulse diagnosis model relies heavily on the doctor's personal practical experience. It requires doctors to accumulate clinical experience over a long period of time to accurately grasp the relationship between tongue and pulse characteristics and syndrome types. Moreover, the diagnostic results are highly subjective, and the consistency of diagnoses among different doctors is poor. At the same time, there is a shortage of TCM doctors with rich experience in tongue and pulse diagnosis, and the training period is long, making it difficult to meet the diagnostic and treatment needs of primary medical institutions and large numbers of ankylosing spondylitis patients. This seriously limits the widespread promotion and popularization of the integrated TCM and Western medicine treatment model in the daily diagnosis of ankylosing spondylitis.

[0005] Therefore, this invention provides an AI-assisted tongue and pulse system and method for treating ankylosing spondylitis using a combination of traditional Chinese and Western medicine, in order to solve the above-mentioned problems. Summary of the Invention

[0006] To address the aforementioned issues, this invention provides an AI-assisted tongue and pulse diagnosis system and method for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine. This system enables the objective quantitative extraction of tongue and pulse characteristics and the deep integration of traditional Chinese and Western medicine diagnostic data. It accurately determines the TCM syndrome type and Western medicine disease activity of ankylosing spondylitis patients, automatically matches personalized integrated traditional Chinese and Western medicine treatment plans, meets the needs of primary care and large-scale treatment, and improves the standardization, precision, and efficiency of disease diagnosis and treatment.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows: an AI-assisted decision-making system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine, comprising: The tongue and pulse acquisition module is used to acquire tongue images and pulse signals at the cun, guan, and chi positions of patients with ankylosing spondylitis. The tongue and pulse preprocessing module is used to perform image preprocessing and feature extraction on the tongue image and to perform preprocessing and quantization feature extraction on the pulse signal to obtain the first combined data. The Western medicine data module is used to connect to the medical database, obtain Western medicine diagnosis and treatment data of patients with ankylosing spondylitis, and standardize the Western medicine diagnosis and treatment data to obtain the second combined data; The AI ​​analysis module is used to output the TCM syndrome type and Western medicine disease activity score of patients with ankylosing spondylitis based on the input of the first combined data and the second combined data, according to the preset AI analysis model. The integrated TCM and Western medicine decision-making module is used to match corresponding integrated TCM and Western medicine treatment plans based on TCM syndrome types and Western medicine disease activity scores, and output the corresponding integrated TCM and Western medicine treatment plans.

[0008] Furthermore, the Western medicine diagnostic data includes HLA-B27 index, CRP index, ESR index, and sacroiliac joint MRI score.

[0009] Furthermore, the TCM syndrome types include kidney deficiency and depletion of the Du meridian, spleen deficiency and dampness accumulation, damp-heat obstruction, cold-dampness obstruction, and blood stasis obstructing the collaterals.

[0010] Furthermore, the tongue vein preprocessing module's preprocessing procedure for tongue images includes: The first step is to use a Gaussian filtering algorithm to denoise the tongue image, with the standard deviation of the Gaussian filter set to 1.0-2.0. The second step is to use the U-Net network to segment the tongue region of the denoised tongue image; The third step is to perform color space conversion on the segmented tongue image and extract four core features: tongue color, tongue coating color, tongue shape, and tongue coating texture.

[0011] Furthermore, the tongue color characteristics include four quantitative parameters: light red, red, crimson, and purple; the coating color characteristics include four quantitative parameters: white, yellow, gray, and black.

[0012] Furthermore, the preprocessing process of the pulse signal in the tongue pulse preprocessing module includes: The first step is to perform baseline correction on the original pulse signal using a linear fitting algorithm; The second step is to use a filtering algorithm to filter the corrected pulse signal, with the filter cutoff frequency set to 10Hz-20Hz. The third step is to extract time-domain features from the filtered pulse signal to obtain eight quantitative features: pulse amplitude peak, pulse rate, pulse rate variation coefficient, pulse width, rise time, fall time, diastolic filling degree, and dicrotic wave depth, forming a pulse feature dataset.

[0013] Furthermore, in the tongue and pulse preprocessing module, the four core features of tongue appearance extracted—tongue color, coating color, tongue shape, and coating texture—are normalized to obtain a standardized tongue appearance feature dataset. Then, the standardized tongue appearance feature dataset and the pulse appearance feature dataset are aligned and fused according to a preset dimension. During the fusion process, a feature splicing method is used to sequentially combine the quantitative parameters of the tongue appearance features with the eight quantitative features of the pulse appearance features to form a feature matrix with unified dimensions. This feature matrix is ​​the first combined data.

[0014] Furthermore, the process by which the AI ​​analysis module outputs TCM syndrome types and Western medicine disease activity scores includes: The first step is to fuse the features of the first and second combined data to obtain the integrated Chinese and Western medicine feature dataset; The second step is to input the integrated TCM and Western medicine feature dataset into the preset AI analysis model; the preset AI analysis model is a CNN-LSTM fusion model; the preset AI analysis model includes a TCM syndrome recognition sub-model and a Western medicine disease assessment sub-model; The third step is that the TCM syndrome identification sub-model outputs the matching probability of each TCM syndrome based on the tongue and pulse features in the integrated TCM and Western medicine feature dataset and the correlation features of Western medicine diagnosis and treatment data, and selects the syndrome with the highest probability as the final TCM syndrome. The fourth step involves the Western medicine disease assessment sub-model calculating and outputting the patient's BASDAI score based on the Western medicine diagnosis and treatment data and tongue and pulse quantitative feature correlation data in the integrated Chinese and Western medicine feature dataset. This score is the Western medicine disease activity score.

[0015] Furthermore, in the TCM-Western medicine decision-making integration module, a three-dimensional knowledge graph of "TCM syndrome type - Western medicine disease activity score - treatment plan" is constructed based on TCM and Western medicine clinical guidelines. The knowledge graph stores standardized treatment plan templates corresponding to different TCM syndrome types and different BASDAI score ranges, and each standardized treatment plan template is associated with corresponding diagnostic and treatment basis. Then, the final TCM syndrome type and BASDAI score output by the AI ​​analysis module are received. First, the BASDAI score is matched to the corresponding score range. Using "TCM syndrome type + score range" as the joint search condition, the matching is performed in the three-dimensional knowledge graph, and the corresponding standardized treatment plan template is called.

[0016] Furthermore, the AI-assisted tongue and pulse therapy method for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine includes the following specific steps: Step 1: Multi-source data acquisition: The tongue image and pulse signals of the three positions of cun, guan, and chi of the ankylosing spondylitis patient are acquired through the tongue and pulse acquisition module. At the same time, the Western medicine data module is connected to the medical database to obtain the Western medicine diagnosis and treatment data of the ankylosing spondylitis patient. Step 2: Data Preprocessing and Fusion: The tongue and pulse preprocessing module is used to denoise, segment, and extract features from the acquired tongue images, and to perform baseline correction, filtering, and quantization feature extraction on the pulse signals. The processed data are then fused to obtain the first combined data. At the same time, the Western medicine data module is used to standardize the acquired Western medicine diagnosis and treatment data to obtain the second combined data. Step 3: AI Collaborative Analysis: Input the first and second combined data into the AI ​​analysis module to output the TCM syndrome type and BASDAI score; Step 4: Integrated Output of TCM and Western Medicine Treatment Plans: Based on the output TCM syndrome type and BASDAI score, the integrated TCM and Western medicine decision-making module generates treatment plans and related diagnostic and treatment evidence.

[0017] The above approach has the following beneficial effects: 1. This solution constructs a standardized tongue and pulse data quantification system and an AI intelligent analysis model. In tongue image processing, Gaussian filtering ensures precise noise reduction, and U-Net network tongue segmentation guarantees accurate extraction of four core features, including tongue color and tongue coating, transforming subjective visual descriptions into quantifiable parameters. Pulse image processing achieves objective quantification of pulse signals through baseline correction, 10Hz-20Hz filtering, and extraction of eight time-domain features. The two types of data are normalized and feature-concatenated to form a unified feature matrix, which is then deeply integrated with standardized Western medicine data such as HLA-B27 and CRP. After inputting into a CNN-LSTM fusion model, TCM syndrome identification achieves accurate judgment through probability ranking, and BASDAI scoring achieves quantitative assessment of the condition. The system can complete professional-level diagnostic analysis without requiring physicians' long-term experience accumulation (especially in TCM's "observation, auscultation, inquiry, and palpation"), reducing reliance on high-quality TCM resources. It can be quickly deployed in primary healthcare institutions, effectively covering a large patient population and improving the standardization and accuracy of integrated TCM and Western medicine diagnosis. 2. This solution constructs a three-dimensional knowledge graph of "Traditional Chinese Medicine (TCM) syndrome type - Western medicine disease activity score - treatment plan" to achieve personalized matching and standardized output of integrated TCM and Western medicine treatment plans for ankylosing spondylitis, thereby improving clinical diagnosis and treatment efficiency and promotional value. The knowledge graph is built based on authoritative clinical guidelines, storing standardized treatment templates and diagnostic criteria corresponding to different syndrome types and BASDAI score ranges. Through joint retrieval of "syndrome type + score range," suitable solutions can be quickly retrieved, reducing the arbitrariness of traditional treatment plan formulation and providing physicians with clear decision support, thus shortening the treatment cycle. This is particularly suitable for scenarios with high outpatient volume and limited resources.

[0018] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of an embodiment of the AI ​​tongue and pulse assist system and method for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine, as described in this invention. Figure 2 This is a schematic diagram of an embodiment of the AI ​​tongue and pulse assist system and method for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine. Detailed Implementation

[0020] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0022] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0023] The following detailed description illustrates the specific implementation method: Example 1:

[0024] like Figure 1 As shown, an AI-assisted tongue and pulse diagnosis system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine includes a tongue and pulse acquisition module, a tongue and pulse preprocessing module, a Western medicine data module, an AI analysis module, and a fusion module for integrated traditional Chinese and Western medicine decision-making, as detailed below: The tongue and pulse acquisition module is used to acquire tongue images and pulse signals at the cun, guan, and chi positions of patients with ankylosing spondylitis, providing high-quality raw data for subsequent analysis. Tongue image acquisition utilizes a high-definition digital image acquisition device with a resolution of at least 3000×4000 pixels. During acquisition, the ambient light intensity is controlled at 500-800 lux using a supplemental lighting device to ensure the tongue image is free of reflection and has true color. The acquisition angle is perpendicular to the center of the tongue surface to avoid distortion. Pulse acquisition employs a three-channel flexible pulse sensor, which is respectively attached to the radial artery at the cun, guan, and chi positions of the patient. The sampling frequency is set to 200Hz, and pulse signals are continuously acquired for 30 seconds. Basic patient information (such as age, gender, and disease course) and the acquisition timestamp are recorded simultaneously to ensure data traceability. For example, for a 35-year-old male patient with ankylosing spondylitis, the tongue and pulse acquisition module can simultaneously acquire a clear image of his tongue and continuous 30-second pulse fluctuation signals at the cun, guan, and chi positions.

[0025] The tongue and pulse preprocessing module is used to perform image preprocessing and feature extraction on tongue images, as well as preprocessing and quantification feature extraction on pulse signals, transforming the raw data into structured features that can be used for AI analysis, and finally obtaining the first combined data.

[0026] Specifically, the tongue vein preprocessing module performs the following preprocessing steps on the tongue image: The first step is to use a Gaussian filtering algorithm to denoise the tongue image. The standard deviation of the Gaussian filter is set to 1.0-2.0, which effectively filters out environmental noise and image sensor noise, while preserving key features such as tongue edges and tongue coating texture. The second step involves using a pre-trained U-Net network to segment the tongue body region of the denoised tongue image. This U-Net network is trained on 10,000 labeled tongue images (covering tongue images of ankylosing spondylitis patients of different ages and syndromes as well as healthy individuals). It can automatically remove non-tongue areas such as oral mucosa, teeth, and background, and output only the region of interest (ROI) containing the tongue body. The third step involves converting the segmented tongue image from RGB to HSV color space. Then, using algorithms such as color histogram analysis and morphological feature extraction, four core features are extracted: tongue color, tongue coating color, tongue shape, and tongue coating texture. Tongue color characteristics include four quantitative parameters: light red (quantitative parameter range: H∈0-15°, S∈20%-40%), red (H∈0-15°, S∈40%-60%), crimson (H∈0-15°, S∈60%-80%), and purple (H∈280-360°, S∈30%-50%). The moss color characteristics include four quantitative parameters: white (H∈0-30°, S∈10%-30%), yellow (H∈30-60°, S∈40%-60%), gray (H∈0-60°, S∈10%-20%), and black (H∈0-60°, S∈0-10%). In addition, tongue shape characteristics are quantified by parameters such as the length-to-width ratio of the tongue body, the smoothness of the edges, and the presence of cracks / tooth marks; tongue coating characteristics are quantified by parameters such as the thickness of the tongue coating, the degree of moisture, and the uniformity of its distribution.

[0027] Specifically, the preprocessing process of the tongue pulse preprocessing module for pulse signals is as follows: The first step is to perform baseline correction on the original pulse signal using a linear fitting algorithm; The second step is to use a filtering algorithm to filter the corrected pulse signal, with the filter cutoff frequency set to 10Hz-20Hz. The third step is to extract time-domain features from the filtered pulse signal to obtain eight quantitative features: pulse amplitude peak, pulse rate, pulse rate variation coefficient, pulse width, rise time, fall time, diastolic filling degree, and dicrotic wave depth, forming a pulse feature dataset.

[0028] In the tongue and pulse preprocessing module, after preprocessing the tongue image and pulse signal, the extracted four core tongue features—tongue color, coating color, tongue shape, and coating texture—are first subjected to min-max normalization (normalization interval: [0,1]) to obtain a standardized tongue feature dataset (12 dimensions, including 4-dimensional tongue color, 4-dimensional coating color, 2-dimensional tongue shape, and 2-dimensional coating texture). Then, the standardized tongue feature dataset and the pulse feature dataset are aligned and fused according to the "patient ID-collection timestamp" dimension. During the fusion process, a feature splicing method is used to sequentially combine the 12 quantization parameters of the tongue features with the 8 quantization parameters of the pulse features to form a feature matrix with a dimension of 20. This feature matrix is ​​the first combined data.

[0029] The Western medicine data module is used to connect to hospital electronic medical record systems, laboratory test databases, and other medical databases. It acquires Western medicine diagnostic and treatment data of ankylosing spondylitis patients via the HL7 interface protocol and performs standardized processing to obtain the second-linked data. This Western medicine diagnostic and treatment data includes core diagnostic indicators: HLA-B27 markers (positive / negative), CRP (C-reactive protein) markers (unit: mg / L, normal reference range: 0-10 mg / L), ESR (erythrocyte sedimentation rate) markers (unit: mm / h, normal reference range for men: 0-15 mm / h, normal reference range for women: 0-20 mm / h), and sacroiliac joint MRI scores (using the ASspiMRI-a scoring standard, range: 0-72 points). The standardization process includes: filling missing data with the K-nearest neighbor algorithm, correcting outlier data (data exceeding 3 times the standard deviation) with the Winsorize method, converting categorical data (such as HLA-B27 positive / negative) into one-hot codes (positive: [1,0], negative: [0,1]), and normalizing continuous data (such as CRP, ESR, MRI scores) to the [0,1] interval, ultimately forming a second-binding data with a dimension of 6 (HLA-B27: 2-dimensional, CRP: 1-dimensional, ESR: 1-dimensional, MRI score: 1-dimensional, disease course: 1-dimensional).

[0030] The AI ​​analysis module is used to output the TCM syndrome type and Western medicine disease activity score of patients with ankylosing spondylitis based on the input of the first and second combined data, using a preset AI analysis model. The TCM syndrome types include kidney deficiency and depletion of the Du meridian, spleen deficiency and dampness accumulation, damp-heat obstruction, cold-dampness obstruction, and blood stasis obstructing the collaterals.

[0031] Specifically, the process by which the AI ​​analysis module outputs TCM syndrome types and Western medicine disease activity scores is as follows: The first step is to fuse the features of the first combined data (20 dimensions) and the second combined data (6 dimensions) and use the feature splicing method to form a 26-dimensional integrated Chinese and Western medicine feature dataset. The second step is to input the integrated Chinese and Western medicine feature dataset into the preset AI analysis model (CNN-LSTM fusion model). This model includes a Chinese medicine syndrome recognition sub-model and a Western medicine disease assessment sub-model, which can perform two types of analysis tasks simultaneously. The third step involves the TCM syndrome identification sub-model, which uses the quantitative features of tongue and pulse (such as crimson tongue color, increased pulse rate, and increased pulse rate variation coefficient) in the integrated TCM and Western medicine feature dataset and the correlation features of Western medicine diagnosis and treatment data (such as elevated CRP and high MRI score) to output the matching probability of 5 TCM syndrome types through the Softmax function. For example, it outputs "Kidney Deficiency and Governing Vessel Deficiency Syndrome: 0.85, Spleen Deficiency and Dampness Syndrome: 0.08, Damp-Heat Obstruction Syndrome: 0.03, Cold-Dampness Obstruction Syndrome: 0.02, Blood Stasis and Collaterals Obstruction Syndrome: 0.02". The TCM syndrome type with the highest probability, "Kidney Deficiency and Governing Vessel Deficiency Syndrome", is selected as the final TCM syndrome type. The fourth step involves the Western medicine disease assessment sub-model, which uses Western medicine diagnosis and treatment data (such as CRP=25mg / L, ESR=30mm / h, MRI score=45 points) and tongue and pulse quantitative feature correlation data (such as pulse amplitude peak=1.1V, diastolic filling=0.3) from the integrated Chinese and Western medicine feature dataset to output the patient's BASDAI score through linear regression. For example, it outputs "BASDAI score: 5.8 points", which is the Western medicine disease activity score.

[0032] The TCM-Western medicine decision fusion module is used to match corresponding integrated TCM and Western medicine treatment plans based on TCM syndrome types and Western medicine disease activity scores, and output the corresponding integrated TCM and Western medicine treatment plans. Specifically, in the TCM-Western medicine decision fusion module, a three-dimensional knowledge graph of "TCM syndrome type - Western medicine disease activity score - treatment plan" is constructed based on TCM and Western medicine clinical guidelines. The knowledge graph stores standardized treatment plan templates corresponding to different TCM syndrome types and different BASDAI score intervals, and each standardized treatment plan template is associated with corresponding diagnostic and treatment basis. Then, it receives the final TCM syndrome type and BASDAI score output by the AI ​​analysis module, first matches the BASDAI score to the corresponding score interval, and then performs matching in the three-dimensional knowledge graph using "TCM syndrome type + score interval" as the joint search condition, and calls the corresponding standardized treatment plan template.

[0033] Example 2:

[0034] As attached Figure 2 As shown, the AI-assisted tongue and pulse therapy method for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine includes the following specific steps: Step 1: Multi-source diagnostic and treatment data acquisition: The tongue image and pulse signals of the three positions of cun, guan, and chi of the ankylosing spondylitis patient are collected through the tongue and pulse acquisition module. At the same time, the Western medicine data module is connected to the medical database to obtain the patient's Western medicine diagnostic and treatment data such as HLA-B27 index, CRP index, ESR index and sacroiliac joint MRI score. Step 2, Data Preprocessing and Fusion: The tongue and pulse preprocessing module is used to perform denoising, tongue segmentation, and feature extraction on the acquired tongue images, and baseline correction, filtering, and quantization feature extraction on the pulse signals. The two types of processed feature data are fused to obtain the first combined data; the Western medicine data module is used to standardize the acquired Western medicine diagnosis and treatment data to obtain the second combined data. Step 3: AI Collaborative Diagnosis and Treatment Analysis: Input the first and second combined data into the CNN-LSTM fusion model of the AI ​​analysis module. The TCM syndrome identification sub-model outputs the final TCM syndrome with the highest probability, and the Western medicine disease assessment sub-model calculates and outputs the corresponding BASDAI score (i.e., Western medicine disease activity score). Step 4: Integrated Output of TCM and Western Medicine Treatment Plans: The TCM and Western medicine decision-making integration module uses the output TCM syndrome type and BASDAI score as search criteria to match the corresponding standardized treatment plan template in the three-dimensional knowledge graph of the TCM and Western medicine decision-making integration module. This generates a personalized TCM and Western medicine integrated treatment plan that includes Western medicine medication and TCM intervention plans. After rationality verification, the treatment plan and related diagnostic and treatment evidence are output in the form of a structured report.

[0035] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. An AI-assisted tongue and pulse therapy system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine, characterized in that: include: The tongue and pulse acquisition module is used to acquire tongue images and pulse signals at the cun, guan, and chi positions of patients with ankylosing spondylitis. The tongue and pulse preprocessing module is used to perform image preprocessing and feature extraction on the tongue image and to perform preprocessing and quantization feature extraction on the pulse signal to obtain the first combined data. The Western medicine data module is used to connect to the medical database, obtain Western medicine diagnosis and treatment data of patients with ankylosing spondylitis, and standardize the Western medicine diagnosis and treatment data to obtain the second combined data. The AI ​​analysis module is used to output the TCM syndrome type and Western medicine disease activity score of patients with ankylosing spondylitis based on the input of the first combined data and the second combined data, according to the preset AI analysis model. The integrated TCM and Western medicine decision-making module is used to match corresponding integrated TCM and Western medicine treatment plans based on TCM syndrome types and Western medicine disease activity scores, and output the corresponding integrated TCM and Western medicine treatment plans.

2. The AI ​​tongue and pulse assist system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine as described in claim 1, characterized in that, Western medicine diagnostic data includes HLA-B27 index, CRP index, ESR index and sacroiliac joint MRI score.

3. The AI ​​tongue and pulse assist system and method for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine as described in claim 2, characterized in that, Traditional Chinese medicine syndromes include kidney deficiency and depletion of the Du meridian, spleen deficiency and dampness accumulation, damp-heat obstruction, cold-dampness obstruction, and blood stasis obstructing the collaterals.

4. The AI ​​tongue and pulse assist system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine as described in claim 3, characterized in that, The tongue vein preprocessing module performs the following preprocessing steps on the tongue image: The first step is to use a Gaussian filtering algorithm to denoise the tongue image, with the standard deviation of the Gaussian filter set to 1.0-2.

0. The second step is to use the U-Net network to segment the tongue region of the denoised tongue image; The third step is to perform color space conversion on the segmented tongue image and extract four core features: tongue color, tongue coating color, tongue shape, and tongue coating texture.

5. The AI ​​tongue and pulse assist system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine as described in claim 4, characterized in that, Tongue color characteristics include four quantitative parameters: light red, red, crimson, and purple; tongue coating characteristics include four quantitative parameters: white, yellow, gray, and black.

6. The AI ​​tongue and pulse assist system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine as described in claim 5, characterized in that, The preprocessing process of the tongue pulse preprocessing module for pulse signals includes: The first step is to perform baseline correction on the original pulse signal using a linear fitting algorithm; The second step is to use a filtering algorithm to filter the corrected pulse signal, with the filter cutoff frequency set to 10Hz-20Hz. The third step is to extract time-domain features from the filtered pulse signal to obtain eight quantitative features: pulse amplitude peak, pulse rate, pulse rate variation coefficient, pulse width, rise time, fall time, diastolic filling degree, and dicrotic wave depth, forming a pulse feature dataset.

7. The AI ​​tongue and pulse assist system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine as described in claim 6, characterized in that, In the tongue and pulse preprocessing module, the four core features of tongue appearance—tongue color, coating color, tongue shape, and coating texture—are normalized to obtain a standardized tongue appearance feature dataset. Then, the standardized tongue appearance feature dataset and the pulse appearance feature dataset are aligned and fused according to a preset dimension. During the fusion process, a feature splicing method is used to sequentially combine the quantitative parameters of the tongue appearance features with the eight quantitative features of the pulse appearance features to form a feature matrix with unified dimensions. This feature matrix is ​​the first combined data.

8. The AI ​​tongue and pulse assist system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine as described in claim 7, characterized in that, The process by which the AI ​​analysis module outputs TCM syndrome types and Western medicine disease activity scores includes: The first step is to fuse the features of the first and second combined data to obtain the integrated Chinese and Western medicine feature dataset; The second step is to input the integrated TCM and Western medicine feature dataset into the preset AI analysis model; the preset AI analysis model is a CNN-LSTM fusion model; the preset AI analysis model includes a TCM syndrome recognition sub-model and a Western medicine disease assessment sub-model; The third step is that the TCM syndrome identification sub-model outputs the matching probability of each TCM syndrome based on the tongue and pulse features in the integrated TCM and Western medicine feature dataset and the correlation features of Western medicine diagnosis and treatment data, and selects the syndrome with the highest probability as the final TCM syndrome. The fourth step involves the Western medicine disease assessment sub-model calculating and outputting the patient's BASDAI score based on the Western medicine diagnosis and treatment data and tongue and pulse quantitative feature correlation data in the integrated Chinese and Western medicine feature dataset. This score is the Western medicine disease activity score.

9. The AI ​​tongue and pulse assist system for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine as described in claim 8, characterized in that, In the TCM-Western medicine decision-making integration module, a three-dimensional knowledge graph of "TCM syndrome type - Western medicine disease activity score - treatment plan" is constructed based on TCM and Western medicine clinical guidelines. The knowledge graph stores standardized treatment plan templates corresponding to different TCM syndrome types and different BASDAI score ranges, and each standardized treatment plan template is associated with corresponding diagnostic and treatment basis. Then, the final TCM syndrome type and BASDAI score output by the AI ​​analysis module are received. First, the BASDAI score is matched to the corresponding score range. Using "TCM syndrome type + score range" as the joint search condition, the matching is performed in the three-dimensional knowledge graph, and the corresponding standardized treatment plan template is called.

10. An AI-assisted tongue and pulse therapy method for treating ankylosing spondylitis using integrated traditional Chinese and Western medicine, as described in claim 1, characterized in that... The specific steps include: Step 1, Multi-source data acquisition: The tongue image and pulse signals of the three positions of cun, guan, and chi of the ankylosing spondylitis patient are collected through the tongue and pulse acquisition module. At the same time, the Western medicine data module is connected to the medical database to obtain the Western medicine diagnosis and treatment data of the ankylosing spondylitis patient. Step 2, Data Preprocessing and Fusion: The tongue and pulse preprocessing module is used to denoise, segment, and extract features from the acquired tongue images, and to perform baseline correction, filtering, and quantization feature extraction on the pulse signals. The processed data are then fused to obtain the first combined data. At the same time, the Western medicine data module is used to standardize the acquired Western medicine diagnosis and treatment data to obtain the second combined data. Step 3, AI Collaborative Analysis: Input the first and second combined data into the AI ​​analysis module, and output the TCM syndrome type and BASDAI score; Step 4, Integrated Output of TCM and Western Medicine Treatment Plans: The TCM and Western medicine decision-making integration module generates treatment plans and related diagnostic and treatment evidence based on the output TCM syndrome types and BASDAI scores.

Images