Tongue diagnosis image intelligent storage search method and system and readable storage medium

Abstract

The application discloses a tongue diagnosis image intelligent storage search method, system and readable storage medium, the method comprises the following steps: extracting the characteristic quantity in each first tongue diagnosis image and its numerical value, and storing the characteristic quantity in association with the first tongue diagnosis image; a plurality of numerical values corresponding to each first tongue diagnosis image are combined to form a plurality of numerical combinations, and the numerical combinations are stored in association with the corresponding first tongue diagnosis image; the characteristic quantity in the second tongue diagnosis image is obtained and is stored in association with the second tongue diagnosis image after being numerical; each numerical value corresponding to the second tongue diagnosis image is combined to form a plurality of search arrays; the numerical combination consistent with the search array is found, and the corresponding first tongue diagnosis image is obtained; the characteristic quantity in the tongue diagnosis image is automatically identified and numerical, which effectively improves the search efficiency of the tongue diagnosis image; the numerical value corresponding to each different characteristic quantity is combined as the basis for searching, which can further improve the search efficiency, and the search result can be directly associated with the disease data and output.

Description

Technical Field

[0001] This invention relates to the field of data storage and search technology, and more specifically, to a method, system, and readable storage medium for intelligent storage and search of tongue diagnosis images. Background Technology

[0002] Tongue diagnosis is a simple and effective method that uses observation of changes in the color and shape of a patient's tongue to aid in diagnosis and differentiation of diseases. Experienced physicians can make accurate judgments about diseases based on images of the patient's tongue combined with their experience accumulated over a long period of work.

[0003] With the continuous improvement of big data and image algorithms, automated diagnosis and treatment systems based on tongue diagnosis images are emerging in large numbers. Their working principle is typically set as follows: multiple tongue diagnosis images are collected, then manually categorized and associated with specific diseases, forming a database. Later, during automated diagnosis, the patient's tongue diagnosis image is directly collected, and image recognition software automatically matches similar tongue diagnosis images and corresponding diseases in the database. While this process is simple, the numerous features of tongue diagnosis images and the complex associated diseases often lead to incorrect image recognition but inaccurate disease associations, resulting in low accuracy of automated diagnosis. Furthermore, as the number of tongue diagnosis images in the database increases, the speed of finding similar images during comparison slows down, affecting the efficiency of tongue diagnosis. Summary of the Invention

[0004] To address the problem of the inability to quickly and accurately obtain similar tongue diagnosis images and their corresponding symptoms in practical applications, this application aims to propose an intelligent storage and search method for tongue diagnosis images. This method not only enables rapid retrieval of similar tongue diagnosis images from a database based on a specific image, but also provides more accurate diagnoses and their medical records, thus improving the accuracy of diagnostic results. The second objective is to provide an intelligent tongue diagnosis image search system. The third objective is to propose a readable storage medium loaded with a program for implementing the aforementioned intelligent storage and search method for tongue diagnosis images. The specific solution is as follows:

[0005] A method for intelligent storage and retrieval of tongue diagnosis images includes:

[0006] The feature quantities in each first tongue diagnosis image are extracted and quantified according to the numerical model. Then, they are associated with and stored with the corresponding first tongue diagnosis image.

[0007] Based on the diagnostic model, multiple numerical values ​​corresponding to each of the first tongue diagnosis images are combined to form multiple numerical combinations to characterize the disease. Then, the multiple numerical combinations are associated with and stored with their corresponding first tongue diagnosis images.

[0008] The feature quantities in the second tongue diagnosis image are obtained and each feature quantity is quantified according to the numerical model. Then, they are associated with and stored in the corresponding second tongue diagnosis image.

[0009] Based on the diagnostic model, the numerical values ​​corresponding to the second tongue diagnosis image are combined to form multiple arrays to be checked;

[0010] Find the numerical combination that matches the array to be searched and retrieve the corresponding first tongue diagnosis image.

[0011] The above technical solution abandons the existing approach of searching for similar tongue diagnosis images based on a single feature. When storing and searching tongue diagnosis images, the feature values ​​of each tongue diagnosis image are combined according to the symptoms they represent, and then stored and searched in the form of numerical combinations. This can greatly improve the efficiency of searching and searching, and at the same time, it can quickly obtain tongue diagnosis images and their corresponding symptom data.

[0012] Furthermore, the features include one or more combinations of tongue coating LAB, tongue coating RGB, tongue body LAB, tongue body RGB, tongue coating thickness, and tongue body slenderness.

[0013] Furthermore, the diagnostic model stores the correlation between various feature quantities and their combinations and various disease data;

[0014] The diagnostic model combines multiple numerical values ​​corresponding to each of the first tongue diagnosis images to form multiple numerical combinations used to characterize the disease, including:

[0015] The numerical values ​​corresponding to multiple feature quantities in the first tongue diagnosis image are sorted to form a first sequence;

[0016] Select the first value in the first sequence as the search basis, and search for its corresponding disease data in the self-diagnosis model;

[0017] If corresponding disease data exists, the current value is stored as a single value combination. Then, the first value is combined with the second value of the next digit and the value combination is used as the search basis until the first value is combined with each value in the first sequence. If the value combination has corresponding disease data, each value combination is stored.

[0018] If no corresponding disease data is found for the first numerical value in the first sequence and its combination with the remaining numerical values, then the next numerical value in the first sequence is selected as the search basis, and the aforementioned steps are repeated until all numerical values ​​in the first sequence have been searched and combined to obtain the numerical combination corresponding to the first tongue diagnosis image.

[0019] Through the above technical solution, the method of searching for a first tongue diagnosis image based on a second tongue diagnosis image has changed from feature comparison to numerical combination comparison, thus improving search efficiency. Furthermore, since each numerical combination is associated with a specific symptom, obtaining the numerical combination allows for the acquisition of related symptom data, which is beneficial for improving subsequent diagnostic efficiency.

[0020] Furthermore, the multiple combinations of the stated values ​​are associated and stored with their corresponding first tongue diagnosis images, including:

[0021] Determine the number of numerical values ​​contained in the numerical combination, and store and number each numerical combination sequentially based on the number of numerical values ​​to form a feature vector that represents each numerical combination.

[0022] The above technical solution allows the numerical combination itself to be used as a feature vector, making the subsequent search more convenient and faster. At the same time, since the output is only possible when the feature values ​​of the first and second tongue diagnosis images are mostly the same or similar, sorting the numerical combinations according to the number of numerical values ​​they contain ensures that the first tongue diagnosis images obtained in the early stages of the search are images with feature values ​​that are mostly the same or similar to those of the second tongue diagnosis images, thus improving the accuracy of the search and enabling the target tongue diagnosis image to be obtained in the shortest possible time.

[0023] Further, finding the numerical combination that matches the array to be searched and retrieving the corresponding first tongue diagnosis image includes:

[0024] Sort the arrays according to the number of values ​​in each array to form a second sequence;

[0025] Retrieve the array with the largest number of values ​​and use it as a comparison reference to obtain the stored identical value combinations and retrieve the corresponding first tongue diagnosis image.

[0026] If there is no numerical combination that is the same as the array to be checked, the next array to be checked in the second column is used as the comparison reference to obtain the stored numerical combination that is the same, and retrieve the corresponding first tongue diagnosis image, until all the arrays to be checked contained in the second column have been compared.

[0027] The first tongue diagnosis images obtained using the above method are sorted and output.

[0028] By using the above technical solution, the search array with the largest number of values ​​is used as the basic comparison data for the initial search. This allows for the rapid and accurate identification of the first tongue diagnosis image that is similar to the second tongue diagnosis image. At the same time, the first tongue diagnosis image is automatically sorted during the output process, meaning that the search is performed first and the output is performed first, which is more efficient in the subsequent data presentation.

[0029] Furthermore, finding the numerical combination that matches the array to be searched and retrieving the corresponding first tongue diagnosis image also includes:

[0030] Based on the sorting of the first tongue diagnosis images, all feature quantities of the first tongue diagnosis images with a predetermined number of positions are extracted, and they are compared with all feature quantities corresponding to the second tongue diagnosis images to obtain feature similarity data between each first tongue diagnosis image and the second tongue diagnosis image.

[0031] Calculate the average of multiple similarity data;

[0032] The average value is then bound to the first tongue diagnosis image obtained and output.

[0033] With the above technical solution, when the first tongue diagnosis image sequence is obtained, the similarity value between each first tongue diagnosis image and the second tongue diagnosis image can be seen intuitively, which facilitates the selection and judgment of subsequent data and provides a reference for the accuracy of disease diagnosis.

[0034] A smart storage and search system for tongue diagnosis images includes:

[0035] The feature extraction module is configured to extract feature values ​​from tongue diagnosis images.

[0036] The numericalization module is data-connected to the feature extraction module and is configured to quantify the feature quantities of the tongue diagnosis image based on a specific numericalization model to generate numerical quantities.

[0037] The numerical combination module is data-connected to the numericalization module, receives the numerical values ​​output by the numericalization module, and combines them into numerical combinations or arrays to be checked based on a specific diagnostic model and outputs them.

[0038] The storage module is configured to store the first tongue diagnosis image for comparison and search, as well as its corresponding feature quantity, numerical quantity, and numerical combination;

[0039] The query output module is configured to receive the second tongue diagnosis image and convert it into a query array, compare it with the numerical combinations in the storage module, and find and output the first tongue diagnosis image corresponding to the same numerical combination.

[0040] The above technical solution first quantifies the first tongue diagnosis image used for search and comparison, and then combines different numerical values ​​according to the disease data. Using the numerical combination as the basis for comparison can greatly improve the search efficiency, avoid invalid search results, and quickly obtain the first tongue diagnosis image and its disease data corresponding to the second tongue diagnosis image.

[0041] Furthermore, the feature extraction module includes an image recognition algorithm and stored feature images for recognition;

[0042] The numerical combination module is equipped with a numerical combination algorithm and a sorting algorithm, which are used to combine different numerical quantities in a set order and then output them sequentially.

[0043] The above technical solution can first output a tongue diagnosis image with high similarity, which helps to improve the search speed.

[0044] Furthermore, the system also includes:

[0045] The ratio calculation module is configured to calculate the similarity of feature quantities between each first tongue diagnosis image and the second tongue diagnosis image, and output the similarity data corresponding to each first tongue diagnosis image.

[0046] The binding output module is connected to the query output module and configured to combine the similarity data with its corresponding first tongue diagnosis image for output.

[0047] The above technical solution enables the similarity labeling of each first tongue diagnosis image obtained through the search, which facilitates the subsequent evaluation of the reliability of the automatic diagnosis results.

[0048] A computer-readable storage medium loaded with a program algorithm for implementing the intelligent storage and search method for tongue diagnosis images as described above.

[0049] Compared with the prior art, the beneficial effects of this application are as follows:

[0050] By automatically identifying and quantifying the features in tongue diagnosis images, the search efficiency of tongue diagnosis images is effectively improved. Furthermore, combining the numerical values ​​corresponding to different features and using these combinations as the basis for the search further enhances search efficiency. Moreover, since the numerical combinations are associated with disease data, invalid search results resulting from simple feature comparisons can be avoided—for example, the searched features may not correspond to any disease data. This improves both search efficiency and accuracy, and the search results can be directly associated with disease data for output. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the overall process of the method in this application;

[0052] Figure 2 A schematic diagram illustrating the method for obtaining numerical combinations;

[0053] Figure 3 A schematic diagram illustrating the method for obtaining the first tongue diagnosis image;

[0054] Figure 4 This is a schematic diagram of the functional modules of the intelligent storage search system of this application.

[0055] Figure reference numerals: 100, Feature extraction module; 200, Numericalization module; 300, Numerical combination module; 400, Storage module; 500, Query output module. Detailed Implementation

[0056] The present application will be further described in detail below with reference to the embodiments and figures, but the implementation of the present application is not limited thereto.

[0057] A method for intelligent storage and retrieval of tongue diagnosis images, such as Figure 1 As shown, the main steps include the following:

[0058] Storage steps:

[0059] S100: Extract the feature quantities from each first tongue diagnosis image and quantify each feature quantity according to the numerical model, and then associate and store them with the corresponding first tongue diagnosis image.

[0060] S200: Based on the diagnostic model, multiple numerical values ​​corresponding to each first tongue diagnosis image are combined to form multiple numerical combinations used to characterize the disease. Then, the multiple numerical combinations are associated with and stored with their corresponding first tongue diagnosis images.

[0061] Search steps:

[0062] S300: Obtain the feature quantities in the second tongue diagnosis image and quantify each feature quantity according to the numerical model, and then associate and store it with the corresponding second tongue diagnosis image.

[0063] S400, based on the diagnostic model, combines the numerical values ​​corresponding to the second tongue diagnosis image to form multiple arrays to be checked;

[0064] S500: Find the numerical combination that matches the array to be searched and retrieve the corresponding first tongue diagnosis image.

[0065] In the above storage steps, the first tongue diagnosis image is an existing tongue diagnosis image stored in the system database for retrieval and comparison. In the search step, the second tongue diagnosis image is the tongue diagnosis image to be diagnosed, which can be obtained on-site through an image acquisition device, such as the camera inside a tongue diagnosis instrument.

[0066] The feature values ​​corresponding to the first tongue diagnosis image used for comparison are quantified to simplify storage and facilitate later retrieval. For example, the color and texture of the tongue coating can be combined to more comprehensively judge the cold or heat nature of the disease and the depth of the pathogenic factors. Yellow and greasy tongue coating: yellow is a reflection of heat, and greasy is a symbol of dampness. Therefore, the disease indicated by yellow and greasy tongue coating is called "damp heat" in traditional Chinese medicine. If the value "010" is used to represent a pale yellow tongue coating and "011" is used to represent a thin tongue coating, then the combination of the values ​​"010011" can represent the symptoms of "damp heat".

[0067] In this embodiment of the application, the features of the tongue diagnosis image include one or more combinations of tongue coating LAB, tongue coating RGB, tongue body LAB, tongue body RGB, tongue coating thickness, and tongue body slenderness. The above features can be extracted by image recognition, or in the feature embodiment, they can be obtained by manual input.

[0068] As mentioned above, since a disease often corresponds to multiple feature quantities, in this application, step S200 is to combine the feature quantities of a first tongue diagnosis image after numericalization to characterize the various diseases corresponding to the first tongue diagnosis image.

[0069] In detail, in the embodiments of this application, the diagnostic model stores the correlation between various feature quantities and their combinations and various disease data. These correlations link the various feature quantities and their combinations to the diseases and are stored in the system database as an analytical formula model for easy retrieval and use.

[0070] In step S200, based on the diagnostic model, multiple numerical values ​​corresponding to each first tongue diagnosis image are combined to form multiple numerical combinations used to characterize the disease, such as... Figure 2 As shown, it includes:

[0071] S210, sort the numerical values ​​corresponding to multiple feature quantities in the first tongue diagnosis image to form a first sequence;

[0072] S220, Select the first value in the first sequence as the search basis, and search for the corresponding disease data in the self-diagnosis model;

[0073] S221, if there is corresponding disease data, the current value is stored as a single value combination. Then, the first value is combined with the second value of the next position and the value combination is used as the search basis until the first value is combined with each value in the first sequence. If the value combination has corresponding disease data, each value combination is stored.

[0074] S222, if there is no corresponding disease data for the first numerical value in the first sequence and its combination with the remaining numerical values, then select the next numerical value in the first sequence as the search basis, and repeat steps S220~S222 until all numerical values ​​in the first sequence have been searched and combined, and obtain the numerical combination corresponding to the first tongue diagnosis image.

[0075] In practice, since not all features of each tongue diagnosis image correspond to specific disease data, and many disease data require combinations of multiple features for confirmation, the method for searching the first tongue diagnosis image based on the second tongue diagnosis image in the above technical solution has changed from comparing features to comparing numerical combinations, thus improving search efficiency. Furthermore, since each numerical combination is associated with a specific disease, obtaining the numerical combination allows for the acquisition of related disease data, which is beneficial for improving subsequent diagnostic efficiency and avoiding invalid searches.

[0076] In step S200, multiple numerical combinations are associated and stored with their corresponding first tongue diagnosis images, including:

[0077] S230, determine the number of numerical values ​​contained in the numerical combination, and store and number each numerical combination in sequence based on the number of numerical values ​​to form a feature vector to represent each of the above numerical combinations.

[0078] When multiple features in the second tongue diagnosis image are the same as those in the first tongue diagnosis image, the probability that the corresponding disease data are the same is relatively high. Based on the technical solution of step S230 above, the first tongue diagnosis image with a high degree of similarity can be placed at the forefront, which helps to quickly search for the target image.

[0079] Similarly, in step S500, after the feature values ​​of the first tongue diagnosis image are quantified, the array to be checked is also sorted according to the number of value values.

[0080] In step S500, the numerical combination that matches the array to be searched is found and the corresponding first tongue diagnosis image is retrieved, such as... Figure 3 As shown, it includes:

[0081] S510, Sort the arrays to be checked according to the number of values ​​in each array to be checked, forming a second sequence;

[0082] S520, retrieve the array to be searched with the largest number of values ​​and use it as a comparison reference to obtain the stored identical value combinations and retrieve the corresponding first tongue diagnosis image;

[0083] S521, if there is no numerical combination that is the same as the array to be checked, then the next array to be checked in the second column is used as the comparison reference to obtain the stored numerical combination that is the same, and retrieve the corresponding first tongue diagnosis image, until all the arrays to be checked contained in the second column have been compared.

[0084] S530, sort and output the first tongue diagnosis image obtained according to the above method.

[0085] The above technical solution uses the search array with the most numerical values ​​as the basic comparison data for the initial search, which can quickly and accurately find the first tongue diagnosis image similar to the second tongue diagnosis image. At the same time, it automatically sorts the first tongue diagnosis image during the output process, that is, it searches first and outputs first, which is more efficient in subsequent data presentation.

[0086] In the optimized step S500, finding the numerical combination that matches the array to be searched and retrieving the corresponding first tongue diagnosis image also includes:

[0087] S540, based on the sorting of the first tongue diagnosis images, extract all feature quantities of the first tongue diagnosis images with a predetermined number of positions, compare them with all feature quantities corresponding to the second tongue diagnosis images, and obtain feature similarity data between each first tongue diagnosis image and the second tongue diagnosis image.

[0088] S550 calculates the average of multiple similarity data;

[0089] S560, the average value is bound to the acquired first tongue diagnosis image and output.

[0090] The above steps are actually search and verification steps. By comparing the features of a predetermined number of first tongue diagnosis images obtained from the search, similarity data is obtained. This allows for an overall assessment of the search accuracy, facilitating subsequent data selection and evaluation, and providing a reference for the accuracy of disease diagnosis. The aforementioned similarity data can be obtained based on the proportion of identical features in the first and second tongue diagnosis images.

[0091] To realize the above-mentioned intelligent storage and search method for tongue diagnosis images, this application proposes an intelligent storage and search system for tongue diagnosis images, such as... Figure 4 As shown, it mainly includes a feature extraction module 100, a numericalization module 200, a numerical combination module 300, a storage module 400, and a query output module 500.

[0092] The feature extraction module 100 is configured to extract feature quantities from a tongue diagnosis image, and to extract feature quantities from a first tongue diagnosis image and a second tongue diagnosis image. The feature extraction module 100 includes an image recognition algorithm and a stored feature image for recognition, and extracts relevant feature quantities from a tongue diagnosis image.

[0093] The numericalization module 200 is data-connected to the feature extraction module 100 and is configured to quantify the feature quantities of the tongue diagnosis image based on a specific numericalization model, generating numerical values. The aforementioned numericalization model is the association model between each feature quantity and its numerical value, used to convert each graphical feature quantity into a numerical value, facilitating subsequent combination and comparison searches.

[0094] The numerical combination module 300 is data-connected to the numericalization module 200. As described in the aforementioned method step S200, it receives the numerical values ​​output by the numericalization module 200 and combines them into numerical combinations or arrays to be checked based on a specific diagnostic model, and then outputs them.

[0095] The aforementioned feature extraction module 100, numericalization module 200, and numerical combination module 300 are all executed through the data processing motherboard in the server, and the corresponding algorithm models can be stored in a database connected to the server.

[0096] The storage module 400 is configured to associate and store a first tongue diagnosis image and its corresponding feature values, numerical values, and numerical combinations for comparative searching. In a specific embodiment, it can also be associated with disease data for storage, facilitating comparison and retrieval during later searches. In practical applications, the storage module 400 can be configured as a database connected to a server.

[0097] The query output module 500 is configured to receive the second tongue diagnosis image, convert it into a query array, compare it with the numerical combinations in the storage module 400, and find and output the first tongue diagnosis image corresponding to the same numerical combination. In practical applications, the query output module 500 also includes a human-computer interaction component for inputting the second tongue diagnosis image information to be queried and outputting the first tongue diagnosis image information obtained from the query.

[0098] In order to output the first tongue diagnosis image with high similarity first and improve the search speed, the numerical combination module 300 is equipped with a numerical combination algorithm and a sorting algorithm, which are used to combine different numerical quantities in a set order and then output them sequentially. The sorting is based on the number of numerical quantities contained in each numerical combination or the array to be searched.

[0099] In order to enable similarity labeling for each first tongue diagnosis image obtained from the search, and to facilitate the subsequent evaluation of the reliability of the automatic diagnosis results, the system also includes a ratio calculation module and a binding output module.

[0100] The ratio calculation module is configured to calculate the similarity of feature quantities between each first tongue diagnosis image and a second tongue diagnosis image, and output similarity data corresponding to each first tongue diagnosis image. In one embodiment, the system retrieves a predetermined number of first tongue diagnosis images ranked first and extracts their corresponding feature quantities, compares them with the second tongue diagnosis images, and compares them based on the feature quantities rather than the array to be searched. Similarity data is obtained based on the number of identical feature quantities and the number of feature quantities corresponding to the second tongue diagnosis images. If multiple first tongue diagnosis images are involved in the comparison, the average of the multiple similarity data is taken.

[0101] The binding output module is data-connected to the query output module 500 and configured to output the similarity data in combination with its corresponding first tongue diagnosis image. Thus, the user can intuitively see the first tongue diagnosis image similar to the second tongue diagnosis image, as well as the similarity data. Usually, there are multiple first tongue diagnosis images.

[0102] To expand the application of the method, this application also proposes a computer-readable storage medium loaded with a program algorithm for implementing the intelligent storage and search method for tongue diagnosis images as described above, including a computer hard disk, optical disk, or SD card.

[0103] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A tongue image intelligent storage search method, characterized in that, The method comprises the following steps: extracting feature quantities in each first tongue diagnosis image and numerically converting each feature quantity according to a numerical conversion model, and then storing each feature quantity in association with the corresponding first tongue diagnosis image; combining multiple numerical quantities corresponding to each first tongue diagnosis image based on a diagnosis model to form multiple numerical combinations for representing diseases, and then storing the multiple numerical combinations in association with the corresponding first tongue diagnosis image; extracting feature quantities in a second tongue diagnosis image and numerically converting each feature quantity according to a numerical conversion model, and then storing each feature quantity in association with the corresponding second tongue diagnosis image; combining multiple numerical quantities corresponding to each second tongue diagnosis image based on a diagnosis model to form multiple arrays to be searched; finding a numerical combination consistent with the array to be searched and calling the corresponding first tongue diagnosis image; the diagnosis model stores the association between each feature quantity and its combination and each disease data; the method of combining multiple numerical quantities corresponding to each first tongue diagnosis image based on a diagnosis model to form multiple numerical combinations for representing diseases comprises: sorting the numerical quantities corresponding to multiple feature quantities in the first tongue diagnosis image to form a first sequence; selecting a first numerical quantity in the first sequence as a search basis to search for corresponding disease data from the diagnosis model; if there is corresponding disease data, the current numerical quantity is stored as a single numerical combination, and then the first numerical quantity is combined with the second numerical quantity next to it, and the numerical combination is used as a search basis until the combination of the first numerical quantity with each numerical quantity in the first sequence is completed, and if the numerical combination has corresponding disease data, each numerical combination is stored; if the first numerical quantity in the first sequence and its combination with the remaining numerical quantities do not have corresponding disease data, the next numerical quantity in the first sequence is selected as a search basis, and the above steps are repeated until the numerical quantities in the first sequence are searched and combined, and the numerical combination corresponding to the first tongue diagnosis image is obtained; storing the multiple numerical combinations in association with the corresponding first tongue diagnosis image comprises: determining the number of numerical quantities contained in the numerical combination, and storing each numerical combination in sequence based on the number of numerical quantities and numbering to form a feature vector for representing each numerical combination; finding a numerical combination consistent with the array to be searched and calling the corresponding first tongue diagnosis image comprises: sorting each array to be searched according to the number of numerical quantities in each array to form a second sequence; calling the array to be searched with the most numerical quantities as a comparison reference, obtaining the same numerical combination that has been stored, and calling the corresponding first tongue diagnosis image; if there is no numerical combination identical to the array to be searched, the next array to be searched in the second sequence is used as a comparison reference to obtain the same numerical combination that has been stored, and the corresponding first tongue diagnosis image is called until all the arrays to be searched contained in the second sequence are compared; sorting and outputting the first tongue diagnosis image obtained according to the above method.

2. The tongue image intelligent storage search method according to claim 1, wherein, The feature quantities include one or more combinations of tongue coating LAB, tongue coating RGB, tongue texture LAB, tongue texture RGB, tongue coating thickness, and tongue body fatness.

3. The tongue image intelligent storage search method according to claim 1, characterized in that, The method for searching the numerical combination consistent with the to-be-searched array and calling the first tongue diagnosis image corresponding to the numerical combination further comprises: Based on the sorting of the first tongue diagnosis image, extracting all feature quantities of the first tongue diagnosis image of the front set position, comparing them with all feature quantities corresponding to the second tongue diagnosis image, and obtaining feature similarity data of each first tongue diagnosis image and the second tongue diagnosis image; Calculating the average value of the plurality of similarity data; Binding and outputting the average value and the obtained first tongue diagnosis image.

4. A tongue image intelligent storage search system, characterized in that, It comprises: A feature extraction module (100) configured to extract feature quantities of tongue diagnosis images, and configured to extract feature quantities of the first tongue diagnosis image and the second tongue diagnosis image; A numerical module (200) connected with the feature extraction module (100) and configured to numerically value the feature quantities of the tongue diagnosis image based on a numerical value model, and generate numerical quantities; A numerical combination module (300) connected with the numerical module (200) and receiving the numerical quantities output by the numerical module (200) and combining them into numerical combinations or to-be-searched arrays based on a diagnosis model and outputting them; The diagnosis model stores the correlation between each feature quantity and its combination and each disease data; The diagnosis model stores the correlation between each feature quantity and its combination and each disease data; The method for searching the numerical combination consistent with the to-be-searched array and calling the first tongue diagnosis image corresponding to the numerical combination further comprises: Sorting the numerical quantities corresponding to the plurality of feature quantities in the first tongue diagnosis image to form a first sequence; Selecting the first numerical quantity in the first sequence as the searching basis to search for the corresponding disease data from the diagnosis model; If there is corresponding disease data, the current numerical quantity is stored as a single numerical combination, and then the first numerical quantity and the second numerical quantity behind it are combined, and the numerical combination is used as the searching basis until the combination of the first numerical quantity and each numerical quantity in the first sequence is completed, and if the numerical combination has corresponding disease data, each numerical combination is stored; If the first numerical quantity in the first sequence and its combination with the remaining numerical quantities do not have corresponding disease data, the next numerical quantity in the first sequence is selected as the searching basis, and the foregoing steps are repeated until the combination of the numerical quantities in the first sequence is completed, and the numerical combination corresponding to the first tongue diagnosis image is obtained; A storage module (400) configured to store the first tongue diagnosis image and its corresponding feature quantity, numerical quantity, and numerical combination for comparison and search; Storing the plurality of numerical combinations and their corresponding first tongue diagnosis images, comprising: Judging the number of numerical quantities in the numerical combination, sequentially storing and numbering each numerical combination based on the number of numerical quantities to form a feature vector representing each numerical combination; A query output module (500) configured to receive the second tongue diagnosis image and convert it into a to-be-searched array, compare it with the numerical combinations in the storage module (400), search for and output the first tongue diagnosis image corresponding to the same numerical combination; The method for searching the numerical combination consistent with the to-be-searched array and calling the first tongue diagnosis image corresponding to the numerical combination further comprises: According to the number of numerical quantities in each to-be-searched array, sorting each to-be-searched array to form a second sequence; The number of values of the array to be searched is the most, and the same number combination is obtained by taking it as a comparison reference, and the corresponding first tongue diagnosis image is obtained; If there is no same number combination with the array to be searched, the next array to be searched in the second number sequence is taken as a comparison reference, the same number combination is obtained, and the corresponding first tongue diagnosis image is obtained until the arrays to be searched in the second number sequence are all compared. The first tongue diagnosis images obtained are sorted and output.

5. The tongue image intelligent storage search system according to claim 4, wherein, The feature extraction module (100) includes an image recognition algorithm and its stored feature images for recognition; The number combination module (300) is configured with a number combination algorithm and a sorting algorithm, which is used to combine different numerical values in a set order and then output them in order.

6. The tongue image intelligent storage search system according to claim 4, wherein, It also includes: The proportion calculation module is configured to calculate the similarity of the feature quantities of each first tongue diagnosis image and second tongue diagnosis image, and output the similarity data corresponding to each first tongue diagnosis image; The binding output module is connected with the query output module (500) data, and is configured to output the similarity data combined with the corresponding first tongue diagnosis image.

7. A computer-readable storage medium, characterized in that, Load the program algorithm for realizing the tongue diagnosis image intelligent storage search method as claimed in any one of claims 1-3.

Images