A method and system for collecting and analyzing information from a tongue diagnosis

Abstract

The application discloses a tongue surface diagnosis information collection diagnosis analysis method and system, belongs to the medical auxiliary diagnosis technical field, the tongue surface image collected by the application is matched with the patient information, the symptom associated tongue surface area is determined, the priority of the associated tongue surface area is divided, whether the collected image meets the clear standard is judged, the image index evaluation model is constructed based on the tongue tremor and the image texture, the image to be processed is evaluated in terms of tongue body tremor, texture stability and water vapor blur interference, the stable frame evaluation model is constructed, the tremor intensity, the texture stability and the water vapor proportion are introduced into the stable frame evaluation model to evaluate the image area stability, and the stable frame set is subjected to image registration, multi-frame registration and fusion are carried out based on the image stable area, the area stability and processing information are recorded, high-quality images for tongue image analysis are generated, the image clarity of the key diagnosis area is improved, and the accuracy and stability of tongue image analysis are improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of medical auxiliary diagnosis, in particular to a tongue surface diagnosis information collection diagnosis analysis method and system. BACKGROUND

[0002] Physicians understand the physiological functions and pathological changes of the human body by observing the tongue, tongue fur and tongue base changes to determine the health status of patients. In tongue image collection, mouth water vapor transpiration and tongue tremor are common interference factors, which can easily cause image blurring, detail loss, texture distortion and other problems, thereby affecting subsequent tongue image recognition and pathological analysis. The tongue structure of different individuals differs greatly, and traditional image processing methods are difficult to stably segment the tongue tip, tongue middle and tongue edge regions, affecting subsequent tongue feature extraction and positioning diagnosis.

[0003] The application extracts diagnosis keywords from patient complaints and case information by collecting images, matches the tongue image database, reversely locates the tongue surface area that should be analyzed, comprehensively evaluates the image stability through the tongue tremor intensity, texture stability and image water vapor reflection degree, improves the image clarity of the key diagnosis area, and improves the accuracy and stability of tongue image analysis. SUMMARY

[0004] In view of the deficiencies of the prior art, the application provides a tongue surface diagnosis information collection diagnosis analysis method and system.

[0005] To achieve the above purpose, the application provides the following technical solutions:

[0006] A tongue surface diagnosis information collection diagnosis analysis method and system, comprising the following specific steps:

[0007] Match the collected tongue surface image with the patient information, determine the symptom-related tongue surface area, divide the priority of the related tongue surface area, and judge whether the collected image meets the clear standard;

[0008] Based on the tongue tremor and image texture, an image index evaluation model is constructed to evaluate the tongue tremor, texture stability and water vapor blur interference of the image to be processed;

[0009] A stable frame evaluation model is constructed, the tremor intensity, texture stability and water vapor proportion are introduced into the stable frame evaluation model to evaluate the image area stability, and the stable frame set is subjected to image registration;

[0010] Based on the image stable area, multi-frame registration and fusion are performed, the area stability and processing information are recorded, and a high-quality image for tongue image analysis is generated.

[0011] Preferably, the matching of the collected tongue surface image with the patient information, determining the symptom-associated tongue surface area, prioritizing the associated tongue surface area, and determining whether the collected image meets the clarity standard include the following specific steps:

[0012] S11, continuously collecting a sequence of tongue surface video frames, simultaneously acquiring a thermal imaging map, a near-infrared spectrum map, acquiring patient clinical complaint information and past cases, extracting diagnostic keywords and symptom key information from the patient's clinical complaint information and past case descriptions through natural language processing technology, and matching the keywords with a TCM tongue symptom database to preliminarily determine the patient's symptom-associated tongue surface area range;

[0013] S12, using a trained tongue segmentation network model to segment the collected tongue surface image, extracting the tongue edge, eliminating background, teeth, and lip interference, and forming a standardized tongue surface contour, taking the longitudinal center axis and the transverse width center line of the tongue surface geometry as the segmentation reference, constructing a standardized rectangular coordinate system, and dividing the tongue into a tongue tip area, a tongue middle area, a tongue edge area, and a tongue root area;

[0014] S13, matching the key tongue surface area corresponding to the symptom keywords identified in S11 with the standard tongue surface area segmented in S12, assigning diagnostic importance weights to the multiple tongue surface areas associated with symptoms according to historical data, indicating the degree of contribution of the area to the diagnosis of the current symptom, determining the key area in the image that needs to be analyzed first, and forming a unified clarity score by normalizing the tongue surface area fusion frame through Laplacian variance and high-frequency energy ratio processing. Multiply the diagnostic importance of each area with the image clarity to get the area diagnosis quality score. Set the image quality judgment threshold, divide the weight, set the area diagnosis importance threshold, which is used to distinguish key areas and secondary areas. If the diagnostic contribution is higher than the area diagnosis importance threshold, and the area clarity is higher than the image quality judgment threshold, the image is directly used for subsequent diagnosis. If the diagnostic contribution is higher than the area diagnosis importance threshold, and the area clarity is lower than the image quality judgment threshold, the image is processed. If the diagnostic contribution is lower than the area diagnosis importance threshold, and the area clarity is higher than the image quality judgment threshold, the image area can be used for subsequent auxiliary diagnosis.

[0015] Preferably, the image index evaluation model is constructed based on tongue tremor and image texture, and the tongue tremor, texture stability, and water vapor blur interference evaluation of the image to be processed include the following specific steps:

[0016] S21, during the shooting process, the tongue involuntary movement causes the image to be out of focus locally and as a whole, the tongue tremor is evaluated by analyzing the position change of the key points of the tongue surface area between the continuous frames, the quantitative evaluation of the stability of the tongue body is realized, the coordinate positions of the related areas in the continuous frame images are extracted, the coordinate positions are substituted into the tongue tremor intensity calculation formula to evaluate the average offset amplitude of the key points in the time sequence, wherein the tongue tremor intensity calculation formula is: , wherein, is the key point coordinate of the region in the i-th frame image, is the average position coordinate of the key points of the whole sequence, is the Euclidean distance between vectors, n is the total number of frames, the higher the tongue tremor intensity value, the more unstable the region in the collection process, and the lower the image reliability;

[0017] S22, the tongue surface texture features of each region are extracted, and the texture consistency between the continuous frames is calculated by the texture stability calculation formula, wherein the texture stability calculation formula is: , wherein, is the image of the tongue surface area R in the i-th frame, wherein the structural similarity index calculation formula is: , wherein, is the average value of the region brightness in the i-th frame, is the region brightness variance in the i-th frame, is the covariance, and are constants, which can quantitatively describe whether the texture structure is stable, and pay more attention to local structure and texture integrity;

[0018] S23, the reflection of water vapor or saliva in the mouth will cause part of the region in the image to be blurred or appear bright spots, which will affect the texture, color and detail texture recognition of the tongue coating, based on the image definition detection of texture sharpness and frequency domain response, the blurred part in the image region is identified, the high reflection part in the image region is detected, combined with the heat map and near-infrared image, the bright spot part caused by saliva transpiration is identified, and the water vapor part ratio is calculated.

[0019] Preferably, the stable frame evaluation model is constructed, the tremor intensity, texture stability and water vapor ratio are introduced into the stable frame evaluation model to evaluate the image region stability, and the stable frame set is image registered, which includes the following specific steps:

[0020] S31, the tremor intensity, texture stability and water vapor area ratio are substituted into the comprehensive stability score calculation formula to evaluate the image stability, wherein the comprehensive stability score calculation formula is: , wherein, alpha is a tremor sensitive adjustment parameter, which is used to control the weight of the influence of the jitter amplitude on the stability, and H is the water vapor area ratio;

[0021] S32, sort the same associated areas of different frames from high to low according to the comprehensive stability score, select the top k% frames with the highest scores as the relatively stable frame set, the frames in the set have the characteristics of small tongue tremor, complete texture structure and high image clarity, perform image registration on the stable frame set to eliminate image misplacement caused by tongue tremor or shooting angle change.

[0022] Preferably, the multi-frame registration and fusion based on the image stable area, recording the area stability and processing information, generating high-quality images for tongue image analysis includes the following specific steps:

[0023] S41, image fusion based on image stable area, key point feature matching and registration are performed on consecutive frames, affine transformation and perspective transformation are used to align the images, avoiding ghosting and misplacement in the fusion process, frame average fusion is used to improve image signal-to-noise ratio and clarity, wavelet enhancement is used to enhance the texture contrast of low clarity areas, the enhanced images are re-input into the tongue image feature extraction and symptom matching process, and the area stability score and enhancement processing record are marked in the image metadata.

[0024] A tongue surface diagnosis information collection and diagnosis analysis system based on the above-mentioned tongue surface diagnosis information collection and diagnosis analysis method, which specifically includes:

[0025] A symptom association matching module is used to match the collected tongue surface image with the patient information, determine the symptom-associated tongue surface area, divide the priority of the associated tongue surface area, and judge whether the collected image meets the clarity standard;

[0026] An image index evaluation module is used to evaluate the tongue tremor, texture stability and water vapor blur interference of the image to be processed;

[0027] A stable frame evaluation module is used to evaluate the image area stability through tremor intensity, texture stability and water vapor proportion, and perform image registration on the stable frame set;

[0028] An image processing module is used to perform multi-frame registration and fusion on the image stable area, record the area stability and processing information, and generate high-quality images for tongue image analysis.

[0029] An electronic device, comprising a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0030] The processor executes the above-mentioned tongue surface diagnosis information collection and diagnosis analysis method by calling the computer program stored in the memory.

[0031] A computer readable storage medium, characterized in that, store instructions, when the instructions run on the computer, make the computer execute the tongue surface diagnosis information collection diagnosis analysis method.

[0032] Compared with the prior art, the beneficial effects of the present application are:

[0033] The tongue surface image collected by the present application is matched with the patient information to determine the symptom-associated tongue surface area, the priority of the associated tongue surface area is divided, it is judged whether the collected image meets the clear standard, an image index evaluation model is constructed based on tongue tremor and image texture, the image to be processed is evaluated in terms of tongue tremor, texture stability and water vapor blur interference, a stable frame evaluation model is constructed, the tremor intensity, texture stability and water vapor proportion are introduced into the stable frame evaluation model to evaluate the image area stability, and the stable frame set is image registered, multi-frame registration and fusion are performed based on the stable image area, the area stability and processing information are recorded, and a high-quality image for tongue image analysis is generated, the image clarity of the key diagnosis area is improved, and the accuracy and stability of tongue image analysis are improved. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 The present application is a tongue surface diagnosis information collection diagnosis analysis method whole flow schematic diagram;

[0035] Figure 2 The present application is a comprehensive stability score calculation flowchart;

[0036] Figure 3 The present application is a tongue surface image partition schematic diagram;

[0037] Figure 4 The present application is a tongue surface diagnosis information collection diagnosis analysis system whole framework schematic diagram. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all embodiments.

[0039] Embodiment 1

[0040] Please refer to Figures 1-3 The present application provides an embodiment: a tongue surface diagnosis information collection diagnosis analysis method, which comprises the following specific steps:

[0041] The collected tongue surface image is matched with the patient information to determine the symptom-associated tongue surface area, the priority of the associated tongue surface area is divided, and it is judged whether the collected image meets the clear standard;

[0042] An image index evaluation model is constructed based on tongue tremor and image texture, and the images to be processed are evaluated in terms of tongue tremor, texture stability and water vapor interference;

[0043] A stable frame evaluation model is constructed, the tremor intensity, texture stability and water vapor proportion are introduced into the stable frame evaluation model to evaluate the image region stability, and the stable frame set is subjected to image registration;

[0044] Based on the image stable region, multi-frame registration and fusion are carried out, the region stability and processing information are recorded, and high-quality images for tongue image analysis are generated.

[0045] In the embodiment, the collected tongue surface image is matched with the patient information, the symptom associated tongue surface region is determined, the priority of the associated tongue surface region is divided, and whether the collected image meets the clear standard is judged, including the following specific steps:

[0046] S11, continuously collect tongue surface video frame sequence, simultaneously acquire thermal imaging map, near-infrared spectrum map, acquire patient clinical complaint information and past cases, extract diagnosis key words and symptom key information in patient clinical complaint information and past case description through natural language processing technology, match key words with traditional Chinese medicine tongue image symptom database, preliminarily determine patient symptom associated tongue surface region range, and improve the pertinence of image region analysis;

[0047] S12, using a trained tongue segmentation network model such as U-Net, DeepLab, etc., the collected tongue surface image is segmented, the tongue edge is extracted, the background, teeth and lip interference are eliminated, the standardized tongue surface contour is formed, the longitudinal center axis and the transverse width center line of the tongue surface geometric shape are taken as the segmentation reference, a standardized rectangular coordinate system is constructed, the tongue is divided into tongue tip area, tongue middle area, tongue edge area and tongue root area to provide structured region label, and subsequent region analysis is facilitated;

[0048] S13, the key tongue surface area corresponding to the symptom keyword identified in S11 is matched with the standard tongue surface area segmented in S12, a plurality of tongue surface areas associated with the symptom are assigned a diagnostic importance weight according to historical data, indicating the degree of contribution of the area to the diagnosis of the current symptom, derived from historical case statistics or expert annotation, to determine the key area that needs to be analyzed preferentially in the image, the uniform clarity score is formed after the tongue surface area fusion frame is normalized through Laplacian variance (edge sharpness) and high-frequency energy ratio (detail retention degree) processing, the diagnostic importance of each area is multiplied by the image clarity to obtain the area diagnostic quality score, the weight is divided by setting the image quality judgment threshold, and the area diagnostic importance threshold is set to distinguish between key areas and secondary areas, if the diagnostic contribution degree is higher than the area diagnostic importance threshold, and the area clarity is higher than the image quality judgment threshold, the image is directly used for subsequent diagnosis, if the diagnostic contribution degree is higher than the area diagnostic importance threshold, and the area clarity is lower than the image quality judgment threshold, the image is processed, if the diagnostic contribution degree is lower than the area diagnostic importance threshold, and the area clarity is higher than the image quality judgment threshold, the image area can be used for subsequent auxiliary diagnosis, avoiding misjudgment caused by good overall image clarity but blurred key area.

[0049] It needs to be specifically described in the present embodiment that the image index evaluation model is constructed based on tongue tremor and image texture, and the image to be processed is evaluated for tongue tremor, texture stability and water vapor blur interference, including the following specific steps:

[0050] S21, during the shooting process, the tongue moves involuntarily, causing the image to be out of focus locally and as a whole, the tongue tremor is evaluated by analyzing the position change of the key points of the tongue surface area in the image sequence between consecutive frames, realizing quantitative evaluation of the stability of the tongue body, the coordinate positions of the associated areas in the consecutive frame images are extracted, and the coordinate positions are substituted into the tongue tremor intensity calculation formula to evaluate the average displacement amplitude of the key points in the time sequence, wherein the tongue tremor intensity calculation formula is: wherein, is the key point coordinate of the region in the i-th frame image, i.e. the region centroid or feature point position, is the average position coordinate of the key points of the entire sequence, is the Euclidean distance between vectors, n is the total number of frames, the higher the tongue tremor intensity value, the more unstable the region in the acquisition process, the lower the image reliability, the stability of different regions and different frames in the image sequence is compared, the motion stability in the image sequence is represented by the trajectory jitter amplitude of the region key points, the mean deviation can measure the fluctuation, the tongue tremor intensity calculation formula is based on the average spatial displacement of the key points in the image sequence, and the motion stability of the tongue body in the shooting process is objectively quantified;

[0051] S22, extract the texture features of the tongue surface of each region, including tongue lines, tooth marks, cracks, and thickness of tongue coating, and calculate the texture consistency between consecutive frames by a texture stability calculation formula, wherein the texture stability calculation formula is: , wherein, is the image of the tongue surface region R in the i-th frame, wherein the structural similarity index calculation formula is: , wherein, is the average brightness of the region in the i-th frame, is the brightness variance of the region in the i-th frame, is the covariance, and are constants, which can quantitatively describe whether the texture structure is stable, and pay more attention to local structure and texture integrity. The structural similarity index is used as a measure of the texture stability of the tongue surface region. By comparing the brightness, contrast and structural similarity of the region in consecutive image frames, the consistency change of the tongue surface texture during shooting is evaluated.

[0052] S23, the reflection of water vapor or saliva in the mouth can cause part of the region in the image to be blurred or to have bright spots, which affects the texture, color and detail texture recognition of the tongue coating. Based on the image sharpness and frequency domain response, the blurred part in the image region is identified, the high reflection part in the image region is detected, that is, the part with abnormally high local brightness value and low saturation, the bright spot part caused by saliva transpiration is identified combined with the heat map and near-infrared image, and the proportion of water vapor part is calculated.

[0053] In this embodiment, it needs to be specifically explained that a stable frame evaluation model is constructed, the tremor intensity, texture stability and water vapor proportion are introduced into the stable frame evaluation model to evaluate the stability of the image region, and the stable frame set is image registered, including the following specific steps:

[0054] S31, the tremor intensity, texture stability and water vapor region proportion are substituted into a comprehensive stability score calculation formula to evaluate the image stability, wherein the comprehensive stability score calculation formula is: , wherein a is a tremor sensitive adjustment parameter, which is used to control the weight of the influence of the shaking amplitude on the stability, H is the water vapor region proportion, which integrates the tongue tremor intensity, region texture stability and water vapor interference influence factors. The exponential decay function is used to simulate the nonlinear influence of tremor on image quality, and the multiplicative weighting is used to reflect the joint action of texture consistency and image occlusion on diagnostic availability.

[0055] S32, sort the same associated areas of different frames from high to low according to the comprehensive stability score, select the top k% (such as 30%) frames with the highest score as the relatively stable frame set, the frames in the set have the characteristics of small tongue tremor, complete texture structure and high image clarity, perform image registration on the stable frame set, verify the registration effect through the mean square deviation and structural similarity of the overlapping area, ensure that all stable frames have good spatial consistency in the key area, eliminate image misplacement caused by tongue tremor or changes in shooting angle, reduce the number of frames to be processed, reduce the calculation amount of registration and fusion, ensure the spatial consistency and texture integrity of image data through stable frames, and improve the accuracy of image-based tongue feature extraction and symptom recognition.

[0056] It needs to be specifically explained in the present embodiment that generating high-quality images for tongue image analysis based on image stable areas includes the following specific steps:

[0057] S41, perform image fusion based on the image stable area, perform feature matching and registration on the key points of consecutive frames, use affine transformation and perspective transformation to align the images, avoid ghosting and misplacement in the fusion process, use frame average fusion to improve the signal-to-noise ratio and clarity of the image, use wavelet enhancement to improve the texture contrast in low clarity areas, input the enhanced image into the tongue feature extraction and symptom matching process again, and mark the stability score and enhancement processing record of each area in the image metadata.

[0058] It needs to be explained here that the value of the various set parameters in the present embodiment is as follows: representative historical tongue image acquisition data is obtained, and patient historical symptom data is obtained, experts are invited to artificially judge the clarity of the tongue image, and the calculation results and judgment results of each step in the present embodiment are input into the fitting software, and the values of various set parameters that meet the highest judgment accuracy are output;

[0059] The benefits of the present embodiment relative to the prior art are as follows:

[0060] The present application matches the collected tongue image with the patient information, determines the symptom-associated tongue area, divides the priority of the associated tongue area, judges whether the collected image meets the clarity standard, constructs an image index evaluation model based on tongue tremor and image texture, evaluates the tongue tremor, texture stability and water vapor interference of the image to be processed, constructs a stable frame evaluation model, imports the tremor intensity, texture stability and water vapor proportion into the stable frame evaluation model to evaluate the image area stability, and performs image registration on the stable frame set, performs multi-frame registration and fusion based on the image stable area, records the area stability and processing information, generates high-quality images for tongue image analysis, improves the image clarity of the key diagnostic area, and improves the accuracy and stability of tongue image analysis.

[0061] Example 2

[0062] like Figure 4 As shown, a tongue surface diagnostic information acquisition, diagnosis, and analysis system is implemented based on the aforementioned tongue surface diagnostic information acquisition, diagnosis, and analysis method. Specifically, it includes a symptom association matching module, an image index evaluation module, a stable frame evaluation module, and an image processing module. The symptom association matching module matches the acquired tongue surface images with patient information, determines symptom-related tongue surface regions, prioritizes these regions, and determines whether the acquired images meet clarity standards. The image index evaluation module evaluates the images to be processed for tongue tremor, texture stability, and moisture blur interference. The stable frame evaluation module evaluates the stability of image regions through tremor intensity, texture stability, and moisture content, and performs image registration on the stable frame set. The image processing module performs multi-frame registration and fusion of stable image regions, records region stability and processing information, and generates high-quality images for tongue image analysis.

[0063] Example 3

[0064] This embodiment provides an electronic device, including: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0065] The processor executes the aforementioned method for collecting, diagnosing, and analyzing tongue surface diagnostic information by calling computer programs stored in memory.

[0066] The electronic device can vary considerably depending on its configuration or performance. It may include one or more Central Processing Units (CPUs) and one or more memories, wherein the memory stores at least one computer program, which is loaded and executed by the processor to implement the tongue surface diagnostic information acquisition and analysis method provided in the above-described embodiment. The electronic device may also include other components for implementing its functions; for example, it may have wired or wireless network interfaces and input / output interfaces for data input and output. Details will not be elaborated upon in this embodiment.

[0067] Example 4

[0068] This embodiment proposes a computer-readable storage medium on which an erasable and rewritable computer program is stored.

[0069] When a computer program runs on a computer device, it causes the computer device to perform the aforementioned method for collecting, diagnosing, and analyzing tongue surface diagnostic information.

[0070] For example, the computer readable storage medium can be read-only memory (ROM), random access memory (RAM), compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, optical data storage device, etc.

[0071] The above-described embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, the above-described embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions according to the embodiments of the present application are entirely or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through a wired network or / and a wireless network. The computer readable storage medium can be any available medium accessible by a computer or a data storage device such as a server, data center, etc. containing one or more available medium sets. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid state disk.

Claims

1. A method for collecting, diagnosing, and analyzing information from the tongue surface, characterized in that, It includes the following specific steps: The acquired tongue images are matched with patient information to determine the tongue regions associated with symptoms. These associated tongue regions are then prioritized, and the clarity and diagnostic importance of the image regions are used to determine whether the acquired images can be used for subsequent auxiliary diagnosis. If the images do not meet the requirements, image processing continues; if the images meet the requirements, they are directly used for symptom assessment. An image index evaluation model is constructed based on tongue tremor and image texture. The images to be processed are evaluated for tongue tremor, texture stability, and water vapor blur interference. The specific steps include: evaluating tongue tremor by analyzing the positional changes of key points in the tongue surface region between consecutive frames in the image sequence; extracting the coordinate positions of the associated regions in the consecutive frame images; and substituting these coordinate positions into the tongue tremor intensity calculation formula to evaluate the average offset of key points in the time series. The tongue tremor intensity calculation formula is as follows: ,in, Let be the coordinates of the key points in the region of the i-th frame image. The average position coordinates of the key points in the entire sequence. Let be the Euclidean distance between vectors, and n be the total number of frames. Extract the tongue-like texture features of each region, and calculate the texture consistency between consecutive frames using the texture stability calculation formula. The texture stability calculation formula is as follows: ,in, This is an image of the tongue region R in the i-th frame. Let be the structural similarity index between frame i and frame i+1, where the structural similarity index is calculated using the following formula: ,in, The average brightness of the region in the i-th frame. Let V be the region brightness variance in the i-th frame. For covariance, and As a constant, it can quantitatively describe whether the texture structure is stable. Based on texture sharpness and frequency domain response, it can detect image clarity, identify blurred parts in the image region, detect high reflectivity parts in the image region, and combine thermal images and near-infrared images to identify bright spots caused by saliva evaporation and calculate the proportion of water vapor. A stable frame evaluation model is constructed, and jitter intensity, texture stability, and water vapor content are imported into the stable frame evaluation model to evaluate the stability of image regions. Image registration is then performed on the stable frame set. Multi-frame registration and fusion are performed based on stable regions of the image to record regional stability and processing information, generating high-quality images for tongue image analysis.

2. The method for collecting, diagnosing, and analyzing tongue surface diagnostic information as described in claim 1, characterized in that, The process of matching the acquired tongue images with patient information to determine the tongue regions associated with symptoms, prioritizing these regions, and determining whether the acquired images can be used for subsequent auxiliary diagnosis based on image clarity and diagnostic importance weights, involves the following specific steps: S11. Continuously acquire video frame sequences of the tongue surface, and simultaneously obtain thermal imaging images and near-infrared spectral images. Obtain the patient's clinical complaint information and previous medical records. Use natural language processing technology to extract diagnostic keywords and key symptom information from the patient's clinical complaint information and previous medical records. Perform keyword matching with the TCM tongue image symptom database to preliminarily determine the range of tongue surface areas associated with the patient's symptoms. S12. Use the trained tongue segmentation network model to segment the acquired tongue surface image to form a standardized tongue surface contour. Use the longitudinal central axis and the transverse width central axis of the tongue surface geometry as the segmentation benchmark to construct a standardized rectangular coordinate system and divide the tongue body into the tongue tip area, tongue middle area, tongue side area and tongue root area. S13. Match the key tongue regions corresponding to the symptom keywords identified in S11 with the standard tongue regions segmented in S12. Assign diagnostic importance weights to multiple tongue regions associated with symptoms based on historical data. The diagnostic importance weight represents the degree of contribution of the region to the diagnosis of the current symptom. Normalize the fused frames of the tongue regions by processing Laplacian variance and high-frequency energy ratio to form a unified clarity score. Set an image quality judgment threshold and a region diagnostic importance threshold. If the diagnostic contribution is higher than the region diagnostic importance threshold and the region clarity score is higher than the image quality judgment threshold, the image is directly used for subsequent diagnosis. If the diagnostic contribution is higher than the region diagnostic importance threshold and the region clarity score is lower than the image quality judgment threshold, the image undergoes image processing. If the diagnostic contribution is lower than the region diagnostic importance threshold and the region clarity score is higher than the image quality judgment threshold, the image region can be used for subsequent auxiliary diagnosis.

3. The method for collecting, diagnosing, and analyzing tongue surface diagnostic information as described in claim 2, characterized in that, The construction of the stable frame evaluation model, which incorporates jitter intensity, texture stability, and water vapor content into the stable frame evaluation model to evaluate the stability of image regions, and the image registration of the stable frame set includes the following specific steps: S21. Substitute the jitter intensity, texture stability, and water vapor region proportion into the comprehensive stability score calculation formula to evaluate image stability. The comprehensive stability score calculation formula is as follows: Where α is the vibration-sensitive adjustment parameter and H is the proportion of the water vapor region; S22. Sort the same associated region in different frames from high to low according to the comprehensive stability score, select the frame with the highest k% score as the relatively stable frame set, perform image registration on the stable frame set, and eliminate image misalignment caused by tongue shaking or changes in shooting angle.

4. The method for collecting, diagnosing, and analyzing tongue surface diagnostic information as described in claim 3, characterized in that, The process of multi-frame registration and fusion based on image stability regions, recording region stability and processing information, and generating high-quality images for tongue image analysis includes the following specific steps: Image fusion is performed based on stable regions of the image. Feature matching and registration are performed on key points of consecutive frames. Affine transformation and perspective transformation are used to align the images. Frame averaging fusion is used to improve the signal-to-noise ratio and sharpness of the image. Wavelet enhancement is used to improve texture contrast in low-resolution regions. The enhanced images are then re-input into the tongue image feature extraction and symptom matching process. The stability score of each region and the enhancement processing record are marked in the image metadata.

5. A tongue surface diagnostic information acquisition, diagnosis, and analysis system, implemented based on the tongue surface diagnostic information acquisition, diagnosis, and analysis method as described in any one of claims 1-4, characterized in that, Specifically, it includes: The symptom association matching module is used to match the acquired tongue images with patient information, determine the symptom-associated tongue regions, prioritize the associated tongue regions, and determine whether the acquired images can be used for subsequent auxiliary diagnosis based on the clarity and diagnostic importance weight of the image regions. If the image does not meet the requirements, image processing continues; if the image meets the requirements, it is directly used for symptom assessment. The image index evaluation module is used to evaluate the tongue tremor, texture stability and water vapor blur interference of the image to be processed; The stable frame evaluation module is used to evaluate the stability of image regions by jitter intensity, texture stability and water vapor content, and to perform image registration on the stable frame set; The image processing module is used to perform multi-frame registration and fusion of stable regions in the image, record the region stability and processing information, and generate high-quality images for tongue image analysis.

6. An electronic device, comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; The processor is characterized in that it executes a tongue surface diagnostic information acquisition and analysis method as described in any one of claims 1-4 by calling a computer program stored in the memory.

7. A computer-readable storage medium, characterized in that, The device stores instructions that, when executed on a computer, cause the computer to perform a method for collecting, diagnosing, and analyzing tongue surface diagnostic information as described in any one of claims 1-4.

Images