Intelligent acupoint positioning method and system for Chinese and Western medicine massage equipment

Abstract

The present invention relates to the field of multi-image analysis technology, and in particular to an intelligent acupoint positioning method and system for Chinese and Western medicine massage equipment. The method obtains an image pyramid of a user's human body depth image and the reference skeleton points and analysis feature points of each image in the pyramid, obtains the acupoint skeleton direction consistency value of each analysis point according to the position distribution of each analysis point compared with the remaining analysis points, obtains the acupoint spatial position value of the reference skeleton point according to the number and position distribution of the analysis feature points within the preset neighborhood range of the reference skeleton point, obtains the acupoint registration domain in each image in the image pyramid in combination with the acupoint skeleton direction consistency value of the analysis point and the acupoint spatial position value of the reference skeleton point corresponding to the analysis point, and locates the user's acupoints based on the overlapping area of ​​the acupoint registration domains in all images in the image pyramid. The present invention effectively improves the accuracy of locating the user's acupoints.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-image analysis, and in particular to an intelligent acupoint positioning method and system for traditional Chinese and western medicine massage equipment. Background Art

[0002] Chinese and Western medicine massage equipment achieves the effects of health preservation, meridian dredging, disease prevention and treatment by massaging and stimulating the acupuncture points of the human body. If the acupuncture points are not accurately located, it is easy to make the stimulation site inaccurate, and the expected treatment effect cannot be achieved, which may lead to the condition not being relieved or even adverse reactions. Therefore, acupuncture point positioning is the key technology of Chinese and Western medicine massage equipment.

[0003] Existing methods obtain human body image information through depth cameras, perform key point detection on human body images, and match the detection results with standard acupoint templates to locate human acupoints. However, individual differences in body shape, muscle distribution, fat thickness, etc. will significantly affect the actual position of acupoints, resulting in low accuracy in acupoint positioning. Summary of the invention

[0004] In order to solve the technical problem that individual differences affect the actual position of acupoints, resulting in low accuracy of acupoint positioning, the purpose of the present invention is to provide an acupoint intelligent positioning method and system for Chinese and Western medicine massage equipment. The technical solutions adopted are as follows:

[0005] In a first aspect, an embodiment of the present invention provides an intelligent acupoint positioning method for a Chinese and Western medicine massage device, the method comprising:

[0006] Acquire a human body depth image of a user using a Chinese and Western medicine massage device;

[0007] Obtain an image pyramid of a human body depth image, as well as reference skeleton points and analysis feature points of each image in the image pyramid; record the reference skeleton points and the analysis feature points as analysis points, and obtain the acupoint skeleton direction consistency value of each analysis point according to the position distribution of each analysis point in each image in the image pyramid compared with the remaining analysis points;

[0008] Analyze the number and position distribution of feature points within a preset neighborhood of each reference bone point to obtain the acupoint spatial position value of each reference bone point in each image in the image pyramid;

[0009] According to the acupoint skeleton direction consistency value of each analysis point and the acupoint spatial position value of the reference skeleton point corresponding to the preset neighborhood range of each analysis point, the acupoint registration domain in each image in the image pyramid is obtained; based on the overlapping area of ​​the acupoint registration domains of all images in the image pyramid, the user's acupoints are located.

[0010] Furthermore, the method for obtaining the reference skeleton points and analysis feature points of each image in the image pyramid includes:

[0011] Selecting a reference structure image from an image pyramid; matching the reference structure image with a standard human body model to obtain key bone points in the reference structure image;

[0012] Select one image from the remaining images in the image pyramid except the reference structure image as the target image, and record the pixel point at the center position of the matching pixel point corresponding to each key skeleton point of the reference structure image in the target image as the reference skeleton point of the target image;

[0013] Feature extraction is performed on each image in the image pyramid to obtain feature points in each image; the pixel points at the center position of the matching pixel points corresponding to the feature points in the target image and the feature points in other images in the image pyramid except the reference structure image and the target image in the target image are recorded as the analysis feature points of the target image.

[0014] Furthermore, the step of obtaining the acupoint-bone direction consistency value of each analysis point includes:

[0015] For each image in the image pyramid, an analysis point is randomly selected from the image and recorded as the target point. The line segments obtained by connecting the target point with the remaining analysis points in the image except the target point are recorded as the analysis line segments of the target point. The angle between the analysis line segment and the preset direction is used as the distribution angle of each analysis line segment of the target point.

[0016] The cumulative sum of the absolute values ​​of the differences in the distribution angles of the pairwise analysis line segments of the target point is negatively correlated and mapped to obtain the acupoint bone direction consistency value of the target point.

[0017] Furthermore, the step of obtaining the acupuncture point spatial position value of each reference skeleton point in each image in the image pyramid includes:

[0018] For each image in the image pyramid, the analysis feature points within the preset neighborhood range of each reference skeleton point of the image are recorded as the neighborhood points of each reference skeleton point;

[0019] For each reference bone point in the image, a neighborhood point of the reference bone point is randomly selected and recorded as a reference point, a neighborhood point closest to the reference point is selected from the neighborhood points of the reference bone point and recorded as a judgment point of the reference point, an angle between a line segment connecting the reference point and its judgment point and a preset direction is recorded as a judgment angle of the reference point, and the judgment angle of the judgment point is obtained;

[0020] The absolute value of the difference between the judgment angle of the reference point and its judgment point is recorded as the direction consistency value of the reference point; the cumulative sum of the direction consistency values ​​of all neighboring points of each reference skeleton point is used as the neighborhood direction value of the corresponding reference skeleton point;

[0021] According to the total number of neighborhood points of each reference skeleton point in each image in the image pyramid and the neighborhood direction value, the acupuncture point spatial position value of the corresponding reference skeleton point is obtained; the total number of neighborhood points of each reference skeleton point is positively correlated with the acupuncture point spatial position value, and the neighborhood direction value is negatively correlated with the acupuncture point spatial position value.

[0022] Furthermore, the step of obtaining the acupoint registration domain in each image in the image pyramid includes:

[0023] For each analysis point of each image in the image pyramid, the product of the acupoint-bone direction consistency value of the analysis point and the acupoint spatial position value of the reference bone point corresponding to the preset neighborhood range of the analysis point is normalized to obtain the acupoint-bone association value of the analysis point;

[0024] Record the analysis points whose acupoint-bone correlation values ​​are greater than the preset correlation threshold as standard points; weight the preset length using the acupoint-bone correlation value of each standard point to obtain the adjusted length of each standard point;

[0025] The area formed by taking each standard point as the center and the adjustment length as the radius is used as the acupoint registration domain of each image in the image pyramid.

[0026] Furthermore, locating the acupuncture points of the human body based on the overlapping areas of the acupuncture point registration domains of all images in the image pyramid includes:

[0027] The largest image in the image pyramid is the human body depth image;

[0028] Selecting an acupoint registration domain from the human body depth image as the target domain, recording the acupoint registration domain where the pixel points in the target domain correspond to the matching pixel points in each of the target images as the analysis registration domain of the target domain, and recording the target image where the analysis registration domain is located as the analysis image of the target domain;

[0029] If the total number of analysis images of the target domain is greater than a preset number threshold, then for the analysis registration domain of each analysis image of the target domain, a curve fitting is performed on the pixel point at the center position of the matching pixel point corresponding to each edge pixel point on the edge of the analysis registration domain in the human body depth image, and the closed area formed by the fitting curve is recorded as the alignment area of ​​the analysis registration domain in the human body depth image;

[0030] The part that exists in common between the target domain and the aligned regions of all its analyzed registration domains is regarded as the user's acupoint region.

[0031] Furthermore, the reference structure image is the image with the smallest size in the image pyramid.

[0032] Furthermore, the downsampling ratio in the image pyramid construction process is an odd number.

[0033] Furthermore, the method for obtaining feature points in each image is a scale-invariant feature transformation algorithm.

[0034] In a second aspect, another embodiment of the present invention provides an intelligent acupoint positioning system for Chinese and Western medicine massage equipment, the system comprising:

[0035] A data acquisition module for acquiring a human body depth image of a user using the Chinese and Western medicine massage device;

[0036] A direction analysis module is used to obtain an image pyramid of a human body depth image, as well as reference skeleton points and analysis feature points of each image in the image pyramid; the reference skeleton points and the analysis feature points are recorded as analysis points, and according to the position distribution of each analysis point of each image in the image pyramid compared with the remaining analysis points, the acupoint skeleton direction consistency value of each analysis point is obtained;

[0037] A spatial analysis module, used to analyze the number and position distribution of feature points within a preset neighborhood of each reference bone point, and obtain the acupoint spatial position value of each reference bone point in each image in the image pyramid;

[0038] The acupoint positioning module is used to obtain the acupoint registration domain in each image in the image pyramid based on the acupoint skeleton direction consistency value of each analysis point and the acupoint spatial position value of the reference skeleton point corresponding to the preset neighborhood range of each analysis point; and locate the acupoints of the human body based on the overlapping area of ​​the acupoint registration domains of all images in the image pyramid.

[0039] The present invention has the following beneficial effects:

[0040] In an embodiment of the present invention, in order to reduce the influence of individual differences on acupoint positioning, an image pyramid of a human body depth image is obtained, and acupoint detection is performed on multi-size images; the distribution of acupoints around different key bones of the human body is analyzed, and two distribution characteristics of acupoints and bones are summarized, specifically: the acupoints are arranged along the bone distribution direction, and the acupoints near the key bones are distributed in a certain direction of the bones; this scheme refers to the first feature to obtain the acupoint-bone direction consistency value, and while referring to the second feature, considers the number of feature points analyzed within the preset neighborhood range of the reference bone point, presents the importance of the reference bone point in the detailed structure analysis of the human body, and improves the accuracy of the acupoint spatial position value in reflecting the distribution of acupoints near the reference bone point; the acupoint bone direction consistency value and the acupoint spatial position value obtained by combining the above two features are determined to determine the acupoint registration domain in each image in the image pyramid, and the user's acupoints are located based on the overlapping area of ​​the acupoint registration domains of all images in the image pyramid. Higher-resolution image layers can provide finer local features, while lower-resolution image layers can capture a wider range of global information, improving the reliability of acupoint area determination; confirming acupoints through overlapping areas in multi-scale images effectively avoids misjudgment or omissions that may occur at a single scale, and effectively reduces the impact of individual differences on acupoint positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are 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 these drawings without paying creative work.

[0042] Figure 1 A flowchart of a method for intelligently locating acupoints for Chinese and Western medicine massage equipment provided by one embodiment of the present invention;

[0043] Figure 2 A system structure diagram of an intelligent acupoint positioning system for Chinese and Western medicine massage equipment provided by one embodiment of the present invention;

[0044] Figure 3 A schematic diagram of a computer device for an intelligent acupoint positioning device for Chinese and Western medicine massage equipment provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the intelligent acupoint positioning method and system for Chinese and Western medicine massage equipment proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0046] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0047] The specific scheme of the intelligent acupoint positioning method and system for Chinese and Western medicine massage equipment provided by the present invention is described in detail below with reference to the accompanying drawings.

[0048] Embodiment 1:

[0049] The present invention proposes an intelligent acupoint positioning method for Chinese and Western medicine massage equipment, please refer to Figure 1 , which shows a flowchart of the steps of a method for intelligent acupoint positioning for Chinese and Western medicine massage equipment provided by an embodiment of the present invention, the method comprising:

[0050] Step S1: Acquire a human body depth image of a user using the Chinese and Western medicine massage equipment.

[0051] Specifically, the patient lies naturally on the Chinese and Western medicine massage equipment without obvious movement features that block key parts, and the depth camera installed on the equipment is used to collect patient information to obtain an original human body depth image. In order to remove noise in the image caused by equipment accuracy or external environment, the original human body depth image is denoised to obtain a human body depth image.

[0052] It should be noted that the depth resolution of the depth camera used in this embodiment is or , the depth sensing range is 0.2 meters to 3 meters; Gaussian filtering is used for denoising. The specific method will not be introduced here and is a technical means well known to those skilled in the art.

[0053] Step S2: Obtain an image pyramid of a human body depth image, as well as reference bone points and analysis feature points of each image in the image pyramid; record the reference bone points and analysis feature points as analysis points, and obtain the acupoint bone direction consistency value of each analysis point according to the position distribution of each analysis point of each image in the image pyramid compared with the remaining analysis points.

[0054] Due to individual differences, such as inconsistent body shape, muscle distribution, and fat thickness, the key points determined by the feature point detection algorithm cannot be effectively detected due to occlusion by fat and muscle, which in turn causes errors in the results of acupoint matching between the key points and the standard human body model. In order to reduce the impact of individual differences on acupoint positioning, this solution obtains an image pyramid of human depth images, analyzes each image in the image pyramid in combination with the distribution characteristics of bones and acupoints, and improves the accuracy of user acupoint positioning.

[0055] It should be noted that, in the embodiment of the present invention, the Gaussian image pyramid algorithm is selected to downsample the human body depth image, and the downsampling ratio in the image pyramid construction process is an odd number; the downsampling ratio set in the downsampling process is 3, and the number of downsampling is 20, which can be set by the implementer according to the specific situation. Among them, the Gaussian image pyramid algorithm is a well-known technology for those skilled in the art and will not be described in detail here.

[0056] Considering that the downsampling process will blur the details in the image, the largest image in the image pyramid has the least blur, and the smallest image has the least impact on the overall structure of the image after downsampling. Therefore, the smallest image in the image pyramid of the human depth image is used as the reference structure image. The human skeleton structure in the reference structure image is the most complete, and only the external visible contours of the user, such as the user's limbs, are retained.

[0057] The reference structure image is matched with the standard human body model to obtain key bone points in the reference structure image, which represent the key bone parts of the user. In this embodiment, the key bone points in the reference structure image are obtained by first using a feature extraction algorithm such as the OpenPose (Realtime Multi-Person 2D PoseEstimation using Part Affinity Fields) algorithm, the Deep High-Resolution Representation Learning for Visual Recognition, HRNet) and other algorithms are used to identify the key points in the reference structure image and find the positions of the human joints in the three-dimensional standard human body model; secondly, the reference structure image is converted into a three-dimensional voxel model to extract the three-dimensional features of the joints; then, a standardized three-dimensional human body model containing important parts (such as shoulders, elbows and knees, etc.) is constructed, usually with joints as key points, and the positional relationship of each joint is defined; then, the ICP algorithm is used to match the joint points extracted from the reference structure image with the joint points in the standard human body model; the least squares method is used to fine-tune the position of the key points to reduce the error between the key points in the reference structure image and the standard human body model; the constraint condition is: use the geometric relationship between the joints to ensure that the matching process meets the biomechanical constraints; finally, mark the corresponding key points in the reference structure image and use these key points as the key bone points in the reference structure image.

[0058] In other embodiments, the reference structure image may also be input into a pre-trained neural network to output a reference structure image with key bone points annotated thereon.

[0059] Select any one image from the remaining images in the image pyramid except the reference structure image and record it as the target image. The pixel point at the center position of the matching pixel point corresponding to each key skeleton point of the reference structure image in the target image is recorded as the reference skeleton point of the target image. Perform feature extraction on each image in the image pyramid to obtain the feature points in each image. The feature points in the target image and the pixel points at the center position of the matching pixel points corresponding to the feature points in other images in the image pyramid except the reference structure image and the target image are recorded as the analysis feature points of the target image.

[0060] In order to ensure the accuracy of key human skeletons and other parts in each image in the image pyramid, the key skeleton points of the reference structure image are mapped to each image in the image pyramid to obtain the reference skeleton points of each image; the reference skeleton points represent important skeleton parts such as shoulders, elbows, knees and spine of the human body. The scale-invariant feature transformation algorithm is used to obtain the feature points in each image in the image pyramid. The feature points represent the attached points of the skeleton or small bones. The feature points provide spatial position reference assistance for acupoint detection. Among them, the scale-invariant feature transformation algorithm is a well-known technology for those skilled in the art and will not be repeated here.

[0061] It should be noted that, since the downsampling ratio in this scheme is 3, for any two adjacent images in the image pyramid, the larger image The pixel area will be downsampled into a single pixel in the smaller image. Therefore, each pixel in the smaller image corresponds to multiple matching pixels in the larger image and these matching pixels have a center position; each pixel in the larger image corresponds to only one matching pixel in the smaller image, and the center position of the matching pixel is the matching pixel itself. The content that all pixels in each layer of the image pyramid can find corresponding matching pixels in other images is a well-known technology and will not be described in detail here.

[0062] It should be noted that the number of reference skeleton points in each image in the image pyramid is equal to the number of key skeleton points in the reference structure image. The reference structure image does not participate in the subsequent analysis and is only used to determine the reference skeleton points in the remaining images in the image pyramid except the reference structure image.

[0063] Acupoints around the shoulders, such as Jianjing and Jianliao, are usually located near the shoulder joints; acupoints around the elbow joints, such as Quchi and Tianjing, are usually located on the surface or near the elbow bones; acupoints around the knees, such as Xiyan and Dubi, are usually located on both sides or below the knee joints, and these acupoints are distributed depending on the knee bones; acupoints on both sides of the back spine, such as Dazhui and Mingmen, are usually distributed along the spine bones, and these acupoints are directly related to the convex parts of the spine. Therefore, the distribution characteristics of acupoints near key bones are: the acupoints are located on both sides of the key bones, below them, or arranged along the distribution direction of the bones.

[0064] Referring to the distribution of acupuncture points on the spine: the acupuncture points on both sides of the spine, such as Dazhui acupoint and Mingmen acupoint, are usually distributed along the direction of the spine, and it is concluded that the acupuncture points are arranged in a fixed direction or along the direction of several analysis points; if the overall distribution of the analysis points is more linear or streamlined, the more likely it is that the acupuncture points are distributed in the main direction of the analysis point position distribution. Therefore, according to the spatial distribution of each analysis point compared with the other analysis points, the acupuncture point bone direction consistency value of the analysis point is obtained.

[0065] Preferably, in some possible implementation modes of the embodiments of the present invention, the method for obtaining the acupoint skeleton direction consistency value includes: for each image in the image pyramid, randomly selecting an analysis point from the image as the target point, and connecting the target point with the remaining analysis points in the image except the target point to obtain line segments as analysis line segments of the target point; taking the angle between the analysis line segment and the preset direction as the distribution angle of each analysis line segment of the target point; and performing negative correlation mapping on the cumulative sum of the absolute values ​​of the difference in the distribution angles of the pairwise analysis line segments of the target point to obtain the acupoint skeleton direction consistency value of the target point.

[0066] It should be noted that, in the embodiment of the present invention, the preset direction is the horizontal direction, and the value range of the distribution angle is , implementers can set it according to specific circumstances.

[0067] In a specific implementation of the embodiment of the present invention, for each image in the image pyramid, an analysis point is selected from the image and recorded as a target point. The acupoint-bone direction consistency value of the target point is expressed by the formula:

[0068]

[0069] Where, F is the consistent value of the acupoint bone direction of the target point; A is the total number of analysis segments of the target point, which is equivalent to the difference between the total number of analysis points in the image where the target point is located and the constant 1; is the analysis angle of the ath analysis line segment of the target point; is the analysis angle of the bth analysis line segment of the target point; is the absolute value function; exp is the exponential function with a natural constant as the base.

[0070] It should be noted that if the sum of the absolute values ​​of the differences in the distribution angles of the pairwise analysis segments of the target point is smaller, the analysis angles of all the analysis segments of the target point are closer, which means that the overall distribution of the target point and other analysis points is closer to a straight line or streamline, and the possibility that the acupoint is in the main direction of the analysis point position distribution is greater, then the greater the acupoint skeleton direction consistency value of the target point, the greater the possibility that there is an acupoint in the vicinity of the target point.

[0071] The method for obtaining the consistent value of the acupoint skeleton direction of each analysis point in each image in the image pyramid is the same as the method for obtaining the consistent value of the acupoint skeleton direction of the target point.

[0072] Step S3: Analyze the number and position distribution of feature points within a preset neighborhood range of each reference skeleton point to obtain the acupoint spatial position value of each reference skeleton point in each image in the image pyramid.

[0073] Refer to the distribution of acupoints in the shoulder joint, elbow joint and knee: the acupoints around the shoulder, such as Jianjing point and Jianliao point, are usually located near the shoulder joint; the acupoints around the elbow joint, such as Quchi point and Tianjing point, are usually located on the surface or near the elbow bone; the acupoints around the knee, such as Xiyan point and Dubi point, are usually located on both sides or below the knee joint. These acupoints depend on the knee bones for distribution, and it is concluded that the acupoints are near the key bones, and the acupoints near the key bones are distributed in a certain direction of the bones.

[0074] Since acupoints are usually located near bone structures, the position distribution of analysis feature points representing small bones within the preset neighborhood of the benchmark bone point can be used to present the distribution of acupoints near key bones; the number of distributed feature points within the preset neighborhood of the benchmark bone point presents the importance of the benchmark bone point in the analysis of human body detailed structure. The above two factors are combined to analyze the spatial position value of the acupoints, thereby improving the accuracy of the distribution of acupoints near the benchmark bone point.

[0075] Preferably, in some possible implementation modes of the embodiments of the present invention, the method for obtaining the spatial position value of the acupoint includes: for each image in the image pyramid, recording the analysis feature points within the preset neighborhood range of each reference bone point of the image as the neighborhood point of each reference bone point; for each reference bone point in the image, optionally recording a neighborhood point of the reference bone point as a reference point, selecting a neighborhood point closest to the reference point from the neighborhood points of the reference bone point as a judgment point of the reference point, recording the angle between the line segment connecting the reference point and its judgment point and the preset direction as the judgment angle of the reference point, and obtaining the judgment angle of the judgment point; recording the absolute value of the difference between the judgment angles of the reference point and its judgment point as the direction consistency value of the reference point; taking the cumulative sum of the direction consistency values ​​of all the neighborhood points of each reference bone point as the neighborhood direction value of the corresponding reference bone point; and obtaining the spatial position value of the acupoint corresponding to the reference bone point according to the total number of neighborhood points and the neighborhood direction value of each reference bone point in each image in the image pyramid.

[0076] For each reference bone point of each image in the image pyramid, the analysis feature points in the area formed by the reference bone point as the center and the preset length as the radius are recorded as the neighborhood points of the reference bone point. It should be noted that the preset length in the embodiment of the present invention is 10, and the implementer can set it according to the specific situation. If a certain analysis feature point is within the preset neighborhood range of multiple reference bone points, the distance between the analysis feature point and these reference bone points is calculated, and the analysis feature point is used as the neighborhood point of the reference bone point corresponding to the minimum distance.

[0077] It is known that acupuncture points are usually located near bones, and acupuncture points near key bones are distributed in a certain direction of bones. For the neighborhood points of the reference bone points, this scheme obtains the distribution of acupuncture points in the local range of each neighborhood point by analyzing the position distribution of other neighborhood points that are close to each neighborhood point. The difference between the judgment angle of the reference point and the judgment angle of the reference point's judgment point is used to measure the position distribution of the analysis feature points in the local area of ​​the reference point; if the difference between the judgment angles of the reference point and its judgment point, that is, the direction consistency value, is smaller, it means that the characteristics of the position distribution of the reference point and the neighborhood points in its local range in a certain direction are more obvious, and the position distribution of the analysis feature points in the local area of ​​the reference point is more consistent with the distribution characteristics of acupuncture points near key bones. The neighborhood direction value comprehensively analyzes the distribution characteristics of the analysis feature points in the local area of ​​all neighborhood points of the reference bone point; if the neighborhood direction value is smaller, it means that the position distribution of the analysis feature points in the local area of ​​all neighborhood points is more consistent with the distribution characteristics of acupuncture points near bones, and the characteristics of the position distribution of the neighborhood points of the reference bone point along a certain direction are more obvious. It should be noted that the method for obtaining the direction consistency value of each neighborhood point of the reference bone point is the same as that of the reference point.

[0078] It is known that there are usually more acupuncture points around the bones of key parts, and the acupuncture points are near the bones; if the total number of neighborhood points of the reference bone point is larger, the reference bone point is more important in the human body structure analysis, and the credibility of the position distribution characteristics of the neighborhood points of the reference bone point is higher.

[0079] In summary, if the total number of neighborhood points of the reference skeleton point is larger, and the position distribution of the neighborhood points is more regular, that is, the neighborhood direction value is smaller, the reference skeleton point corresponds to the acupoint more accurately, and the acupoint spatial distribution of the reference skeleton point is more accurate, then the acupoint spatial position value is larger, and the possibility that there are acupoints in the vicinity of the reference skeleton point is greater. Therefore, the total number of neighborhood points of each reference skeleton point is positively correlated with the acupoint spatial position value, and the neighborhood direction value is negatively correlated with the acupoint spatial position value.

[0080] It should be noted that the judgment angle of the judgment point of the reference point is similar to the method for obtaining the judgment angle of the reference point. The difference is: the judgment point of the reference point is selected from the other neighborhood points of the benchmark skeleton point except the reference point, while the judgment point of the reference point is selected from the remaining neighborhood points based on the skeleton point except the reference point and its judgment point. The other steps are the same.

[0081] In a specific implementation of the embodiment of the present invention, for each reference skeleton point in each image in the image pyramid in the human body depth image, the acupoint spatial position value of the reference skeleton point is expressed by the formula:

[0082]

[0083] Where D is the spatial position value of the acupoint of each reference bone point; Num is the total number of neighboring points of each reference bone point; U is the total number of neighboring points of each reference bone point; is the judgment angle of the uth neighboring point of each reference skeleton point; is the judgment angle of the judgment point of the u-th neighborhood point of each reference skeleton point; is the direction consistency value of the u-th neighborhood point of each reference skeleton point; is the absolute value function; exp is the exponential function with a natural constant as the base.

[0084] Step S4: According to the acupoint skeleton direction consistency value of each analysis point and the acupoint spatial position value of the reference skeleton point corresponding to the preset neighborhood range of each analysis point, the acupoint registration domain in each image in the image pyramid is obtained; based on the overlapping area of ​​the acupoint registration domains of all images in the image pyramid, the user's acupoints are located.

[0085] By analyzing the distribution characteristics of acupoints around different bones, the acupoint-bone direction consistency value and the acupoint-spatial position value are obtained, and then the possibility of the existence of acupoints in the vicinity of the reference bone point and the analysis feature point is determined, and the acupoint-bone association value is obtained to help users locate acupoints.

[0086] Preferably, in some possible implementation modes of the embodiments of the present invention, the method for obtaining the acupoint-skeletal association value includes: for each analysis point of each image in the image pyramid, normalizing the product of the acupoint-skeletal direction consistency value of the analysis point and the acupoint spatial position value of the reference bone point corresponding to the preset neighborhood range of the analysis point, to obtain the acupoint-skeletal association value of the analysis point.

[0087] The larger the acupoint skeleton direction consistency value is, the greater the possibility that the acupoint is in the main direction of the analysis point position distribution is, and the greater the possibility that the acupoint exists in the vicinity of the target point is; the larger the acupoint spatial position value is, the more accurate the acupoint spatial distribution of the reference skeleton point is, and the greater the possibility that the acupoint exists in the vicinity of the reference skeleton point is. Therefore, the acupoint skeleton direction consistency value and the acupoint spatial position value are both positively correlated with the acupoint skeleton association value.

[0088] It should be noted that in the embodiment of the present invention, the Norm function is used for normalization processing. In the embodiment of the present invention, other normalization methods may also be selected, such as function conversion, maximum and minimum normalization, etc., which are not limited here. For each image in the image pyramid of the human body depth image, if there is an analysis point in the image that is not within the preset neighborhood range of all the reference bone points, then the analysis point does not need to be subsequently analyzed.

[0089] Since the possibility of acupoints existing near the analysis point with a larger acupoint-bone correlation value is greater, the analysis point with acupoint-bone correlation value greater than the preset correlation threshold is recorded as the standard point; the preset length is weighted by the acupoint-bone correlation value of each standard point to obtain the adjusted length of each standard point; the area formed by taking each standard point as the center and the adjusted length as the radius is used as the acupoint registration domain of each image in the image pyramid. There are acupoints around the standard point.

[0090] It should be noted that in the embodiment of the present invention, the preset association threshold is an empirical value of 0.9, and the preset length is an empirical value of 10, which can be set by the implementer according to the specific situation.

[0091] The largest image in the image pyramid is the human body depth image; any acupoint registration domain is selected from the human body depth image and recorded as the target domain, and the acupoint registration domain where the matching pixel points corresponding to the pixel points in the target domain in each target image are located is recorded as the analysis registration domain of the target domain, and the target image where the analysis registration domain is located is recorded as the analysis image of the target domain; if the total number of analysis images of the target domain is greater than the preset number threshold, then for the analysis registration domain of each analysis image of the target domain, a curve fitting is performed on the pixel points at the center position of the matching pixel points corresponding to each edge pixel point on the edge of the analysis registration domain in the human body depth image, and the closed area formed by the fitted curve is recorded as the alignment area of ​​the analysis registration domain in the human body depth image; the part that exists in common in the target domain and the alignment areas of all its analysis registration domains is taken as the user's acupoint area.

[0092] It should be noted that the largest human depth image and the smallest reference structure image in the image pyramid do not participate in the analysis of the analysis image of the target domain, so the maximum total number of analysis images of the target domain in this embodiment is 18. The preset number threshold is half of the total number of images in the image pyramid. In this embodiment, the preset number threshold is an empirical value of 10, which can be set by the implementer according to the specific situation.

[0093] It should be noted that for each target image, if the matching pixel points corresponding to the pixel points in the target domain in the target image are in multiple acupoint registration domains, the number of matching pixel points corresponding to all pixel points in the target domain in the target image in each acupoint registration domain is counted, and the acupoint registration domain with the largest number is taken as the analysis registration domain of the target domain in the target image.

[0094] Different images in the image pyramid have different scales. In order to overlap the target domain with its analysis and registration domain, it is necessary to map the analysis and registration domain of the target domain to the human depth image to obtain the alignment area of ​​the analysis and registration domain in the human depth image.

[0095] It should be noted that this embodiment uses the least square method for curve fitting, which is a well-known technique and will not be described in detail here. The common part of the target domain and the alignment areas of all its analysis and registration domains is equivalent to the intersection of the alignment areas of the target domain and all its analysis and registration domains.

[0096] This scheme performs acupoint detection on multi-scale images of human body depth images. Higher-resolution image layers can provide more refined local features, while lower-resolution image layers can capture a wider range of global information, thereby improving the reliability of acupoint area determination. Acupoints are confirmed by overlapping areas in multi-scale images, effectively avoiding misjudgment or omissions that may occur at a single scale, and effectively reducing the impact of individual differences on acupoint positioning.

[0097] So far, the present invention is completed.

[0098] Embodiment 2:

[0099] The present invention proposes an intelligent acupoint positioning system for Chinese and Western medicine massage equipment. Figure 2 , which shows a system structure diagram of an acupoint intelligent positioning system for Chinese and Western medicine massage equipment provided by an embodiment of the present invention, the system comprising:

[0100] The data acquisition module 510 is used to obtain a human body depth image of a user using the Chinese and Western medicine massage device;

[0101] The direction analysis module 520 is used to obtain an image pyramid of a human body depth image, as well as reference skeleton points and analysis feature points of each image in the image pyramid; the reference skeleton points and the analysis feature points are recorded as analysis points, and according to the position distribution of each analysis point of each image in the image pyramid compared with the remaining analysis points, the acupoint skeleton direction consistency value of each analysis point is obtained;

[0102] A spatial analysis module 530 is used to analyze the number and position distribution of feature points within a preset neighborhood range of each reference bone point, and obtain the acupoint spatial position value of each reference bone point in each image in the image pyramid;

[0103] The acupoint positioning module 540 is used to obtain the acupoint registration domain in each image in the image pyramid based on the acupoint skeleton direction consistency value of each analysis point and the acupoint spatial position value of the reference skeleton point corresponding to the preset neighborhood range of each analysis point; and locate the acupoints of the human body based on the overlapping area of ​​the acupoint registration domains of all images in the image pyramid.

[0104] It should be noted that the device provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the acupoint intelligent positioning system for Chinese and Western medicine massage equipment and the acupoint intelligent positioning method embodiment for Chinese and Western medicine massage equipment provided in the above embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0105] Embodiment 3:

[0106] Figure 3 A computer device schematic diagram of an acupoint intelligent positioning device for Chinese and Western medicine massage equipment provided by an embodiment of the present invention. Figure 3 As shown, the computer device includes: a memory 601, a processor 602, and a computer program 603 stored in the memory 601 and running on the processor 602, wherein when the processor 602 executes the computer program 603, the computer device can execute any one of the aforementioned intelligent acupoint positioning methods for Chinese and Western medicine massage equipment.

[0107] In addition, an embodiment of the present application also protects a device, which may include a memory and a processor, wherein the memory stores executable program code, and the processor is used to call and execute the executable program code to execute an intelligent acupoint positioning method for Chinese and Western medicine massage equipment provided in an embodiment of the present application.

[0108] In this embodiment, the functional modules of the device can be divided according to the above method example. For example, each functional module can be corresponded, or two or more functions can be integrated into one processing module. The above integrated module can be implemented in the form of hardware. It should be noted that the division of modules in this embodiment is schematic and is only a logical function division. There may be other division methods in actual implementation.

[0109] It should be understood that the device provided in this embodiment is used to execute the above-mentioned intelligent acupoint positioning method for Chinese and Western medicine massage equipment, and thus can achieve the same effect as the above-mentioned implementation method.

[0110] In the case of an integrated unit, the device may include a processing module and a storage module. When the device is applied to a device, the processing module may be used to control and manage the actions of the device. The storage module may be used to support the device to execute mutual program codes, etc.

[0111] The processing module may be a processor or a controller, which may implement or execute various exemplary logic blocks, modules and circuits included in the disclosure of this application. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and a microprocessor, etc. The storage module may be a memory.

[0112] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0113] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0114] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent acupoint positioning method for Chinese and Western medicine massage equipment, characterized in that: The method includes: Acquire a human body depth image of a user using a Chinese and Western medicine massage device; Obtain an image pyramid of a human body depth image, as well as reference skeleton points and analysis feature points of each image in the image pyramid; record the reference skeleton points and the analysis feature points as analysis points, and obtain the acupoint skeleton direction consistency value of each analysis point according to the position distribution of each analysis point in each image in the image pyramid compared with the remaining analysis points; Analyze the number and position distribution of feature points within a preset neighborhood of each reference bone point to obtain the acupoint spatial position value of each reference bone point in each image in the image pyramid; According to the acupoint skeleton direction consistency value of each analysis point and the acupoint spatial position value of the reference skeleton point corresponding to the preset neighborhood range of each analysis point, the acupoint registration domain in each image in the image pyramid is obtained; based on the overlapping area of ​​the acupoint registration domains of all images in the image pyramid, the acupoints of the user are located; The method for obtaining the reference skeleton points and analysis feature points of each image in the image pyramid includes: Selecting a reference structure image from an image pyramid; matching the reference structure image with a standard human body model to obtain key bone points in the reference structure image; Select one image from the remaining images in the image pyramid except the reference structure image as the target image, and record the pixel point at the center position of the matching pixel point corresponding to each key skeleton point of the reference structure image in the target image as the reference skeleton point of the target image; Perform feature extraction on each image in the image pyramid to obtain feature points in each image; record the pixel points at the center position of the matching pixel points corresponding to the feature points in the target image and the feature points in other images in the image pyramid except the reference structure image and the target image in the target image as the analysis feature points of the target image; The step of obtaining the acupoint-bone direction consistency value of each analysis point includes: For each image in the image pyramid, an analysis point is randomly selected from the image and recorded as the target point. The line segments obtained by connecting the target point with the remaining analysis points in the image except the target point are recorded as the analysis line segments of the target point. The angle between the analysis line segment and the preset direction is used as the distribution angle of each analysis line segment of the target point. The cumulative sum of the absolute values ​​of the distribution angle differences of the pairwise analysis line segments of the target point is negatively correlated to obtain the acupoint bone direction consistency value of the target point; The step of obtaining the acupuncture point spatial position value of each reference skeleton point of each image in the image pyramid includes: For each image in the image pyramid, the analysis feature points within the preset neighborhood range of each reference skeleton point of the image are recorded as the neighborhood points of each reference skeleton point; For each reference bone point in the image, a neighborhood point of the reference bone point is randomly selected and recorded as a reference point, a neighborhood point closest to the reference point is selected from the neighborhood points of the reference bone point and recorded as a judgment point of the reference point, an angle between a line segment connecting the reference point and its judgment point and a preset direction is recorded as a judgment angle of the reference point, and the judgment angle of the judgment point is obtained; The absolute value of the difference between the judgment angle of the reference point and its judgment point is recorded as the direction consistency value of the reference point; the cumulative sum of the direction consistency values ​​of all neighboring points of each reference skeleton point is used as the neighborhood direction value of the corresponding reference skeleton point; According to the total number of neighborhood points of each reference skeleton point in each image in the image pyramid and the neighborhood direction value, the acupoint spatial position value of the corresponding reference skeleton point is obtained; the total number of neighborhood points of each reference skeleton point is positively correlated with the acupoint spatial position value, and the neighborhood direction value is negatively correlated with the acupoint spatial position value; The step of obtaining the acupoint registration domain in each image in the image pyramid comprises: For each analysis point of each image in the image pyramid, the product of the acupoint-bone direction consistency value of the analysis point and the acupoint spatial position value of the reference bone point corresponding to the preset neighborhood range of the analysis point is normalized to obtain the acupoint-bone association value of the analysis point; Record the analysis points whose acupoint-bone correlation values ​​are greater than the preset correlation threshold as standard points; weight the preset length using the acupoint-bone correlation value of each standard point to obtain the adjusted length of each standard point; The area formed by taking each standard point as the center and the adjustment length as the radius is used as the acupoint registration domain of each image in the image pyramid.

2. The intelligent acupoint positioning method for Chinese and Western medicine massage equipment according to claim 1, characterized in that: The method of locating the acupuncture points of the human body based on the overlapping areas of the acupuncture point registration domains of all images in the image pyramid includes: The largest image in the image pyramid is the human body depth image; Selecting an acupoint registration domain from the human body depth image as the target domain, recording the acupoint registration domain where the pixel points in the target domain correspond to the matching pixel points in each of the target images as the analysis registration domain of the target domain, and recording the target image where the analysis registration domain is located as the analysis image of the target domain; If the total number of analysis images of the target domain is greater than a preset number threshold, then for the analysis registration domain of each analysis image of the target domain, a curve fitting is performed on the pixel point at the center position of the matching pixel point corresponding to each edge pixel point on the edge of the analysis registration domain in the human body depth image, and the closed area formed by the fitting curve is recorded as the alignment area of ​​the analysis registration domain in the human body depth image; The part that exists in common between the target domain and the aligned regions of all its analyzed registration domains is regarded as the user's acupoint region.

3. The intelligent acupoint positioning method for Chinese and Western medicine massage equipment according to claim 1, characterized in that: The reference structure image is an image with the smallest size in the image pyramid.

4. The intelligent acupoint positioning method for Chinese and Western medicine massage equipment according to claim 1, characterized in that: The downsampling ratio in the image pyramid construction process is an odd number.

5. The intelligent acupoint positioning method for Chinese and Western medicine massage equipment according to claim 1, characterized in that: The method for obtaining feature points in each image is a scale-invariant feature transformation algorithm.

6. An intelligent acupoint positioning system for Chinese and Western medicine massage equipment, characterized in that: The system includes: A data acquisition module for acquiring a human body depth image of a user using the Chinese and Western medicine massage device; A direction analysis module is used to obtain an image pyramid of a human body depth image, as well as reference skeleton points and analysis feature points of each image in the image pyramid; the reference skeleton points and the analysis feature points are recorded as analysis points, and according to the position distribution of each analysis point of each image in the image pyramid compared with the remaining analysis points, the acupoint skeleton direction consistency value of each analysis point is obtained; A spatial analysis module, used to analyze the number and position distribution of feature points within a preset neighborhood of each reference bone point, and obtain the acupoint spatial position value of each reference bone point in each image in the image pyramid; an acupoint positioning module, for obtaining the acupoint registration domain in each image in the image pyramid according to the acupoint skeletal direction consistency value of each analysis point and the acupoint spatial position value of the reference skeletal point corresponding to the preset neighborhood range of each analysis point; and locating the acupoints of the human body based on the overlapping areas of the acupoint registration domains of all images in the image pyramid; The method for obtaining the reference skeleton points and analysis feature points of each image in the image pyramid includes: Selecting a reference structure image from an image pyramid; matching the reference structure image with a standard human body model to obtain key bone points in the reference structure image; Select one image from the remaining images in the image pyramid except the reference structure image as the target image, and record the pixel point at the center position of the matching pixel point corresponding to each key skeleton point of the reference structure image in the target image as the reference skeleton point of the target image; Perform feature extraction on each image in the image pyramid to obtain feature points in each image; record the pixel points at the center position of the matching pixel points corresponding to the feature points in the target image and the feature points in other images in the image pyramid except the reference structure image and the target image in the target image as the analysis feature points of the target image; The step of obtaining the acupoint-bone direction consistency value of each analysis point includes: For each image in the image pyramid, an analysis point is randomly selected from the image and recorded as the target point. The line segments obtained by connecting the target point with the remaining analysis points in the image except the target point are recorded as the analysis line segments of the target point. The angle between the analysis line segment and the preset direction is used as the distribution angle of each analysis line segment of the target point. The cumulative sum of the absolute values ​​of the distribution angle differences of the pairwise analysis line segments of the target point is negatively correlated to obtain the acupoint bone direction consistency value of the target point; The step of obtaining the acupuncture point spatial position value of each reference skeleton point of each image in the image pyramid includes: For each image in the image pyramid, the analysis feature points within the preset neighborhood range of each reference skeleton point of the image are recorded as the neighborhood points of each reference skeleton point; For each reference bone point in the image, a neighborhood point of the reference bone point is randomly selected and recorded as a reference point, a neighborhood point closest to the reference point is selected from the neighborhood points of the reference bone point and recorded as a judgment point of the reference point, an angle between a line segment connecting the reference point and its judgment point and a preset direction is recorded as a judgment angle of the reference point, and the judgment angle of the judgment point is obtained; The absolute value of the difference between the judgment angle of the reference point and its judgment point is recorded as the direction consistency value of the reference point; the cumulative sum of the direction consistency values ​​of all neighboring points of each reference skeleton point is used as the neighborhood direction value of the corresponding reference skeleton point; According to the total number of neighborhood points of each reference skeleton point in each image in the image pyramid and the neighborhood direction value, the acupoint spatial position value of the corresponding reference skeleton point is obtained; the total number of neighborhood points of each reference skeleton point is positively correlated with the acupoint spatial position value, and the neighborhood direction value is negatively correlated with the acupoint spatial position value; The step of obtaining the acupoint registration domain in each image in the image pyramid comprises: For each analysis point of each image in the image pyramid, the product of the acupoint-bone direction consistency value of the analysis point and the acupoint spatial position value of the reference bone point corresponding to the preset neighborhood range of the analysis point is normalized to obtain the acupoint-bone association value of the analysis point; Record the analysis points whose acupoint-bone correlation values ​​are greater than the preset correlation threshold as standard points; weight the preset length using the acupoint-bone correlation value of each standard point to obtain the adjusted length of each standard point; The area formed by taking each standard point as the center and the adjustment length as the radius is used as the acupoint registration domain of each image in the image pyramid.

Images