AI-based automated acupoint identification and positioning system and method

Abstract

The present invention discloses an AI-based automatic acupoint recognition and positioning system and method, the system comprises an AI server, an image acquisition device, a body shape recognition and comparison module, an AI acupoint recognition module and an acupoint projection module; the AI ​​server is used to establish at least one sample model according to a pre-collected three-dimensional image of a human body shape, wherein the three-dimensional image of the human body shape contains acupoint marking points of various parts of the human body; the image acquisition device is used to collect a three-dimensional target image of the patient's body shape when the positions of various acupoints on the patient's body need to be positioned; the body shape recognition and comparison module is used to compare the three-dimensional target image with multiple three-dimensional images of human body shapes of at least one sample model in the AI ​​server to determine whether the three-dimensional target image can match the size ratio of any three-dimensional image of human body shape; the AI ​​acupoint recognition module uses the matched three-dimensional image of human body shape as a reference to determine the positions of acupoints of various parts of the patient's body; the acupoint projection module is used to project the acupoint marking points of various parts of the human body displayed by the matched three-dimensional image of human body shape to various positions on the patient's body skin surface. Effect: AI technology can quickly and accurately locate acupoints, can adapt to different groups of people, can realize the visualization and intuition of acupoint positioning, and can more easily help people who need to quickly and accurately locate acupoints.

Description

Technical Field

[0001] The present invention relates to the technical field of acupoint positioning, and in particular to an AI-based automated acupoint recognition and positioning system and method. Background Art

[0002] As the pace of modern life continues to accelerate, people are often in a sub-healthy state. Acupuncture and massage, as conventional treatment methods of traditional Chinese medicine, are widely favored by people due to their convenient operation, few side effects and quick results.

[0003] However, users generally do not have basic knowledge of traditional Chinese medicine. Whether it is a doctor or a patient, it is difficult for them to quickly and accurately confirm the exact location of the body's acupuncture points. The acupuncture point positioning methods on the market only stay at text descriptions or directly simulate operations on dummies, or are only displayed on human body model pictures for reference.

[0004] However, for different people, due to the influence of height, weight, fatness, etc., the locations of acupuncture points in various parts of their bodies will be different, so it is indeed difficult to quickly and accurately find the locations of the acupuncture points on the body surface using the existing above-mentioned operation methods. Once the locations of the acupuncture points are not found accurately, the treatment of certain diseases through the acupuncture points of the human body will not have a good effect.

[0005] In this regard, the inventor of this patent combined clinical experience, thought about the problems encountered in clinical work, read a large amount of scientific research materials and literature, and through retrieval and novelty search, gradually conceived and designed this application to solve related technical problems. Summary of the invention

[0006] The present invention aims to solve one of the technical problems in the related art at least to a certain extent. To this end, the purpose of the present invention is to propose an AI-based automatic acupoint identification and positioning system and method.

[0007] To achieve one of the above purposes, the AI-based automatic acupoint identification and positioning system according to an embodiment of the present invention includes:

[0008] AI server; the AI ​​server is used to establish at least one sample model based on a pre-collected three-dimensional image of a human body shape, wherein the three-dimensional image of a human body shape contains acupuncture point markers of various parts of a human body;

[0009] An image acquisition device; the image acquisition device communicates with the AI ​​server via a network connection, and is used to acquire a three-dimensional target image of the patient's body shape when the positions of various acupuncture points on the patient's body need to be located, and upload the three-dimensional target image to the AI ​​server via the network;

[0010] A body shape recognition and comparison module; the body shape recognition and comparison module is connected to the AI ​​server and is used to compare the three-dimensional target image with a plurality of three-dimensional images of human body shapes of at least one sample model in the AI ​​server to determine whether the three-dimensional target image can match any of the three-dimensional images of human body shapes;

[0011] AI acupoint recognition module; the AI ​​acupoint recognition module is connected to the AI ​​server and is used to determine the positions of various acupoints on the patient's body based on the matched three-dimensional image of the human body when the three-dimensional target image is successfully matched with any of the three-dimensional images of the human body.

[0012] In addition, the AI-based automated acupoint identification and positioning system according to the above embodiment of the present invention may also have the following additional technical features:

[0013] According to one embodiment of the present invention, it also includes an acupoint projection module;

[0014] The acupoint projection module is connected to the AI ​​acupoint recognition module, and is used to project the acupoint marking points of various parts of the human body displayed by the matched three-dimensional image to various positions on the patient's skin surface.

[0015] According to one embodiment of the present invention, it also includes a data storage module;

[0016] The data storage module is connected to the AI ​​server and is used to store the collected three-dimensional images of human body shape.

[0017] According to one embodiment of the present invention, it also includes a display module;

[0018] The display module is connected to the image acquisition device and is used to display the three-dimensional target image acquired by the image acquisition device.

[0019] According to one embodiment of the present invention, it also includes a prompt module;

[0020] The prompt module is connected to the image acquisition device. When the three-dimensional target image successfully matches the three-dimensional image of the human body shape in the sample model, the AI ​​server sends a first control signal to the image acquisition device so that the image acquisition device controls the prompt module to issue a prompt.

[0021] According to one embodiment of the present invention, when the three-dimensional target image successfully matches the three-dimensional image of the human body shape in the sample model, the AI ​​server also sends a second control signal to the image acquisition device, so that the image acquisition device marks the acupuncture points of various parts of the patient's body in the three-dimensional target image on the display module and marks the corresponding acupuncture point names next to the acupuncture points of various parts.

[0022] According to one embodiment of the present invention, the image acquisition device comprises:

[0023] A shooting camera, wherein the shooting camera is used to collect the three-dimensional target image;

[0024] A processor, connected to the shooting camera, for processing the three-dimensional target image;

[0025] A communication module, the communication module is connected to the processor and is used to upload the processed three-dimensional target image to the AI ​​server.

[0026] According to an embodiment of the present invention, the AI ​​server is specifically used to establish a sample library storing at least one group of human body three-dimensional images according to the acquired pre-collected at least one group of human body three-dimensional images, and extract multiple human acupoint sample features from the human body three-dimensional images respectively, and all acupoint sample features from the same group of human body three-dimensional images constitute a sample feature set;

[0027] The AI ​​server also uses the features in the same acupoint sample feature set to establish a corresponding sample model, and each acupoint sample feature set corresponds to a sample model.

[0028] According to one embodiment of the present invention, when the three-dimensional target image successfully matches the three-dimensional image of the human body shape in the sample model, the AI ​​server also stores the three-dimensional target image in the sample library as a new three-dimensional image of the human body shape, and establishes a new sample model based on multiple new three-dimensional images of the human body shape.

[0029] To achieve the second objective above, the method for automatic acupoint identification and positioning based on AI according to an embodiment of the present invention comprises the following steps:

[0030] S21. Collection of three-dimensional images of human body shape;

[0031] Obtain a large number of three-dimensional images of human body with marked acupuncture points;

[0032] S22, human body three-dimensional image data annotation;

[0033] Use annotation tools to mark the acupuncture point locations on each 3D image of the human body and create a detailed annotation file, including the image file name, coordinates of each part of the human body and its corresponding acupuncture point name;

[0034] S23, preprocessing of human body three-dimensional image data;

[0035] Standardize the 3D images of human body shape, including resizing and color normalization. Use data augmentation technology to randomly rotate, scale, and flip the 3D images of human body shape to increase data diversity and improve the robustness of the sample model.

[0036] S24, dividing the data set;

[0037] The dataset of human body 3D images is divided into training set, validation set and test set, usually with a ratio of 70% training set, 20% validation set and 10% test set;

[0038] S25, select model;

[0039] Choose a suitable deep learning model based on the complexity of the task, such as convolutional neural network (CNN). Common model architectures include ResNet, VGG, YOLO, and Faster R-CNN.

[0040] S26, build network structure;

[0041] Build or select a suitable CNN model architecture, add appropriate layers according to task requirements, such as convolutional layers, pooling layers, and fully connected layers, and add target detection and key point detection heads to the model to accurately mark human acupuncture points.

[0042] The beneficial effects of the present invention are:

[0043] First, the AI-based automated acupuncture point identification and positioning system provided by the present application, in specific implementation, uses AI artificial intelligence technology to pre-collect multiple three-dimensional images of human body shape to establish at least one sample model, wherein the three-dimensional images of human body shape contain acupuncture point marking points of various parts of the human body, and the collected three-dimensional images of human body shape will include various combinations of people of different heights, weights, and fatness and thinness. When the positions of various acupuncture points on a patient's body need to be located, the image acquisition device can be used to acquire a three-dimensional target image of the patient's body shape, and the three-dimensional target image can be matched with any of the three-dimensional images of human body shape in size ratio. When the match is successful, the matched three-dimensional image of the human body can be used as a reference, and the AI ​​acupoint recognition module can be used to determine the positions of various acupoints on the patient's body, and the acupoint projection module can be used to project the acupoint marking points of various parts of the human body displayed by it to various positions on the patient's skin surface according to the matched three-dimensional image of the human body. In this way, the medical staff can save the need for manual searching and positioning, so that the present application can truly realize the rapid and accurate search and positioning of the positions of the acupoints that need to be located on the patient's body, which is conducive to using acupoint therapy to help patients treat certain diseases quickly and efficiently.

[0044] Secondly, AI technology can be used to quickly and accurately locate acupoints, and when the acupoint marking points on various parts of the human body are projected onto various locations on the patient's skin surface, the acupoint positioning can be visualized and intuitive. Visualized acupoints can assist in learning and acupoint knowledge, which is of great help in expanding the influence of traditional Chinese medicine, and can also more easily help people who need to quickly and accurately locate acupoints.

[0045] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0047] Figure 1 This is the overall block diagram of the AI-based automated acupoint identification and positioning system of the present invention;

[0048] Figure 2 is a flow chart of a method for determining the positions of various acupoints on a patient's body using an AI acupoint recognition module in an embodiment of the present invention;

[0049] Figure 3 It is a flowchart of the steps of the AI-based automated acupoint identification and positioning method of the present invention;

[0050] Figure 4 is a flow chart of the steps of sample model training in an embodiment of the present invention;

[0051] Figure 5 is a flow chart of steps for evaluating and optimizing a sample model in an embodiment of the present invention;

[0052] Figure 6 is a flowchart of the steps of sample model deployment and application in an embodiment of the present invention;

[0053] Figure 7 is a flowchart of the steps of sample AI training in an embodiment of the present invention;

[0054] Reference numerals:

[0055] AI server 10;

[0056] Image acquisition device 20;

[0057] Shooting camera 201;

[0058] Processor 202;

[0059] Communication module 203;

[0060] Body shape recognition and comparison module 30;

[0061] AI acupoint recognition module 40;

[0062] Acupoint projection module 50;

[0063] Data storage module 60;

[0064] Display module 70;

[0065] Prompt module 80;

[0066] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0067] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the drawings of the specification, wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the drawings are exemplary and are intended to be used to explain the present invention, but cannot be understood as limiting the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present invention.

[0068] In the description of the present invention, it is to be understood that the terms “center”, “longitudinal”, “lateral”, “length”, “width”, “thickness”, “up”, “down”, “front”, “back”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inside”, “outside”, “clockwise”, “counterclockwise”, “axial”, “circumferential”, “radial”, etc., indicating orientations or positional relationships are based on the orientations or positional relationships shown in the drawings of the specification, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0069] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0070] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0071] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include that the first and second features are in direct contact, or may include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply indicates that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes that the first feature is directly below and obliquely below the second feature, or simply indicates that the first feature is lower in level than the second feature.

[0072] The following describes in detail the AI-based automatic acupoint recognition and positioning system according to an embodiment of the present invention with reference to the accompanying drawings.

[0073] Reference Figure 1 As shown, the AI-based automatic acupoint recognition and positioning system provided according to an embodiment of the present invention includes an AI server 10, an image acquisition device 20, a body shape recognition and comparison module 30 and an AI acupoint recognition module 40:

[0074] The AI ​​server 10 is used to establish at least one sample model based on a pre-collected three-dimensional image of a human body shape, wherein the three-dimensional image of a human body shape contains acupuncture point markers of various parts of the human body; the image acquisition device 20 communicates with the AI ​​server 10 via a network connection, and is used to collect a three-dimensional target image of the patient's body shape when the positions of various acupuncture points on the patient's body need to be located, and upload the three-dimensional target image to the AI ​​server 10 via the network;

[0075] In addition, the body shape recognition and comparison module 30 is connected to the AI ​​server 10, and is used to compare the three-dimensional target image with multiple human body shape three-dimensional images of at least one sample model in the AI ​​server 10 to determine whether the three-dimensional target image can match any of the human body shape three-dimensional images; the AI ​​acupoint recognition module 40 is connected to the AI ​​server 10, and is used to determine the positions of various acupoints on the patient's body based on the matched three-dimensional human body shape image when the three-dimensional target image successfully matches any of the human body shape three-dimensional images.

[0076] Based on the above, it can be clearly seen that when this application is implemented specifically, it is mainly used as an AI-based automated acupoint identification and positioning system.

[0077] Specifically, when applying the present application, the AI ​​server 10 is used to establish at least one sample model based on a pre-collected three-dimensional image of the human body shape, wherein the three-dimensional image of the human body shape contains acupuncture point marking points of various parts of the human body. When the positions of various acupuncture points on a patient's body need to be located, the image acquisition device 20 can be used to acquire a three-dimensional target image of the patient's body shape, and the three-dimensional target image can be uploaded to the AI ​​server 10 via the network.

[0078] Then, the body shape recognition and comparison module 30 can be used to compare the three-dimensional target image with multiple human body shape three-dimensional images of at least one sample model in the AI ​​server 10 to determine whether the three-dimensional target image can match any of the three-dimensional human body shape images in terms of size ratio. When the three-dimensional target image successfully matches any of the three-dimensional human body shape images in terms of size ratio, the AI ​​acupoint recognition module 40 can be used to determine the positions of various acupoints on the patient's body based on the matched three-dimensional human body shape image.

[0079] Obviously, for the AI-based automated acupoint identification and positioning system provided in the present application, it utilizes AI artificial intelligence technology to pre-collect multiple three-dimensional images of the human body to establish at least one sample model, wherein the three-dimensional images of the human body contain acupoint markers of various parts of the human body, and the collected three-dimensional images of the human body will include various combinations of people of different heights, weights, and fatness or thinness.

[0080] Subsequently, when the positions of various acupuncture points on a patient's body need to be located, the image acquisition device 20 can be used to acquire a three-dimensional target image of the patient's body shape, and the three-dimensional target image can be matched with any three-dimensional image of the human body shape in terms of size ratio. When the match is successful, the matched three-dimensional image of the human body shape can be used as a reference, and the AI ​​acupuncture point recognition module 40 can be used to determine the positions of various acupuncture points on the patient's body. In this way, the positions of acupuncture points in various parts of the body of different patients can be located almost accurately, so as to help patients treat certain diseases through acupuncture point therapy.

[0081] Furthermore, through the above-mentioned optimized design, the overall structure of the present application is highly practical and has good use effect.

[0082] Furthermore, in a specific implementation, according to an embodiment of the present invention, the present application further includes an acupoint projection module 50;

[0083] The acupoint projection module 50 is connected to the AI ​​acupoint recognition module 40, and is used to project the acupoint marking points of various parts of the human body displayed therein to various positions on the patient's skin surface according to the matched three-dimensional image of the human body shape.

[0084] Therefore, when the present application is actually applied, the positions of the patient's body where the acupoints need to be located can be exposed outside the clothes, and the acupoint projection module 50 can be used to project the acupoint marking points of various parts of the human body displayed to various positions on the patient's skin surface according to the matched three-dimensional image of the human body shape. In this way, the medical staff can save the need for manual searching and positioning, so that the present application can truly realize the rapid and accurate search and positioning of the positions of the patient's body where the acupoints need to be located, which is more conducive to using acupoint therapy to help patients treat certain diseases quickly and efficiently.

[0085] Furthermore, through the above-mentioned optimization design, the overall practicality of the present application is stronger, and the use effect can be effectively improved. The acupoints can be quickly and accurately located through AI technology, and when the acupoint marking points in various parts of the human body are projected to various positions on the patient's skin surface, the acupoint positioning can be visualized and intuitive. The visualized acupoints can assist in learning and knowledge of acupoints, which is of great help to expand the influence of traditional Chinese medicine, and can also more easily help people who need to quickly and accurately locate acupoints.

[0086] Furthermore, in a specific implementation, according to an embodiment of the present invention, the present application further includes a data storage module 60;

[0087] The data storage module 60 is connected to the AI ​​server 10 to store the collected three-dimensional image of the human body.

[0088] In addition, in the technical solution, according to an embodiment of the present invention, the present application further includes a display module 70;

[0089] The display module 70 is connected to the image acquisition device 20 to display the three-dimensional target image acquired by the image acquisition device 20 so as to enable visual and intuitive viewing.

[0090] Furthermore, in a specific implementation, according to an embodiment of the present invention, the present application further includes a prompt module 80;

[0091] Among them, the prompt module 80 is connected to the image acquisition device 20. When the three-dimensional target image successfully matches the three-dimensional image of the human body in the sample model in terms of size ratio, the AI ​​server 10 sends a first control signal to the image acquisition device 20, so that the image acquisition device 20 controls the prompt module 80 to issue a prompt.

[0092] In this way, it is convenient to remind medical staff that the three-dimensional target image and the three-dimensional image of the human body in the sample model are successfully matched in terms of size ratio.

[0093] Preferably, in the present technical solution, the prompt module 80 is set as a voice prompt module.

[0094] Furthermore, in a specific implementation, according to an embodiment of the present invention, when the three-dimensional target image successfully matches the three-dimensional image of the human body in the sample model in terms of size ratio, the AI ​​server 10 also sends a second control signal to the image acquisition device 20, so that the image acquisition device 20 marks the acupuncture points of various parts of the patient's body in the three-dimensional target image on the display module 70 and marks the corresponding acupuncture point names next to the acupuncture points of various parts.

[0095] Therefore, when the three-dimensional target image successfully matches the three-dimensional image of the human body in the sample model in terms of size ratio, the image acquisition device 20 will mark the acupuncture points of various parts of the patient's body in the three-dimensional target image on the display module 70 and mark the corresponding acupuncture point names next to the acupuncture points of various parts, so that medical staff can clearly know where the acupuncture points of a certain part of the patient's body are and what the corresponding acupuncture point names are, making it clear at a glance.

[0096] And preferably, in the technical solution, according to an embodiment of the present invention, the image acquisition device 20 includes a shooting camera 201, a processor 202 and a communication module 203;

[0097] Among them, the shooting camera 201 is used to collect the three-dimensional target image, and it is preferably an industrial camera, for example, a Hikvision camera; the processor 202 is connected to the shooting camera 201, and is used to process the three-dimensional target image, such as compressing the shooting data; the communication module 203 is connected to the processor 202, and is used to upload the processed three-dimensional target image to the AI ​​server 10, so that the AI ​​server 10 can decompress the received shooting data, deframe the shot human body shape data, and discard or delete invalid frames.

[0098] Again, in a specific implementation, according to an embodiment of the present invention, the AI ​​server 10 is specifically used to establish a sample library storing at least one group of human body three-dimensional images according to the acquired pre-collected at least one group of human body three-dimensional images, and extract multiple human acupoint sample features from the human body three-dimensional images respectively, and all acupoint sample features from the same group of human body three-dimensional images constitute a sample feature set;

[0099] The AI ​​server 10 also uses the features in the same acupoint sample feature set to establish a corresponding sample model, and each acupoint sample feature set corresponds to a sample model.

[0100] In this way, the number of samples collected in the sample library using the AI-based automated acupoint identification and positioning system provided by the present application will be very diverse, and will be divided into multiple groups based on the arrangement and combination of characteristics such as the patient's height, weight, and fatness. All acupoint sample features from the same group of human body three-dimensional images constitute a sample feature set. Based on this, the AI ​​server 10 described in the present application also uses the features in the same acupoint sample feature set to establish a corresponding sample model, and each acupoint sample feature set corresponds to a sample model.

[0101] In this way, even if the positions of the acupuncture points on a patient's body need to be located later, the image acquisition device 20 can be used to acquire a three-dimensional target image of the patient's body shape, and the three-dimensional target image can be uploaded to the AI ​​server 10 through the network. Then, the body shape recognition and comparison module 30 can be used to compare the three-dimensional target image with multiple three-dimensional images of human body shapes of at least one sample model in the AI ​​server 10 to determine whether the three-dimensional target image can match any of the three-dimensional images of human body shapes in terms of size ratio. When the three-dimensional target image successfully matches any of the three-dimensional images of human body shapes in terms of size ratio, the AI ​​acupuncture point recognition module 40 can be used to determine the positions of the acupuncture points on the patient's body based on the matched three-dimensional image of human body shape.

[0102] In addition, according to one embodiment of the present invention, when the three-dimensional target image successfully matches the three-dimensional image of the human body shape in the sample model in terms of size ratio, the AI ​​server 10 also stores the three-dimensional target image in the sample library as a new three-dimensional image of the human body shape, and establishes a new sample model based on multiple new three-dimensional images of the human body shape.

[0103] In this way, it is clear that the AI ​​server 10 described in the present application can perform deep learning of samples, continuously improve the sample model and continuously optimize the sample model, so that in the process of use, over time, the accuracy of recognition can be continuously improved. As the number of sample models increases, the accuracy of recognition through comparison will also be higher.

[0104] In addition, it should be noted here that for the established sample model, sample training and deep learning can be performed by using neural networks and decision trees. Deep learning can learn more complex features through multi-layer neural networks to improve the accuracy of image recognition. At the same time, deep learning can also automatically extract and select acupuncture point sample features, avoiding the tedious process of manual feature design, making the model performance better.

[0105] It should be added that, in a specific implementation, according to the AI-based automatic acupoint identification and positioning system provided by an embodiment of the present invention, the communication module 203 is preferably a WiFi module, a Bluetooth module, a 4G module or a 5G module.

[0106] Furthermore, it is necessary to add that, referring to Figure 2 As shown, in the present application, the method for determining the positions of the acupuncture points on the patient's body using the AI ​​acupuncture point recognition module 40 is as follows:

[0107] S11, preprocessing the acquired three-dimensional target image, including denoising, enhancement, and normalization algorithm, so as to make the acquired three-dimensional target image clearer and more regular;

[0108] S12, extracting features of human acupuncture point samples from the preprocessed three-dimensional target image, where these features mainly include position, color, texture, shape, etc. There are many methods for extracting features, such as SIFT, HOG, SURF, etc., which are all existing conventional technologies and will not be described in detail here;

[0109] S13, displaying the extracted human acupuncture point sample features at corresponding positions on the preprocessed three-dimensional target image.

[0110] It is further necessary to add that in this application, reference is made to Figure 3 As shown, the AI ​​server 10 is used to establish at least one sample model according to a pre-collected human body sample picture, wherein the human body sample picture contains acupuncture point marking points of various parts of the human body, and the positioning of the acupuncture point marking points of various parts of the human body includes the following steps:

[0111] S21. Collection of three-dimensional images of human body shape;

[0112] Obtain a large number of human body 3D images with labeled acupuncture points. These 3D images should contain acupuncture point markers of different parts of the human body and at different locations. Ensure that the data set contains a variety of human body postures and angles, and cover all possible situations as much as possible.

[0113] S22, human body three-dimensional image data annotation;

[0114] Use annotation tools (such as Labelbox, VGG Image Annotator, etc.) to mark the acupoint locations on each human body 3D image and create detailed annotation files, which are usually stored in JSON or XML format and include the image file name, coordinates of each part of the human body and its corresponding acupoint name.

[0115] S23, preprocessing of human body three-dimensional image data;

[0116] The three-dimensional images of human body shape are standardized, such as size adjustment and color normalization, and data augmentation techniques (Data Augmentation) are used, such as random rotation, scaling and flipping, to increase data diversity and improve the robustness of the sample model.

[0117] S24, dividing the data set;

[0118] The dataset of human body 3D images is divided into a training set, a validation set and a test set, usually with a ratio of 70% training set, 20% validation set and 10% test set.

[0119] S25, select model;

[0120] Select a suitable deep learning model according to the complexity of the task, such as convolutional neural network (CNN). Common model architectures include ResNet, VGG, YOLO, FasterR-CNN, etc.

[0121] S26, build network structure;

[0122] Build or select a suitable CNN model architecture, add appropriate layers according to task requirements, such as convolutional layers, pooling layers, fully connected layers, etc., and add target detection and key point detection heads to the model to accurately mark human acupuncture points.

[0123] In addition, for the training of sample models, refer to Figure 4 As shown, it mainly includes the following steps:

[0124] S31. Define loss function and optimizer;

[0125] Choose an appropriate loss function (such as cross entropy loss, mean squared error) and optimizer (such as Adam, SGD).

[0126] S32, training sample model;

[0127] Use the training set to train the sample model and adjust hyperparameters such as learning rate, batch size, etc. Monitor the performance on the validation set during training to prevent overfitting.

[0128] Also, for sample model evaluation and tuning, refer to Figure 5As shown, it mainly includes the following steps:

[0129] S41, sample model evaluation;

[0130] Use the test set to evaluate the performance of the sample model and check the effect of the sample model through indicators such as accuracy, recall, and F1 score.

[0131] S42, sample model optimization;

[0132] Iteratively adjust the sample model architecture and hyperparameters based on the evaluation results. This may require more data augmentation, additional training data, or trying different network architectures.

[0133] For sample model deployment and application, refer to Figure 6 As shown, it mainly includes the following steps:

[0134] S51, sample model deployment;

[0135] Deploy the trained sample model to the AI ​​server 10 or edge device. Use the sample model to perform actual acupoint recognition tasks to ensure that the system can work properly in real scenarios.

[0136] S52. Continuous monitoring and updating: monitor the performance of sample models in actual use, collect new data and feedback, and continuously update and train sample models to improve recognition accuracy.

[0137] Furthermore, for sample AI training, refer to Figure 7 As shown, it mainly includes the following steps:

[0138] S61, Convolutional Neural Network (CNN);

[0139] CNN extracts the sample features of acupoints in the three-dimensional image of the human body through the convolution layer. The pooling layer reduces the parameters and calculation amount and enhances the generalization ability of the sample model. The convolution kernel learns the sample features of local acupoints in the three-dimensional image of the human body and extracts higher-level sample features of acupoints through layer-by-layer progressive learning.

[0140] S62, back propagation and gradient descent;

[0141] The gradient of the loss function relative to each parameter is calculated through the back-propagation algorithm, and the gradient descent method is used to optimize the sample model parameters and gradually reduce the loss function value.

[0142] S63, data augmentation and regularization;

[0143] Data augmentation techniques increase data diversity and reduce overfitting, and regularization methods such as Dropout and BatchNormalization help sample models generalize better.

[0144] Therefore, by combining the above steps and methods, it is possible to effectively train an AI-based automated acupoint recognition and positioning system that can recognize human acupoints.

[0145] Other embodiments and the like are not described here by way of example.

[0146] In summary, the AI-based automated acupuncture point recognition and positioning system provided by the present application, in its specific implementation, uses AI artificial intelligence technology to pre-collect multiple three-dimensional images of human body shape to establish at least one sample model, wherein the three-dimensional images of human body shape contain acupuncture point marking points of various parts of the human body, and the collected three-dimensional images of human body shape will include various combinations of people of different heights, weights, and fatness and thinness. When the positions of various acupuncture points on a patient's body need to be located, the image acquisition device 20 can be used to acquire a three-dimensional target image of the patient's body shape, and the three-dimensional target image can be matched with any of the three-dimensional images of the human body shape in size ratio. When the match is successful, the matched three-dimensional image of the human body can be used as a reference, and the AI ​​acupoint recognition module 40 can be used to determine the positions of the various acupoints on the patient's body, and the acupoint projection module 50 can be used to project the acupoint marking points of various parts of the human body displayed by it to various positions on the patient's skin surface according to the matched three-dimensional image of the human body. In this way, the medical staff can save the need for manual searching and positioning, so that the present application can truly realize the rapid and accurate search and positioning of the positions of the acupoints that need to be located on the patient's body, which is conducive to using acupoint therapy to help patients treat certain diseases quickly and efficiently.

[0147] Furthermore, through the above-mentioned optimization design, the overall structure of the present application is highly practical and has excellent use effect. The acupoints can be quickly and accurately located through AI technology, and when the acupoint marking points in various parts of the human body are projected to various positions on the patient's skin surface, the acupoint positioning can be visualized and intuitive. The visualized acupoints can assist in learning and knowledge of acupoints, which is of great help to expand the influence of traditional Chinese medicine, and can also more easily help people who need to quickly and accurately locate acupoints.

[0148] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0149] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. An AI-based automated acupoint identification and positioning system, characterized in that: include: AI server; The AI ​​server is used to establish at least one sample model based on a pre-collected three-dimensional image of a human body shape, wherein the three-dimensional image of a human body shape contains acupuncture point markers of various parts of a human body; Image acquisition device; the image acquisition device communicates with the AI ​​server through a network connection, and is used to acquire a three-dimensional target image of the patient's body shape when the positions of various acupuncture points on the patient's body need to be located, and upload the three-dimensional target image to the AI ​​server through the network; A body shape recognition and comparison module; the body shape recognition and comparison module is connected to the AI ​​server and is used to compare the three-dimensional target image with multiple three-dimensional images of human body shapes of at least one sample model in the AI ​​server to determine whether the three-dimensional target image can match the size ratio of any three-dimensional image of human body shape; AI acupoint recognition module; the AI ​​acupoint recognition module is connected to the AI ​​server and is used to determine the acupoint locations of various parts of the patient's body based on the matched three-dimensional image of the human body when the size ratio of the three-dimensional target image is successfully matched with any three-dimensional image of the human body.

2. The AI-based automated acupoint identification and positioning system according to claim 1, characterized in that: Also included is an acupoint projection module; The acupoint projection module is connected to the AI ​​acupoint recognition module, and is used to project the acupoint marking points of various parts of the human body displayed to various positions on the patient's skin surface according to the matched three-dimensional image of the human body shape.

3. The AI-based automated acupoint identification and positioning system according to claim 1, characterized in that: Also included is a data storage module; The data storage module is connected to the AI ​​server to store the collected three-dimensional images of the human body.

4. The AI-based automated acupoint identification and positioning system according to claim 1, characterized in that: Also includes a display module; The display module is connected to the image acquisition device and is used to display the three-dimensional target image acquired by the image acquisition device.

5. The AI-based automated acupoint identification and positioning system according to claim 4, characterized in that: Also includes a prompt module; The prompt module is connected to the image acquisition device. When the size ratio of the three-dimensional target image and the three-dimensional image of the human body in the sample model matches successfully, the AI ​​server sends a first control signal to the image acquisition device so that the image acquisition device controls the prompt module to issue a prompt.

6. The AI-based automated acupoint identification and positioning system according to claim 5, characterized in that: When the size ratio of the three-dimensional target image successfully matches the three-dimensional image of the human body in the sample model, the AI ​​server also sends a second control signal to the image acquisition device, so that the image acquisition device calibrates the acupuncture points of various parts of the patient's body in the three-dimensional target image on the display module and calibrates the corresponding acupuncture point names next to the acupuncture points in various parts.

7. The AI-based automated acupoint identification and positioning system according to claim 1, characterized in that: Image acquisition equipment includes: A shooting camera, which is used to collect a three-dimensional target image; A processor, the processor is connected to the shooting camera and is used to process the three-dimensional target image; The communication module is connected to the processor and is used to upload the processed three-dimensional target image to the AI ​​server.

8. The AI-based automated acupoint identification and positioning system according to claim 1, characterized in that: The AI ​​server is specifically used to establish a sample library storing at least one set of three-dimensional human body images according to the acquired three-dimensional human body images, and extract a plurality of human acupuncture point sample features from the three-dimensional human body images respectively, and all acupuncture point sample features from the same set of three-dimensional human body images constitute a sample feature set; The AI ​​server also uses the features in the same acupoint sample feature set to establish a corresponding sample model, and each acupoint sample feature set corresponds to a sample model.

9. The AI-based automated acupoint identification and positioning system according to claim 8, characterized in that: When the size ratio of the three-dimensional target image successfully matches the three-dimensional image of the human body shape in the sample model, the AI ​​server also stores the three-dimensional target image in the sample library as a new three-dimensional image of the human body shape, and establishes a new sample model based on multiple new three-dimensional images of the human body shape.

10. A method for automatic acupoint identification and positioning based on AI, comprising the following steps: S21. Collection of three-dimensional images of human body shape; Obtain a large number of three-dimensional images of human body with marked acupuncture points; S22, human body three-dimensional image data annotation; Use annotation tools to mark the acupuncture point locations on each 3D image of the human body and create a detailed annotation file, including the image file name, coordinates of each part of the human body and its corresponding acupuncture point name; S23, preprocessing of human body three-dimensional image data; Standardize the 3D images of human body shape, including resizing and color normalization. Use data augmentation technology to randomly rotate, scale, and flip the 3D images of human body shape to increase data diversity and improve the robustness of the sample model. S24, dividing the data set; The dataset of human body 3D images is divided into training set, validation set and test set, usually with a ratio of 70% training set, 20% validation set and 10% test set; S25, select model; Choose a suitable deep learning model based on the complexity of the task, such as convolutional neural network (CNN). Common model architectures include ResNet, VGG, YOLO, and FasterR-CNN. S26, build network structure; Build or select a suitable CNN model architecture, add appropriate layers according to task requirements, such as convolutional layers, pooling layers, and fully connected layers, and add target detection and key point detection heads to the model to accurately mark human acupuncture points.

Images