An auxiliary diagnosis and treatment system based on a large Chinese medicine model

Abstract

The present invention belongs to the technical field of traditional Chinese medicine assisted diagnosis and treatment, and specifically discloses an assisted diagnosis and treatment system based on a traditional Chinese medicine large model. By receiving tongue image uploaded by patients at the same time period regularly, tongue image features of different tongue body regions are extracted. Thus, by using the tongue image features under time series, tongue image change curves are generated, and the syndrome type evolution path is output by using the tongue image - syndrome type model. Then, the clinical indicators of the patients in the same time series are associated with the syndrome type evolution path to form a syndrome type - disease association matrix. Based on this, a prescription compatibility scheme is recommended to ensure that the treatment plan is not only based on the syndrome type, but also fully considers the specific pathological mechanism and clinical manifestations, avoiding the disconnection between the treatment plan and the actual condition, realizing more accurate and personalized traditional Chinese medicine assisted diagnosis and treatment. In addition, automated data processing and analysis reduce manual intervention, improve the diagnosis and treatment efficiency, and save medical resources.

Description

Technical Field

[0001] The present invention belongs to the technical field of traditional Chinese medicine (TCM) assisted diagnosis and treatment, and specifically discloses an assisted diagnosis and treatment system based on a large TCM model. Background Art

[0002] In modern society, with significant changes in lifestyle, factors such as sedentary behavior, unreasonable diet structure, and high stress have led to the increasing prevalence of chronic diseases and sub - health states. This has promoted the growing demand for TCM assisted diagnosis and treatment. Because TCM, with its holistic concept and syndrome differentiation and treatment methods, can provide personalized comprehensive evaluation and treatment plans, especially showing unique advantages in chronic disease management and clinical diagnosis.

[0003] In the TCM diagnosis and treatment system, comprehensively evaluating the patient's health status through the four diagnostic methods of inspection, auscultation and olfaction, inquiry, and palpation is a key step in syndrome differentiation and treatment. The correlation analysis between the results of inquiry and syndrome types provides an important basis for subsequent formula compatibility. Along with the development of information technology, modern technologies such as artificial intelligence, big data, and cloud computing have gradually been integrated into the field of TCM. By constructing a large syndrome type model based on big data, the diagnosis and treatment efficiency has been significantly improved.

[0004] Under the big data - based syndrome type model, there are existing technical solutions for TCM assisted diagnosis and treatment. For example, the system proposed in the Chinese invention patent with the publication number CN119170207A uses a four - diagnostic instrument to collect the pulse and tongue image of the patient, and inputs them into a preset TCM syndrome differentiation model to obtain the syndrome type of the patient, thereby generating corresponding prescription information. Through the interaction between the intelligent terminal and the four - diagnostic instrument, this system assists doctors in collecting four - diagnostic information and performing syndrome differentiation and prescribing, improving the diagnosis and treatment efficiency.

[0005] Another example is the invention patent with the publication number CN112992344A, which proposes a TCM intelligent assisted diagnosis and treatment system. By collecting the symptom information of the patient (such as tongue coating and pulse condition) through an inquiry form, converting it into the corresponding syndrome type, selecting the corresponding prescription from the prescription library, and then determining the final prescription according to the drug addition and subtraction rules, it effectively improves the efficiency of TCM diagnosis and treatment.

[0006] Although the above two solutions have significantly improved the diagnosis and treatment efficiency by using the TCM models of syndrome types and symptoms, they have not fully considered the complex relationship between syndrome types and pathological mechanisms. TCM syndrome types are not only a high - level generalization of disease symptoms, but also a profound revelation of the internal causes and development trends of diseases. Each syndrome type reflects a specific pathological mechanism, and this mechanism determines that there is a complex synergistic relationship between different syndrome types and pathological states. Under this complex synergistic relationship, relying solely on syndrome types for formula compatibility may ignore the detailed pathological information of the patient, resulting in the lack of necessary clinical pathological support for the treatment plan, easily causing the disconnection between the treatment plan and the actual condition, and affecting the final treatment effect. Summary of the Invention

[0007] In view of this, the present invention aims to propose an auxiliary diagnosis and treatment system based on a traditional Chinese medicine large model. In the traditional Chinese medicine diagnosis and treatment system, tongue image analysis, as the core link of the four diagnostic methods, is a key indicator reflecting syndrome types. By focusing on the tongue image-syndrome type model and taking into account the patient's clinical indicators, a more comprehensive syndrome type-clinical indicator correlation model is established, so as to achieve more accurate and personalized traditional Chinese medicine assisted diagnosis.

[0008] The object of the present invention can be achieved by the following technical solutions: An auxiliary diagnosis and treatment system based on a traditional Chinese medicine large model, comprising: A tongue image data acquisition module: receiving tongue images uploaded by patients at the same regular time period, and triggering tongue image enhancement processing based on an image quality detection feedback mechanism for tongue body clarity and color uniformity.

[0009] A tongue image feature extraction module: dividing the enhanced tongue image into three tongue body regions of the tip of the tongue, the middle of the tongue, and the root of the tongue, and respectively extracting the tongue image features of each tongue body region, specifically the HSV value of the tongue color, the pixel ratio of the tongue coating thickness, and the crack density.

[0010] A tongue image change analysis module: storing the tongue image features of regularly uploaded tongue images according to the time series, and generating a tongue image change curve including the corresponding tongue color, tongue coating thickness, and cracks of each tongue body region.

[0011] A traditional Chinese medicine large model prediction module: importing the tongue image change curve into the constructed tongue image-syndrome type model to analyze the syndrome type under the time series, and outputting the syndrome type evolution path.

[0012] A diagnosis and treatment correlation decision module: correlating the patient's clinical indicators under the same time series with the syndrome type evolution path to form a syndrome type-disease correlation matrix, and recommending a prescription compatibility plan accordingly.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The present invention imports the tongue image change curve generated by the corresponding tongue image features of the patient's tongue image under the time series into the tongue image-syndrome type model to output the syndrome type evolution path, and combines it with the patient's clinical indicators to construct a syndrome type-disease correlation matrix, and recommends a personalized prescription compatibility plan accordingly, ensuring that the treatment plan is not only based on the syndrome type, but also fully considers the specific pathological mechanism and clinical manifestations, avoiding the disconnection between the treatment plan and the actual condition, and realizing more accurate and personalized traditional Chinese medicine assisted diagnosis and treatment.

[0014] 2. Before extracting features from the tongue images uploaded by patients, the present invention triggers image enhancement processing based on an image quality detection feedback mechanism for tongue body clarity and color uniformity, which can effectively improve low-quality images and is beneficial to improving the accuracy of subsequent tongue image feature extraction. In addition, this mechanism ensures that all tongue images for analysis have a consistent quality standard, improving the reliability and comparability of time series analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0016] Figure 1 It is a schematic diagram of the module composition and data flow in the system of the present invention.

[0017] Figure 2 It is a flowchart of the implementation for triggering tongue image enhancement processing in the present invention.

[0018] Figure 3 It is a schematic diagram of the syndrome type evolution path in the present invention. Specific implementation manners

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0020] The present invention proposes an auxiliary diagnosis and treatment system based on a traditional Chinese medicine large model, including a tongue image data acquisition module, a tongue image feature extraction module, a tongue image change analysis module, a traditional Chinese medicine large model prediction module, and a diagnosis and treatment correlation decision-making module.

[0021] Please refer to Figure 1 As shown, where the modules are connected end to end, and the data flow relationship between the modules is as follows: The tongue image data acquisition module, as the starting point of the system, provides raw data for the subsequent modules.

[0022] The tongue image feature extraction module processes the tongue image provided by the tongue image data acquisition module and transfers the extracted features to the tongue image change analysis module.

[0023] The tongue image change analysis module analyzes the change of the tongue image over time and transfers the results to the traditional Chinese medicine large model prediction module.

[0024] The traditional Chinese medicine large model prediction module receives and processes the data provided by the tongue image change analysis module, determines the syndrome type evolution path of the patient, and transfers this information to the diagnosis and treatment correlation decision-making module.

[0025] The diagnosis and treatment correlation decision-making module receives the results of the traditional Chinese medicine large model prediction module, integrates the clinical indicators of the patient, and generates treatment suggestions after comprehensive analysis.

[0026] Specifically, the tongue image data acquisition module receives the tongue images uploaded by the patient at the same time period regularly, and triggers the tongue image enhancement process based on the image quality detection feedback mechanism of tongue body clarity and color uniformity.

[0027] It should be noted that the above-mentioned regular data acquisition refers to the acquisition of tongue images at fixed intervals and cycles. This interval can be daily or other time frequencies set according to clinical needs, and the cycle can be one month, two months or longer. Through this regular data acquisition, the changes of the patient's tongue image over time can be effectively monitored, providing continuous and reliable data support for the subsequent generation of the tongue image change curve.

[0028] The same time period means that the patient selects the same time period when uploading the tongue image every day. For example, upload the tongue image between 8:00 and 10:00 every morning. Since the physiological state of the human body may change at different time periods of the day, in order to reduce the impact of this daily fluctuation on the data, selecting to acquire the tongue image at the same time period can ensure the consistency and stability of the data, thereby improving the reliability of the analysis results.

[0029] "Regular" and "the same time period" work together to ensure the regularity and consistency of the acquisition of tongue image data, thereby providing reliable basic data for the TCM auxiliary diagnosis and treatment system.

[0030] In the preferred implementation of the above solution, see Figure 2 As shown, the tongue image enhancement process triggered by the image quality detection feedback mechanism based on tongue body clarity and color uniformity is as follows: perform noise preprocessing operations on the received tongue image.

[0031] The above-mentioned noise preprocessing operation is specifically as follows in one example: perform smoothing processing on the image to remove high-frequency noise and avoid mistaking the noise for an edge. Through noise preprocessing, noise interference can be reduced and the accuracy of edge detection can be improved.

[0032] Use the edge detection algorithm to extract the tongue body edge pixels from the preprocessed image.

[0033] The purpose of extracting the tongue body edge pixels through edge detection above is that the tongue body edge pixels are the dividing line between the tongue body and the background. This operation aims to accurately identify the tongue body contour, which can reflect the edge sharpness and edge continuity, and the edge sharpness and edge continuity are the core indicators for measuring the clarity of the image, that is, the clarity of the tongue body. The edge sharpness reflects the clarity of the image details, while the edge continuity reflects the integrity of the tongue body contour in the image. The two are used in combination to comprehensively evaluate the clarity of the tongue body.

[0034] Evaluate the edge sharpness and edge continuity respectively based on the tongue body edge pixels.

[0035] As an implementation of the above solution, the edge sharpness is evaluated by analyzing the change rate of the pixels at the edge of the tongue body to assess its sharpness. High sharpness means clear edges and distinct boundaries, which helps to accurately extract the tongue body features.

[0036] In a specific example, the edge sharpness evaluation can be carried out by calculating the average value or the maximum value of the gradient amplitude of the edge pixels as a quantitative index of the edge sharpness.

[0037] As another implementation of the above solution, the edge continuity is to evaluate the coherence of the pixels at the edge of the tongue body, that is, whether the edge is complete and without breaks. Good edge continuity indicates that the tongue body structure in the image is complete and can more truly reflect the actual state of the tongue body.

[0038] In a specific example, the edge continuity evaluation can be carried out by calculating the ratio of the length of the complete edge to the total perimeter of the tongue body as a quantitative index of the edge continuity.

[0039] It should be noted that both the edge sharpness and the edge continuity evaluated above are dimensionless data.

[0040] Define the tongue body clarity as the product of the edge sharpness and the edge continuity. Under this definition, the tongue body clarity of the tongue image is obtained based on the evaluated edge sharpness and edge continuity.

[0041] It should be understood that the above definition of the tongue body clarity is based on a comprehensive consideration of the edge sharpness and the edge continuity. Specifically, the edge sharpness focuses on the clarity of the edge, reflecting the sharpness of the details of the tongue body contour boundary in the image; while the edge continuity focuses on the integrity of the edge line, that is, whether the edge is coherent and without breaks. A single index is difficult to comprehensively describe the tongue body clarity, because the usability of an image with high sharpness but low continuity will be greatly reduced in practical applications. Therefore, combining these two indexes can more comprehensively reflect the overall quality of the tongue image.

[0042] In the process of combining the two, the method of multiplying the edge sharpness and the edge continuity is adopted because the multiplication operation can effectively amplify the combined influence of the two indexes. If one of the indexes is low, for example, the edge continuity is very low, even if the other index, such as the edge sharpness, is very high, the final tongue body clarity will be significantly reduced. This mechanism ensures that the tongue body clarity is high only when the edge is both clear and continuous.

[0043] Adopt image segmentation technology to separate the tongue body area from the background and generate a corresponding binary mask, ensuring that the pixel values within the tongue body area are marked as 1, while the background area is marked as 0.

[0044] It should be noted that through image segmentation technology, the tongue body area can be accurately identified and separated, excluding the interference of the background and other irrelevant parts.

[0045] It should be further pointed out that the binary mask of the tongue body region and the background ensures the accurate extraction of the color values of all pixel points within the tongue body region in the HSV color space without being interfered by the background, which is conducive to obtaining a more reliable color uniformity evaluation result.

[0046] Convert the tongue image from the RGB color space to the HSV color space, and apply the generated mask above to accurately extract the color values of all pixel points within the tongue body region.

[0047] It should be known that in the HSV color space, the H channel represents hue, which describes the type of color, such as red, green, or blue, etc. The S channel represents saturation, which reflects the purity of the color, that is, the degree of gray in the color. The higher the saturation, the more vivid the color; the lower the saturation, the closer the color is to gray. The V channel represents brightness, which indicates the brightness of the color. These three channels respectively describe different attributes of the color.

[0048] The above conversion of the color space is for better analysis of the color distribution, because compared with the RGB color space, the HSV color space is more superior in representing the differences in human visual perception of colors.

[0049] Independently calculate the standard deviation of the pixel values for each color channel, and obtain the color uniformity by weighted summing the standard deviations of each color channel.

[0050] The weight values of each color channel above can be set according to requirements.

[0051] Compare the tongue body clarity and color uniformity of the tongue image with the set passing thresholds respectively. If both meet the passing thresholds, the tongue image enhancement process is not triggered. On the contrary, if any of the indicators fails to reach the passing threshold, the enhancement process of the tongue image is automatically triggered.

[0052] The passing thresholds of the tongue body clarity and color uniformity above can be set according to experimental data or experience.

[0053] It should be emphasized that the tongue body clarity meeting the passing threshold means that the tongue body clarity is greater than or equal to the passing threshold, while the color uniformity is calculated based on the standard deviation of the color channel. The larger the standard deviation, the more uneven the color distribution and the more problems with color inconsistency. Therefore, the color uniformity meeting the passing threshold means that the color uniformity is less than or equal to the passing threshold.

[0054] It should be noted that in the quality detection of tongue image, the clarity of the tongue body and the color uniformity are selected as quality indicators because the clarity of the tongue body and the color uniformity respectively reflect the key aspects of the image quality from different angles. The clarity of the tongue body focuses on the structural information of the image, while the color uniformity focuses on the color information. The combination of the two can provide a comprehensive and detailed image quality assessment.

[0055] In a further preferred implementation of the above solution, the tongue image enhancement processing operation is as follows: Define the image enhancement processing rule as: a) When the clarity of the tongue body in the tongue image does not meet the standard threshold, perform edge degradation processing on the tongue image.

[0056] b) When the color uniformity of the tongue image does not meet the standard threshold, perform color correction processing on the tongue image.

[0057] When the enhancement processing of the tongue image is triggered, use the above rules to perform image enhancement processing within the specified time window.

[0058] It should be understood that setting the specified time window when triggering the enhancement processing of the tongue image is based on the following considerations: Patients upload tongue images within a fixed time period. If there is no time window limit and image enhancement processing is allowed to run without restriction, long-term image processing may lead to a backlog in the processing queue, thereby delaying the response speed of the entire system; in addition, even after image enhancement processing, some images may still not meet the standard threshold. In this case, patients need to upload new tongue images. To ensure the consistency and comparability of data, patients' tongue images should be uploaded within the same time period. Without the time window limit, the re-upload time point may become uncertain, thus affecting the time series consistency and reliability of subsequent analysis.

[0059] Applied to the above understanding, the specified time window should be shorter than the fixed time period for uploading tongue images.

[0060] After each image enhancement processing, recalculate the clarity of the tongue body and the color uniformity of the enhanced image, and compare them with the standard threshold again. If they still do not meet the standard, perform multiple iterative enhancements according to the above rules until the standard threshold is met or until the available processing duration within the time window is exhausted.

[0061] If the clarity of the tongue body and the color uniformity of the tongue image still do not meet the standard threshold at the end of the specified time window, a reshoot instruction will be automatically generated to prompt that the tongue image needs to be reshot.

[0062] Before feature extraction using the tongue image uploaded by the patient, the present invention triggers image enhancement processing through an image quality detection feedback mechanism based on tongue body clarity and color uniformity, which can effectively improve low-quality images and is conducive to improving the accuracy of subsequent tongue image feature extraction. In addition, this mechanism ensures that all tongue images for analysis have a consistent quality standard, improving the reliability and comparability of time series analysis.

[0063] The tongue image feature extraction module is used to divide the enhanced tongue image into three tongue body regions: the tip of the tongue, the middle of the tongue, and the root of the tongue, and extract the tongue image features of each tongue body region respectively, specifically the HSV value of the tongue color, the proportion of pixels occupied by the tongue coating thickness, and the crack density.

[0064] It should be noted that the above extraction of tongue image features by dividing the tongue body region is because the three regions of the tip of the tongue, the middle of the tongue, and the root of the tongue have different physiological functions and pathological manifestations in traditional Chinese medicine theory. For example, the tip of the tongue is often related to the heart and lungs, the middle of the tongue is related to the spleen and stomach, and the root of the tongue is closely related to the kidneys. Therefore, extracting features by region can more accurately reflect the specific health status of each part.

[0065] It should also be noted that the selection of the HSV value of the tongue color, the proportion of pixels occupied by the tongue coating thickness, and the crack density as tongue image features is because the tongue color is one of the important bases for traditional Chinese medicine diagnosis, reflecting the state of qi, blood, body fluids, etc. in the body. Through quantitative analysis in the HSV color space, the tongue color state can be more accurately described.

[0066] The thickness of the tongue coating is an important indicator for judging pathological states such as dampness-heat and phlegm-fluid in the body. By quantitatively analyzing the proportion of pixels in the area covered by the tongue coating, the thickness and distribution of the tongue coating can be objectively evaluated.

[0067] The presence and distribution of cracks on the tongue surface are also important reference factors for syndrome differentiation and treatment in traditional Chinese medicine. The crack density reflects the microscopic structure of the tongue body surface and can reflect the functional state and pathological changes of internal organs in the body.

[0068] Through these three features, namely the HSV value of the tongue color, the proportion of pixels occupied by the tongue coating thickness, and the crack density, the tongue image can be comprehensively evaluated from multiple dimensions such as color, texture, and structure, providing more abundant diagnostic information.

[0069] In an optional implementation of the above solution, the division of the tongue body region can follow the following geometric rules: Tip of the tongue region: Approximately the front one-third part of the front end of the tongue body.

[0070] Middle of the tongue region: Approximately the middle one-third part of the tongue body.

[0071] Root of the tongue region: Approximately the back one-third part of the back end of the tongue body.

[0072] In another alternative implementation of the above solution, extracting the HSV value of the tongue color, the proportion of moss thickness pixels, and the crack density includes the following: Perform the conversion from the RGB color space to the HSV color space for the three regions of the tip of the tongue, the middle of the tongue, and the root of the tongue that have been segmented.

[0073] Calculate the average values of the pixels in each region in the H, S, and V channels respectively in the HSV color space, so as to obtain the HSV value of the tongue color in each region.

[0074] Define a specific threshold range in the HSV color space to identify white pixels as the tongue moss marker. For the three regions of the tip of the tongue, the middle of the tongue, and the root of the tongue, respectively count the number of white pixels that meet the threshold condition, and calculate the proportion of the total number of pixels in their respective regions, and use this as the proportion of moss thickness pixels in this region.

[0075] It should be noted that in the HSV color space, the moss thickness usually appears as white or nearly white pixels. By setting a specific threshold range, such as , , to identify white pixels.

[0076] Perform crack recognition on each region, and then calculate the ratio of the cumulative length of all identified cracks to the area of the region as the crack density.

[0077] It should be noted that cracks usually appear as dark linear structures on the surface of the tongue body, and cracks can be identified through edge detection algorithms or morphological gradient operations.

[0078] The tongue image change analysis module is used to store the tongue image features uploaded regularly in time series, and generate a tongue image change curve including the corresponding tongue color, moss thickness, and density of each tongue body region.

[0079] The above method for generating the tongue image change curve is as follows: For each tongue body region, based on its corresponding time series tongue image features, draw the tongue color change curve, the moss thickness change curve, and the crack change curve respectively in a two-dimensional coordinate system with time as the horizontal axis and the HSV value of the tongue color, the proportion of moss thickness pixels, and the crack density as the vertical axis.

[0080] The traditional Chinese medicine large model prediction module is used to import the tongue image change curve into the constructed tongue image - syndrome type model to analyze the syndrome type in the time series, and output the syndrome type evolution path.

[0081] It should be noted that the above-mentioned tongue image - syndrome type model takes the tongue image features corresponding to each tongue body region as input and the syndrome type as output.

[0082] The basis for constructing the above tongue image-syndrome type model is that traditional Chinese medicine emphasizes "syndrome differentiation and treatment". Tongue images are one of the important bases for syndrome differentiation. Different regions of the tongue body correspond to different parts, and tongue color, tongue coating thickness, and crack characteristics are closely related to specific syndrome types.

[0083] In an example, the corresponding syndrome type for a red tip of the tongue, thick and greasy coating in the middle of the tongue, and pale white at the root of the tongue is excessive heart fire combined with dampness-heat in the spleen and stomach and kidney yang deficiency. The specific explanations are as follows: A red tip of the tongue with a thin yellow coating indicates excessive heart fire, which is commonly seen in symptoms such as insomnia, palpitations, and dry mouth.

[0084] A light red tongue in the middle with a thick and greasy coating indicates dampness-heat in the spleen and stomach, which is commonly seen in symptoms such as indigestion, abdominal distension, nausea, and vomiting.

[0085] A pale white tongue root with a white and slippery coating indicates kidney yang deficiency, which is commonly seen in symptoms such as fear of cold and cold limbs, and frequent urination at night.

[0086] When constructing the tongue image-syndrome type model, machine learning methods can be used: First, collect a large number of tongue image samples, and label the syndrome types of each tongue image sample by professional Chinese medicine doctors to ensure the authenticity and accuracy of the data. At the same time, preprocess the collected tongue image samples and extract tongue image features to form feature vectors. Then, use the feature vectors and syndrome type labels of each tongue image sample for model training, and evaluate and adjust the model during the training process to achieve an accurate mapping between tongue image features and syndrome types.

[0087] According to a preferred embodiment, the analysis of the syndrome type under the time series is as follows: Perform time segmentation on the tongue image change curves corresponding to the tongue color, tongue coating thickness, and cracks in each region of the tongue body. The specific segmentation process is as follows: Starting from the starting time, gradually identify the changes in the HSV values of the tongue color, the proportion of tongue coating pixels, and the crack density in adjacent times for the tongue image change curves corresponding to the tongue color, tongue coating thickness, and density in each region of the tongue body. If no changes occur in all tongue image features in a certain adjacent time, then proceed to identify the changes in tongue image features in the next adjacent time until a change occurs in a certain tongue image parameter in the adjacent time. Then, use the subsequent time corresponding to this adjacent time as the change time.

[0088] The above change identification process is as follows: Calculate the difference in the HSV values of the tongue color, the proportion of tongue coating pixels, and the crack density for adjacent times respectively, and compare with the set critical difference. If the difference in a certain tongue image parameter is greater than the critical difference, then it is identified that a change has occurred.

[0089] It should be noted that at least one tongue image feature has changed at each change time.

[0090] Mark each change time as the time point segmented on the tongue image change curve.

[0091] Finally, a series of segmentation points are generated through the above steps, and these segmentation points divide the tongue image change curve into multiple time periods. The tongue image features within each time period remain relatively stable, while at least one feature changes significantly when crossing a segmentation point.

[0092] For each time point, the tongue image features of each tongue body region are extracted from the above change curve respectively.

[0093] The tongue image features corresponding to each time point and each tongue body region are used as the input set and input into the tongue image-syndrome type model to obtain the syndrome types at each time point.

[0094] It should be noted that the above method of determining the syndrome type by extracting tongue image features according to the segmented time points is because at least one tongue image feature changes significantly when crossing a segmentation point. This significant change usually marks an important transformation of the patient's tongue image state. By extracting tongue image features at each segmentation point, these important tongue image state changes can be captured, and the changes in tongue image state are often closely related to the changes in syndrome types. For example, the tongue color changing from pale white to crimson may indicate a transition from qi deficiency to heat syndrome. By extracting features at each segmentation point and inputting them into the model, these changes in syndrome types can be more accurately identified, thereby improving the accuracy of diagnosis. In addition, if the tongue image features do not change significantly within a time period, there is no need to frequently re-evaluate the syndrome type. By setting segmentation points, new evaluations are only carried out when the features change, reducing unnecessary calculations and optimizing the use of system resources.

[0095] According to another preferred embodiment, see Figure 3 As shown, the output of the syndrome type evolution path includes the following content: Use lines to connect the syndrome types at adjacent time points in the time series in sequence to form a continuous syndrome type evolution path.

[0096] For each time point, compare whether its syndrome type is the same as that of the previous time point. If they are different, then identify this time point as a syndrome type change time point and mark the syndrome type change time point in the syndrome type evolution path.

[0097] As an example of the above solution, specific symbols such as asterisks or triangles can be used to mark the syndrome type change time points to highlight them.

[0098] Another example is that the syndrome type change time points can be marked with colors, such as using different colors or line types to distinguish the time periods before and after the change to enhance the visual contrast.

[0099] The diagnosis and treatment association decision module is used to associate the clinical indicators of the patient in the same time series with the syndrome type evolution path to form a syndrome type-pathology association matrix, and based on this, recommend a prescription compatibility plan.

[0100] Among the ways that the above solution can be implemented, the clinical indicators of the patient under the same-period time series are associated with the syndrome type evolution path to form a syndrome type-pathology association matrix, and the specific implementation is as follows: Based on the syndrome type change time points marked on the syndrome type evolution path, the syndrome types corresponding to each syndrome type change time point are extracted.

[0101] For each syndrome type change time point, a time interval is determined under the set time delay condition, and the corresponding clinical indicator data is extracted within this time interval to form a mapping set of the syndrome type corresponding to each syndrome type change time point and the clinical indicators.

[0102] In the example of the above operation, the syndrome type change time point Under the time delay condition of The determined time interval is .

[0103] It should be understood that in the process of extracting clinical indicator data based on the syndrome type evolution path, setting the time delay condition to determine the time interval for clinical indicator extraction is considered because traditional Chinese medicine theory believes that the change of syndrome type is often a comprehensive manifestation of the internal pathological state of the body, and this change usually lags behind the change of some clinical indicators. For example, blood biochemical indicators such as blood sugar and blood lipids may change before tongue image features. By setting the time delay condition, these clinically antecedent-changing indicators can be captured, so as to better understand the potential mechanism of syndrome type change. In addition, the time delay condition allows the system to collect clinical indicator data within a reasonable time interval, so as to obtain more comprehensive and representative information. This method can reduce the error caused by single-time-point sampling.

[0104] It should be noted that the above-mentioned clinical indicators include but are not limited to physical sign indicators, blood biochemical indicators, endocrine hormone indicators, etc. Among them, physical sign indicators include body temperature, heart rate, respiratory rate, blood pressure, etc., blood biochemical indicators include blood sugar, blood lipids, blood routine, electrolytes, etc., and endocrine hormone indicators include thyroid, sex hormones, etc.

[0105] Variation identification is performed on the clinical indicator mapping sets of adjacent syndrome type change time points under the time delay condition to determine the changing clinical indicators and their change amounts corresponding to the syndrome type change.

[0106] The above-mentioned variation identification can be similarly referred to the variation identification method of adjacent tongue image parameters. The specific identification steps are as follows: (1) Each clinical indicator number extracted within the time interval determined under the time delay condition is used to generate a clinical indicator change curve in a coordinate system with time as the horizontal axis and clinical indicators as the vertical axis.

[0107] (2) The stationary phase data is extracted from the generated clinical indicator change curve as the representative data of the clinical indicators. Specifically, the representative data of the clinical indicators can be the average value or the median.

[0108] It should be noted that the stable period data mentioned above refers to the data where the clinical indicators change little within a certain time period and can reflect the typical values of the clinical indicators within the time interval.

[0109] (3) Subtract the representative data of each clinical indicator corresponding to the adjacent syndrome type change time points for the same item to obtain the difference of each clinical indicator, and compare it with the critical difference. If the difference of a certain clinical indicator is greater than the critical difference, it is recognized that there is a change in that clinical indicator.

[0110] Further, the change amount of the changed clinical indicator corresponding to the adjacent syndrome type change time point is the difference of the changed clinical indicator.

[0111] Compare the front and back syndrome types corresponding to each syndrome type change time point, and classify the same front and back syndrome types into one category.

[0112] Statistically analyze the change frequency and average change amount of the changed clinical indicators in each category of front and back syndrome type combinations, and thus assign an association strength to the changed clinical indicators. The specific implementation is as follows: Normalize the change frequency and average change amount of the changed clinical indicators in each category of front and back syndrome type combinations respectively.

[0113] It should be understood that the change frequency refers to the number of times the changed clinical indicator changes repeatedly during the conversion between the front and back syndrome types. The higher the change frequency, the greater the frequency of change of the clinical indicator during the specific syndrome type conversion, indicating that it has a stronger association with the syndrome type conversion.

[0114] The average change amount refers to the average value of each change of the changed clinical indicator during the conversion between the front and back syndrome types. The larger the average change amount, the greater the amplitude of change of the clinical indicator during the specific syndrome type conversion, indicating that its influence during the syndrome type conversion is more significant.

[0115] The above normalization of the change frequency and average change amount is to map these values into a standardized interval, such as the interval [0, 1]. This processing method ensures that data of different scales and magnitudes can be compared and analyzed on the same basis.

[0116] Calculate the association strength by fusing the normalized change frequency and average change amount of the changed clinical indicators.

[0117] In an example, the fusion calculation can use the weighted summation method. The weights of the change frequency and average change amount can be set according to experience, and the sum of the weights of the change frequency and average change amount is 1. This method can adjust the weights according to actual needs and is applicable to situations where it is necessary to comprehensively consider the change frequency and average change amount.

[0118] In another example, the fusion calculation can adopt the product method. This method emphasizes that the correlation strength will increase significantly only when both the change frequency and the average change amount are large, and it is suitable for strictly screening cases of changing clinical indicators.

[0119] It should be understood that during the process of the transformation of the patient's syndrome type, if a certain clinical indicator is observed to change synchronously, it indicates that there may be an internal connection between this clinical indicator and the syndrome type. By means of quantifying the change frequency and the change amount, such a correlation can be systematically measured and evaluated, so as to clarify the relationship strength and significance between the two.

[0120] Construct a syndrome type - pathology correlation matrix based on the correlation strength between the combination of the previous and subsequent syndrome types and the changing clinical indicators, where the rows of the matrix represent the combination of the previous and subsequent syndrome types, the columns represent the clinical indicators, and the matrix elements represent the correlation strength between the combination of the previous and subsequent syndrome types and the changing clinical indicators.

[0121] In the example of the above operation, assume that we have three different combinations of the previous and subsequent syndrome types: qi deficiency syndrome -> damp - heat syndrome, qi deficiency syndrome -> yin - fire hyperactivity syndrome, damp - heat syndrome -> yin - fire hyperactivity syndrome, denoted as A, B, C respectively, and the changing clinical indicators corresponding to these three syndrome type combinations are all: blood glucose, blood lipid, white blood cell count. We will construct a syndrome type - pathology correlation matrix based on this information as follows: 。

[0122] For the syndrome type conversion of "qi deficiency syndrome -> damp - heat syndrome", the correlation strength of blood glucose is 0.75, indicating that the change of blood glucose is relatively significant during this process and may be one of the key factors in the syndrome type conversion.

[0123] For the syndrome type conversion of "qi deficiency syndrome -> yin - fire hyperactivity syndrome", the correlation strength of white blood cell count is 0.70, suggesting that the change of white blood cell count is the most significant during this process and may have an important impact on the syndrome type conversion.

[0124] For the syndrome type conversion of "damp - heat syndrome -> yin - fire hyperactivity syndrome", the correlation strength of blood lipid is 0.85, indicating that the change of blood lipid is relatively significant during this process and may be one of the important signs of the syndrome type conversion.

[0125] In another implementation manner of the above - mentioned scheme, the recommended formula compatibility scheme includes the following content: Set a threshold based on the range of the correlation strength values in the correlation matrix to define a strong correlation.

[0126] Exemplarily, since the association strength is calculated by fusing the normalized change frequency and the average change amount, its value range is between [0, 1]. To emphasize a strong association, a threshold, such as 0.5, can be set to define a strong association. This means that when the association strength exceeds 0.5, it is considered that the clinical index has a strong association with the combination of the previous and current syndrome types, thus highlighting those indicators with significant effects.

[0127] Utilize the syndrome type evolution path and the syndrome type - pathology association matrix to identify the syndrome type evolution of the current patient and the clinical indicators strongly associated with it.

[0128] Screen out the basic prescriptions applicable to the current syndrome type evolution from the basic prescription library constituted by the mapping between syndrome types and common prescriptions.

[0129] Determine the priority of the clinical indicators strongly associated with the current syndrome type evolution in the prescription compatibility according to the association strength in the association matrix, and optimize the basic prescription accordingly.

[0130] Take an example of applying to the syndrome type - pathology association matrix. Assume that the current syndrome type evolution is from qi deficiency syndrome to yin deficiency with flaring fire syndrome. According to this matrix, the clinical indicators strongly associated with this syndrome type transformation include blood lipid and white blood cell count, and their association strengths are 0.6 and 0.7 respectively. Given that the association strength of the white blood cell count is higher, it indicates that it has a greater impact during this syndrome type conversion. Therefore, when adjusting the drug selection in the basic prescription, the change of the white blood cell count should be considered preferentially to optimize the treatment plan.

[0131] The specific drug selection is as follows: If the white blood cell count increases, suggesting possible inflammation or infection, traditional Chinese medicines with the effects of clearing heat and detoxifying, cooling blood and detumescence, such as honeysuckle flower and forsythia fruit, can be added to the basic prescription.

[0132] If the white blood cell count decreases, which may suggest low immune function, traditional Chinese medicines for tonifying qi and nourishing blood and enhancing immunity, such as astragalus root and codonopsis pilosula, can be considered for addition.

[0133] Adjust the drug combination according to the specific change of the white blood cell count on the premise of ensuring the balance of the overall prescription, ensuring that the prescription can effectively respond to the current pathological state and promote the transformation of the syndrome type towards a healthy direction.

[0134] It should be noted that the prescription compatibility plan is not fixed and should be adjusted regularly according to the patient's condition changes and treatment effects.

[0135] The above embodiments can be implemented in whole or in part through software, hardware, firmware or any other arbitrary combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0136] Those of ordinary skill in the art will realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of this application.

[0137] In addition, the functional modules in each embodiment of this application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0138] As mentioned above, the above are only specific embodiments of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0139] Finally, the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. An auxiliary diagnosis and treatment system based on a large-scale traditional Chinese medicine model, characterized in that, Including: Tongue image data acquisition module: Receives tongue images uploaded by patients at the same regular time period, and triggers enhanced processing of tongue images based on an image quality detection feedback mechanism for tongue body clarity and color uniformity; Tongue image feature extraction module: Divides the enhanced tongue image into three tongue body regions: the tip of the tongue, the middle of the tongue, and the root of the tongue, and extracts the tongue image features of each tongue body region respectively, specifically the HSV value of the tongue color, the proportion of moss thickness pixels, and the crack density; Tongue image change analysis module: Stores the tongue image features of regularly uploaded tongue images in a time series, and generates a tongue image change curve including the corresponding tongue color, moss thickness, and cracks in each tongue body region; Traditional Chinese medicine large model prediction module: Imports the tongue image change curve into the constructed tongue image-syndrome type model to analyze the syndrome type in the time series, and outputs the syndrome type evolution path; Diagnosis and treatment association decision module: Associates the clinical indicators of the patient in the same time series with the syndrome type evolution path to form a syndrome type-pathology association matrix, and recommends a prescription compatibility plan based on this. The specific method is as follows: Based on the syndrome type change time point of the syndrome type evolution path, extract clinical indicator data within a set time delay interval, identify the changing indicators and count their change frequencies and average change amounts, and construct a syndrome type-pathology association matrix based on the association strength between the front and back syndrome type combinations and the changing clinical indicators. The rows of the matrix represent the front and back syndrome type combinations, the columns represent the clinical indicators, and the matrix elements represent the association strength between the front and back syndrome type combinations and the changing clinical indicators.

2. The auxiliary diagnosis and treatment system based on the traditional Chinese medicine large model according to claim 1, wherein: The triggering of the enhanced processing of tongue images by the image quality detection feedback mechanism based on tongue body clarity and color uniformity is as follows: Perform noise preprocessing operations on the received tongue images; Use an edge detection algorithm to extract the tongue body edge pixels from the preprocessed image; Evaluate the edge sharpness and edge continuity respectively based on the tongue body edge pixels; Define the tongue body clarity as the product of the edge sharpness and the edge continuity. Under this definition, obtain the tongue body clarity of the tongue image based on the evaluated edge sharpness and edge continuity; Use image segmentation technology to separate the tongue body region from the background and generate a corresponding binary mask, ensuring that the pixel values within the tongue body region are marked as 1, while the background region is marked as 0; Convert the tongue image from the RGB color space to the HSV color space, and use the generated mask above to accurately extract the color values of all pixel points within the tongue body region; Independently calculate the standard deviation of the pixel values for each color channel, and obtain the color uniformity by weighted summing the standard deviations of each color channel; Compare the tongue body clarity and color uniformity of the tongue image with the set passing thresholds respectively. If both meet the passing thresholds, the enhanced processing of the tongue image is not triggered. Conversely, if either indicator fails to reach the passing threshold, the enhanced processing of the tongue image is automatically triggered.

3. The auxiliary diagnosis and treatment system based on the traditional Chinese medicine large model according to claim 2, wherein: The triggering of the enhanced processing of tongue images by the image quality detection feedback mechanism based on tongue body clarity and color uniformity also includes the following: Define the image enhancement processing rules as: a) When the tongue body clarity of the tongue image does not meet the passing threshold, perform edge degradation processing on the tongue image; b) When the color uniformity of the tongue image does not meet the passing threshold, perform color correction processing on the tongue image; When triggering the enhancement processing of the tongue image, perform image enhancement processing using the above rules within the specified time window; After each image enhancement processing, recalculate the clarity of the tongue body and the color uniformity of the enhanced image, and compare them with the compliance threshold again. If they still do not meet the standard, perform multiple iterative enhancements according to the above rules until the compliance threshold is met or until the available processing duration within the time window is exhausted; If the clarity of the tongue body and the color uniformity of the tongue image still do not meet the compliance threshold at the end of the specified time window, automatically generate a reshooting instruction to prompt that the tongue image needs to be reshot.

4. The auxiliary diagnosis and treatment system based on the traditional Chinese medicine large model according to claim 1, wherein: The extraction of the HSV value of the tongue color, the proportion of the pixel area of the tongue coating thickness, and the crack density includes the following content: Perform the conversion from the RGB color space to the HSV color space for the three regions of the tongue tip, the middle of the tongue, and the root of the tongue that have been segmented; Under the HSV color space, calculate the average values of the H, S, and V channels for each pixel point within each region, so as to obtain the HSV value of the tongue color for each region; Define a specific threshold range in the HSV color space to identify white pixels as the tongue coating markers. For the three regions of the tongue tip, the middle of the tongue, and the root of the tongue, respectively count the number of white pixels that meet the threshold condition, and calculate the proportion of the total number of pixels within their respective regions, which is used as the proportion of the pixel area of the tongue coating thickness in this region; Perform crack recognition for each region, and then calculate the ratio of the cumulative length of all recognized cracks to the area of the region as the crack density.

5. The auxiliary diagnosis and treatment system based on the traditional Chinese medicine large model according to claim 1, characterized in that: The analysis of the syndrome types in the time series is as follows: Perform time segmentation on the tongue image change curves corresponding to the tongue color, the tongue coating thickness, and the cracks in each tongue body region; For each time point, extract the tongue image features of each tongue body region from the above change curves respectively; Use the tongue image features corresponding to each time point and each tongue body region as the input set and input it into the tongue image-syndrome type association model to obtain the syndrome types at each time point.

6. The auxiliary diagnosis and treatment system based on the traditional Chinese medicine large model according to claim 5, characterized in that: The time segmentation of the tongue image change curves corresponding to the tongue color, the tongue coating thickness, and the cracks in each tongue body region is as follows: Starting from the start time, gradually perform the change recognition of the HSV value of the tongue color, the proportion of the pixel area of the tongue coating thickness, and the crack density at adjacent times for the tongue image change curves corresponding to the tongue color, the tongue coating thickness, and the density in each tongue body region. If all tongue image features do not change within a certain adjacent time, then perform the change recognition of the tongue image features at the next adjacent time until a change occurs in a certain tongue image parameter within the adjacent time, and use the subsequent time corresponding to this adjacent time as the change time; Mark each change time as the time point segmented on the tongue image change curve.

7. The auxiliary diagnosis and treatment system based on the traditional Chinese medicine large model according to claim 1, wherein: The output of the syndrome type evolution path includes the following content: Use lines to connect the syndrome types at adjacent time points in the time series in sequence to form a continuous syndrome type evolution path; For each time point, compare whether its syndrome type is the same as the syndrome type at the previous time point. If it is different, then identify this time point as a syndrome type change time point and mark the syndrome type change time point in the syndrome type evolution path.

8. The auxiliary diagnosis and treatment system based on the traditional Chinese medicine large model according to claim 7, characterized in that: The specific implementation of constructing the syndrome type-pathology association matrix is as follows: Extract the syndrome type corresponding to each syndrome type change time point based on the syndrome type change time points marked on the syndrome type evolution path; For each syndrome type change time point, a time interval is determined under the set time delay condition. Within this time interval, the corresponding clinical index data are extracted to form a mapping set of the syndrome type and clinical indexes corresponding to each syndrome type change time point; Change identification is performed on the clinical index mapping sets at adjacent syndrome type change time points to determine the changed clinical indexes and their change amounts corresponding to the syndrome type change; The syndrome types before and after each syndrome type change time point are compared, and the same syndrome types before and after are grouped into one category; The change frequencies and average change amounts of the changed clinical indexes in each category of syndrome type combinations before and after are statistically analyzed, and thus association strengths are assigned to the changed clinical indexes.

9. The auxiliary diagnosis and treatment system based on the traditional Chinese medicine large model according to claim 8, wherein: The assignment of association strengths to the changed clinical indexes includes the following: The change frequencies and average change amounts of the changed clinical indexes in each category of syndrome type combinations before and after are respectively normalized; The association strength is calculated by fusing the change frequencies and average change amounts of the changed clinical indexes after normalization processing.

10. The auxiliary diagnosis and treatment system based on the traditional Chinese medicine large model according to claim 8, characterized in that: The recommended formula compatibility scheme is as follows: A threshold is set based on the range of association strength values in the association matrix to define strong associations; The syndrome type evolution of the current patient and the clinical indexes strongly associated with it are identified by using the syndrome type evolution path and the syndrome type-pathology association matrix; The basic formulas applicable to the current syndrome type evolution are screened out from the basic formula library composed of the mapping between syndrome types and common formulas; The clinical indexes strongly associated with the current syndrome type evolution in the association matrix are determined according to their association strengths to determine their priorities in formula compatibility, and the basic formula is optimized accordingly.

Images