Human body meridian and acupoint automatic detection method and system based on AI machine vision

Abstract

The application discloses an automatic human body meridian and acupoint detection method and system based on AI machine vision, relates to the cross field of artificial intelligence and medical health, and comprises the following steps: hand surface direction judgment, judging the palm surface or the back of the hand according to the continuous distribution curve extraction result in the hand image; hand surface feature extraction, segmenting the finger area and the palm area according to the finger root horizontal lines, and extracting feature marks according to the corresponding strategy selected according to the hand surface; global reference establishment, constructing a coordinate system with the midpoint of the wrist horizontal line as the global origin, segmenting the fine area and forming a mapping library; local adaptation, establishing a local coordinate system and calculating the offset from the global origin; candidate point screening, screening and verifying the optimal acupoint from the fine area according to multiple features, and the system comprises the above functional modules; the method improves the automation degree, positioning accuracy and adaptability of acupoint detection, and meets the needs of traditional Chinese medicine diagnosis and treatment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence and medical health, more specifically, it relates to an automatic detection method and system for human meridian and acupoint based on AI machine vision. BACKGROUND

[0002] With the rapid development of artificial intelligence and machine vision technology, the automatic detection method for human meridian and acupoint based on AI machine vision has gradually become an important direction of modernization research of traditional Chinese medicine. This kind of technology can realize the accurate positioning of human meridian and acupoint, and provide technical support for traditional Chinese medicine diagnosis and treatment. However, the existing related technology still has deficiencies in automation degree, positioning accuracy and adaptability, which limits its popularization and application in practical application.

[0003] Through retrieval, a preparation method of a customized mask device based on visual three-dimensional reconstruction is disclosed in patent CN111639553B. This patent marks the facial acupoint points to make a customized mask device that fits the user's face by taking multiple face pictures and combining three-dimensional reconstruction technology. However, this technical solution mainly relies on manual marking of acupoint points, lacks automatic recognition ability for meridian and acupoint, resulting in low efficiency and accuracy limited by manual operation accuracy.

[0004] In addition, a mobile joint type pulse diagnosis instrument is disclosed in patent CN115778320B. This patent realizes automatic recognition and pressing of the inch, Guan and chi pulse points through the image acquisition module and mechanical arm module combined with visual positioning technology. However, this technical solution mainly targets the positioning of specific acupoints on the wrist, with limited application scope, and is difficult to cover the meridian and acupoints of other parts of the human body. At the same time, this method relies on blood vessel imaging technology, and has weak detection ability for non-blood vessel related acupoints, which cannot meet the needs of diversified acupoint detection in traditional Chinese medicine meridian theory.

[0005] The above problems show that the existing acupoint positioning method based on visual technology still has significant deficiencies in automation degree, application scope and adaptability to diversified acupoints. Therefore, the present application provides an automatic detection method for human meridian and acupoint based on AI machine vision, which aims to realize high-precision and automatic detection of human meridian and acupoint throughout the body through intelligent algorithm, improve detection efficiency and applicability, and meet the needs of efficient and accurate acupoint detection technology for modern traditional Chinese medicine diagnosis and treatment. SUMMARY

[0006] In view of the deficiencies of the prior art, the purpose of the present application is to solve the deficiencies of the existing acupoint positioning method based on visual technology in terms of automation degree, application scope and adaptability to diversified acupoints, and therefore to provide an AI machine vision-based automatic detection method for human hand meridian acupoints, which realizes high-precision and automatic detection of human hand meridian acupoints through AI machine vision technology, improves positioning efficiency and applicability, and meets the needs of modern Chinese medicine diagnosis and treatment.

[0007] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0008] The AI machine vision-based automatic detection method for human hand meridian acupoints comprises the following steps:

[0009] The hand surface direction judgment step judges whether a continuous distribution curve can be extracted in the hand image collected by the vision camera, and outputs the palm surface or the back surface according to the extraction result;

[0010] The hand surface feature extraction step divides the hand image into a finger area and a palm area according to the finger root horizontal lines, and selects a corresponding feature extraction strategy according to the surface of the hand image to extract feature marks in the hand image;

[0011] The global reference establishment step selects the midpoint of the wrist horizontal line as the global origin, defines the horizontal and vertical coordinate axes, and divides the hand image according to the feature marks to obtain a refined area, extracts the boundary coordinates of the refined area in the global coordinate system, and forms a mapping library;

[0012] The local adaptation step establishes a local coordinate system with the geometric center of each refined area as the local origin, calculates the offset between the local origin and the global origin, and forms a coordinate conversion, so that the local positioning result can be directly mapped to the global space;

[0013] The candidate point screening step extracts the spatial relationship parameters of the refined area and the feature marks as the initial conditions for subsequent candidate point screening, and combines the line intersection feature, the gray gradient feature and the acupoint group correlation verification to screen out candidate points that meet the proportion from the refined area, and the optimal candidate point is reserved as the target point through concave-convex feature matching verification.

[0014] Further, when the hand image is the palm surface, the feature extraction strategy comprises extracting the palm center main line by using an edge detection algorithm, the palm center main line reflects three feature lines, the palm area is divided into a plurality of macro areas according to the spatial distribution of the palm center main line, and the boundary coordinates are output, and then the macro areas are subjected to palm muscle texture analysis, the muscle texture is extracted by Hough transform, the muscle texture comprises longitudinal, transverse and radial directions, and the preset palm center acupoints are one-to-one mapped into each macro area to obtain acupoint falling area information.

[0015] Further, when the hand image is a dorsal surface, the feature extraction strategy includes using a contour detection algorithm to identify the metacarpal bone apex and the dorsal wrist transverse lines, and using a texture tracking algorithm to analyze the local gradient direction and intensity change of the tendon to extract the texture direction of the extensor tendon, and mapping the preset dorsal acupoints to each macro area to obtain acupoint falling area information.

[0016] Further, the global reference establishment step includes a refined region segmentation strategy, the refined region segmentation strategy includes a palm surface segmentation step and a dorsal surface segmentation step,

[0017] The palm surface segmentation step further divides the macro area into a refined region by combining the texture arrangement direction with the number of texture intersection points per square millimeter;

[0018] The dorsal surface segmentation step divides the macro area into a refined region by combining the metacarpal bone apex and the dorsal wrist transverse lines with the tendon direction and bony landmarks.

[0019] Further, the global reference establishment step includes a reference establishment strategy, the reference establishment strategy includes a coordinate axis construction step and a boundary coordinate determination step,

[0020] The coordinate axis construction step defines the midpoint of the wrist transverse line as the horizontal coordinate axis and the midpoint of the wrist transverse line as the vertical coordinate axis;

[0021] The boundary coordinate determination step identifies the contour edge of each refined region by an edge detection algorithm, and obtains the coordinate values of at least 4 vertices as the boundary coordinates.

[0022] Further, in the local adaptation step, the average value of the horizontal coordinates of all boundary vertices of the refined region is taken as the horizontal coordinate of the local origin, and the average value of the vertical coordinates of all boundary vertices of the refined region is taken as the vertical coordinate of the local origin, and the offset includes a horizontal offset and a vertical offset.

[0023] Further, the spatial relationship parameters include distance parameters and angle parameters, and the candidate point screening step further includes a multi-feature fusion screening strategy, the multi-feature fusion screening strategy includes a texture intersection matching step and a gray gradient analysis step,

[0024] The texture intersection matching step identifies the intersection points, turning points and dense texture clusters of the palm print texture in the refined region, calculates the Euclidean distance of the candidate points from the region boundary and the angle with the feature line, and screens out the top 5 candidate points that meet the preset proportion,

[0025] The gray gradient analysis step retains the candidate points with ring-shaped gradient characteristics according to the gradient change of the candidate points in the hand image.

[0026] Further, the multi-feature fusion screening strategy further includes an acupoint group correlation verification step;

[0027] The acupoint group correlation verification step verifies the spatial topological relationship of the candidate points according to the distribution position correlation between the acupoints in the palm region, and eliminates the candidate points whose relative position deviation from adjacent acupoints exceeds a preset threshold.

[0028] Further, the multi-feature fusion screening strategy further includes a depth verification step;

[0029] The depth verification step samples the depth values of the screened candidate points, extracts their three-dimensional coordinates and analyzes the depth value distribution, and retains the optimal candidate points that meet the conditions through concave-convex feature recognition and feature matching verification.

[0030] The human meridian acupoint automatic detection system based on AI machine vision includes:

[0031] The hand surface direction judgment module judges whether a continuous distribution curve can be extracted in the hand image collected by the vision camera, and outputs the palm surface or the back surface according to the extraction result;

[0032] The hand surface feature extraction module divides the hand image into a finger region and a palm region according to the finger root horizontal lines, and selects a corresponding feature extraction strategy according to the surface of the hand image to extract feature marks in the hand image;

[0033] The global reference establishment module selects the midpoint of the wrist horizontal line as the global origin, defines the horizontal and vertical coordinate axes, and divides the hand image according to the feature marks to obtain a refined region, extracts the boundary coordinates of the refined region in the global coordinate system, and forms a mapping library;

[0034] The local adaptation module establishes a local coordinate system for each refined region with its geometric center as the local origin, calculates the offset between the local origin and the global origin, and forms a coordinate conversion to enable the local positioning result to be directly mapped to the global space;

[0035] The candidate point screening module extracts the spatial relationship parameters of the refined region and the feature marks as initial conditions for subsequent candidate point screening, and screens candidate points that meet the proportion from the refined region in combination with the line intersection feature, the gray gradient feature and the acupoint group correlation verification, and retains the optimal candidate point as the target point through concave-convex feature matching verification of the candidate points.

[0036] The beneficial effects of the present application are: 1. The whole process of hand meridian and acupoint automatic detection is realized by AI machine vision technology, and manual intervention is not needed from hand direction judgment, feature extraction to candidate point screening, which completely solves the problems of low efficiency and poor consistency caused by the dependence of traditional technology on manual marking, and specifically, the palm surface and the back of the hand are accurately distinguished through the hand direction judgment step, and targeted feature extraction strategies are adopted, such as extracting feature lines through edge detection for the palm surface, and identifying metacarpal vertex and tendon texture through contour detection for the back of the hand, and the global reference and local coordinate system are cooperatively constructed to realize hierarchical segmentation from macro area to fine area, ensuring the uniformity of spatial reference for acupoint positioning; at the same time, the candidate point screening link integrates cross feature, gray gradient feature, acupoint group correlation verification and depth concave-convex feature verification, and multi-dimensional interference points are removed, so that the positioning error of the optimal acupoint is controlled within a subtle range, which is significantly superior to the positioning accuracy of the prior art, and meets the strict requirements of traditional Chinese medicine diagnosis and treatment on acupoint accuracy;

[0037] 2. Through the multi-level adaptive mechanism, the influence of different individual hand shape differences (such as palm size, texture depth, tendon variation, etc.) and complex imaging environment is effectively responded, in the global reference establishment step, the midpoint of the wrist transverse line is taken as the origin to construct a standardized coordinate system, and the boundary coordinate mapping library of the feature mark is combined to provide a unified spatial reference for the hand images of different individuals; the local adaptive step realizes the accurate conversion of local and global coordinates through the geometric center offset amount calculation, so that the positioning result of each fine area can be mapped to the global space, avoiding the positioning deviation caused by the individual hand proportion difference, in addition, different feature extraction strategies are designed for the palm surface and the back of the hand (such as focusing on palm muscle texture analysis for the palm surface and focusing on tendon and bony mark identification for the back of the hand), and the region segmentation granularity is dynamically adjusted through parameters such as texture intersection point density and muscle texture direction, to ensure the full coverage of diversified acupoints (such as Laogong acupoint on the palm and Hegu acupoint on the back of the hand), this adaptability not only expands the scope of the technology for the applicable population, but also improves the robustness under different light and shooting angles, providing reliable technical support for traditional Chinese medicine clinical diagnosis and treatment, family health monitoring and other scenes. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is the whole flowchart in the present application;

[0039] Figure 2 is the hand surface feature extraction flowchart in the present application;

[0040] Figure 3 is the candidate point multi-feature screening flowchart in the present application;

[0041] Figure 4 is the system module connection diagram in the present application. DETAILED DESCRIPTION

[0042] The application will be further described in detail below in combination with the accompanying drawings and embodiments. Identical parts are denoted by identical reference numerals. It should be noted that the words "front", "back", "left", "right", "up" and "down" used in the following description refer to the directions in the drawings, and the words "bottom" and "top", "inner" and "outer" refer to the directions towards or away from the geometric center of a particular part.

[0043] The application provides an AI machine vision-based automatic detection method and system for human meridians and acupoints, the specific implementation manner of which is described in detail in combination with the accompanying drawings in the description, and the following content will be gradually unfolded from image acquisition, and the mutual cooperation relationship of each step will be elaborated in detail.

[0044] The hand surface direction judgment step includes hand image acquisition and preprocessing and hand surface direction judgment. First, the hand image of the user is acquired through a visual camera. The visual camera used can be a high-resolution industrial camera, which is installed at a fixed distance from the user's hand to ensure image clarity. In actual operation, the user places his hand on a fixed platform, and a light source is provided below the platform for uniform lighting, thereby reducing the influence of environmental light on image quality. The image is denoised by a Gaussian filter to eliminate the influence of environmental noise, and a grayscale normalization operation is used to reduce the influence of light changes on image quality. The core parameters of the Gaussian filter include kernel size and standard deviation , the kernel size is usually set to 5x5 pixels, , and the standard deviation is dynamically adjusted according to the image noise level, ranging from 0.8 to 1.2. The purpose of grayscale normalization is to map the brightness range of the image to a fixed interval, such as 0 to 255, to improve the contrast of subsequent processing. At this time, the quality of the image has been significantly optimized, providing a foundation for further feature extraction.

[0045] Subsequently, the continuous distribution curve of the palm center main line is extracted by a Canny operator, and if a characteristic line is detected, it is determined to be a palm surface; if not, it can be determined to be a back surface, wherein the characteristic line includes a first characteristic line, a second characteristic line, and a third characteristic line, the first characteristic line is usually the uppermost end of the three characteristic lines in the palm center, and the starting position of the first characteristic line starts from the vicinity of the intersection between the middle finger and the palm area, specifically in the side edge region of the middle finger root and the palm surface connection, the trend and form of which is an overall curved form, which gradually extends away from the thumb after starting from the starting point, and finally stops at the edge position of the palm center away from the thumb, and the first characteristic line is an important reference for dividing the upper macro region of the palm center; the second characteristic line is the lowest line among the three characteristic lines, and the starting position starts from the region between the index finger and the thumb, more specifically, it is located near the connection region of the index finger root and the thumb root, above the horizontal lines of the intersection between the palm area and the finger area, the trend and form of which is to extend in the horizontal line direction of the palm area in a curved posture, which is a key sign for distinguishing the lower macro region of the palm center; the relative position of the second characteristic line is between the first characteristic line and the third characteristic line, and the starting position starts from the edge of the third finger joint cavity of the index finger, and also starts from between the index finger and the thumb, the trend and form of which is to extend to the lower side of the middle finger, ring finger or little finger, and the common termination positions include the middle line of the ring finger, the lower side of the ring finger and the lower side of the little finger, which is an important basis for dividing the middle macro region of the palm center.

[0046] Due to certain reasons, the characteristic line is not recognized, in order to improve the judgment accuracy, the texture tracking algorithm is switched to extract the extensor tendon texture (the texture tracking algorithm uses an improved SUSAN corner detection combined with a gradient descent method, the local gradient direction calculation window is set to 3x3 pixels, and the intensity change threshold is set to 20-30 gray values), the texture tracking algorithm identifies the tendon texture by analyzing the local gradient direction and intensity change, and matches and verifies it in combination with the metacarpal bony markers, the specific matching process includes calculating the spatial consistency of the texture direction and the bony marker point, if the consistency is higher than a set threshold (for example, 90%), it is determined to be a back surface; this process ensures that the image can be accurately determined to belong to a palm surface or a back surface, and during the image acquisition process, the shooting angle of the camera is calibrated according to the position of the reference point mapped in the hand image, to ensure the geometric consistency of the image.

[0047] The hand feature extraction step includes region segmentation and feature extraction, such as Figure 2As shown, first, in the region segmentation, the YOLO or Faster R-CNN target detection algorithm is used to segment the finger region and the palm region, the finger root transverse line is taken as the boundary line, the distal finger region and the proximal palm region are divided, and the boundary coordinates of the finger region and the palm region and the area ratio of the two in the whole image are recorded when the segmentation result is output. The position coordinates of the finger root transverse line extracted by the target detection algorithm are taken as the boundary line to ensure the accuracy of the region segmentation. In this step, the model training data set of the target detection algorithm needs to contain a large number of labeled clear hand images to improve the segmentation accuracy. In addition, the area ratio calculation formula of the finger region and the palm region is the ratio of the pixel number of each region to the total pixel number of the whole image. This ratio is used in the subsequent feature extraction and reference anchoring steps.

[0048] Subsequently, in the feature extraction, according to the surface selection of the hand image, the corresponding strategy is selected. For the palm surface, the first feature line, the second feature line and the third feature line of the palm center main line are extracted by the Hough transform to preliminarily divide the palm region into macro regions, and then the palm region is further divided into fine regions according to the number of texture intersection points per square millimeter. In addition, the fine regions are divided into multiple fine sub-regions according to the difference in texture density, for example, the thenar eminence region is divided into the thenar medial sub-region and the thenar lateral sub-region.

[0049] For the back of the hand, the contour detection algorithm is used to identify the metacarpal vertex and the wrist back transverse line. The bony landmarks include the metacarpal vertex and the wrist back transverse line. The first to fifth metacarpal distal vertex, which is the protrusion near the finger root, is identified by the contour detection algorithm, and is the key positioning landmark of the skeletal structure of the back of the hand. The wrist back transverse line is a transversely distributed texture on the back of the wrist, and its connection position with the metacarpal base is the proximal boundary of the macro region. The macro region of the back of the hand is preliminarily divided into multiple macro functional areas, such as the region between the first and second metacarpal vertices and the wrist back transverse line.

[0050] Since the macro region can only roughly frame the functional area of the acupoint, but the range is wide, it may contain non-acupoint regions or interference features, therefore, the macro region is divided into fine regions in combination with the direction of the extensor tendon texture and the bony landmark, and then the fine regions are divided into multiple fine sub-regions. Specifically, the division of the fine region needs to rely on three rules. First, the spatial distance correlation, taking the metacarpal vertex as the starting point, calculating the vertical distance between the extensor tendon texture and the metacarpal vertex. Second, the angle correlation, the angle between the extensor tendon texture and the metacarpal axis needs to meet the preset range. Third, the texture density correlation, analyzing the data amount of the tendon texture intersection points to distinguish the gap region between the tendons and the tendon itself region.

[0051] Directly match the specific location of the acupoint, reduce the interference of subsequent candidate point screening, in addition, the hand morphology of different individuals is different, when dividing the refined sub-region, it is not fixed size, but dynamic adjustment based on local features, the back of the hand is divided according to the "relative angle of tendon and bony landmarks", which can adapt to the hand features of different individuals, avoid positioning deviation caused by morphological differences, in addition, the boundary coordinates of the refined sub-region (at least 4 vertices) are stored in the mapping library, which can realize millimeter level coordinate conversion through the offset of local origin and global origin, ensure that the positioning result of each acupoint can be mapped to the global space and can retain local fine features, finally realize the hierarchical precision guarantee of "macro positioning to local refinement and then global unification"; The output results of the hand feature extraction step include not only the texture distribution information, but also the boundary coordinates and geometric center of each sub-region, which are transmitted to the global reference step and the local reference step.

[0052] The distribution of the second feature line and the third feature line in the above and the palm center main line together with the position relationship of the metacarpal vertex and the extensor tendon texture constitute the basis of feature extraction, the key of this step lies in the parameter setting of Hough transform and contour detection algorithm, such as the threshold value of Hough transform and the minimum area limit of contour detection, these parameters directly affect the granularity of feature extraction.

[0053] In the global reference establishment step, the midpoint of the wrist transverse line is selected as the global origin, the X axis and Y axis directions are defined, the unit is unified as millimeter, the boundary coordinates of all refined regions in the global coordinate system are extracted, and they are stored in the mapping library, the definition of global coordinate system makes the positioning of all regions have unified reference, which is convenient for subsequent global positioning operation, at the same time, the global reference step is also responsible for the coordinate transformation of the sub-region data output by the feature extraction step, so that it can adapt to the requirements of global coordinate system.

[0054] In the local adaptation step, the geometric center of each refined region is taken as the local origin to establish the local coordinate system, the local origin coordinate calculation formula is as follows: let the boundary vertex coordinates of the refined region be 、 … , the horizontal coordinate of the local origin is , the vertical coordinate is , the offset is , , the offset of the local origin and the global origin is calculated and the coordinate conversion is formed, the global coordinate ; wherein, and They represent the offset between the local origin and the global origin in the X-axis and Y-axis directions respectively. At the same time, the spatial relationship parameters between the refined area and the feature mark, including distance parameters and angle parameters, are extracted and stored in the benchmark database so that the local positioning results can be directly mapped to the global space. For example, "the vertical distance between the upper edge of the medial sub-area of ​​the thenar eminence and the third feature line is 3-5mm" and "the angle between the central axis of the central area of ​​the palm and the second feature line is 30°±5°". These parameters provide an accurate positioning basis for subsequent candidate point screening.

[0055] like Figure 3 As shown in the figure, in the candidate point screening step, the texture intersection feature, grayscale gradient feature and acupoint group correlation verification are combined to screen candidate points that meet the anatomical proportions from the refined area. For example, the texture intersection dense area between the first and second metacarpal bones is identified as the candidate point in the area where the Hegu point is located. The top five candidate points are screened by calculating the Euclidean distance between the candidate point and the area boundary and the angle with the feature line. The gradient changes around the candidate points are analyzed, and the Sobel operator is used to extract the gradient information. The points that meet the annular gradient characteristics are retained. The specific quantitative indicators of the annular gradient can be the grayscale value difference range, the radius range of the annular distribution, and the angular span of the gradient change. The anatomical correlation of the acupoint group is used to verify the spatial topological relationship of the candidate points, and the points that do not meet the relative position deviation threshold are eliminated.

[0056] In addition, the candidate point screening step also includes depth information assistance, extracting the 3D coordinates of the candidate point, collecting the depth value distribution of a 5×5 pixel area centered on the candidate point, and calculating the standard deviation of the depth value in the area. Identify flatness differences, extract depth mutation points, and determine whether they are concave or convex edges. For concave acupuncture points, set the conditions that the depth value of the center of the area is less than the average depth value of the surrounding area and the depth mutation points are distributed in a ring-shaped manner; for convex acupuncture points, set the conditions that the depth value of the center of the area is greater than the average depth value of the surrounding area and the depth mutation points are distributed along the direction of the bony landmarks, and finally retain the optimal candidate point; among them, the standard deviation σ of the depth value is calculated as the square root of the variance of the depth values ​​in the area.

[0057] In multi-dimensional verification, after obtaining the optimal candidate point, the benchmark anchoring and candidate point screening are repeated, and the two results are compared. If the fluctuation unit is located in the same hand image, the candidate point is determined to be accurate, otherwise it is re-analyzed. The key to this step lies in the determination method of the fluctuation unit and the comparison rules of the two results. For example, the steep peak and trough values ​​of the fluctuation unit are calculated through trigonometric functions, and the relative position relationship between the acupuncture point and the adjacent acupuncture points is obtained by combining the external system library index. This process ensures the stability and reliability of the candidate points.

[0058] In the acupoint correlation feedback step, the peak value and the trough value of the fluctuation unit are calculated, the relative position relationship between the acupoint and the adjacent acupoint is obtained through a trigonometric function calculation or an external adjustment system library index, the relative position relationship is compared with a threshold value, and a monitoring instruction and a danger instruction are output. The key of this step lies in the calculation method of the relative position relationship and the setting rule of the threshold value. For example, the relative position relationship of the adjacent acupoint needs to meet a specific anatomical proportion, otherwise a danger instruction is output to prompt reanalysis.

[0059] Through the above steps, the present application realizes multi-level analysis of the hand image, gradually refines from the global to the local, combines various features such as palm print texture, extensor tendon texture and bony landmarks, and constructs a high-precision acupoint positioning system; through multi-dimensional verification of the candidate points, including the introduction of depth information and the verification of anatomical correlation, the reliability of the positioning result is improved; at the same time, through the automatic algorithm, the diversified acupoints are comprehensively covered, solving the problems of low efficiency and insufficient consistency of traditional manual positioning, meeting the needs of modern Chinese medicine diagnosis and treatment for efficient and accurate acupoint detection.

[0060] Based on the above steps, the present application gives an embodiment including:

[0061] Hand surface direction judgment: 1. Image acquisition and preprocessing: using a 1600 million pixel visual camera (resolution 4608x3456) to collect the user's right palm center surface image (palm center upward, fingers naturally stretched), removing noise through Gaussian filtering (kernel size 5x5, σ=0.9), and normalizing the gray scale to 0-255 interval to eliminate light interference;

[0062] 2. Direction judgment logic: using Canny edge detection algorithm (high threshold 70, low threshold 35) to extract continuous distribution curves in the image, successfully identifying three palm center main lines (third feature line, second feature line, first feature line), and determining as the palm center surface.

[0063] Hand surface feature extraction: 1. Region segmentation: taking the finger root horizontal line as the boundary (identified by Faster R-CNN algorithm, boundary coordinate y=1100 pixels), the hand image is divided into finger area (distal end) and palm area (proximal end), and the palm area accounts for about 60% of the area;

[0064] 2. Palm center surface feature extraction strategy:

[0065] ①, palm center main line extraction: the third feature line, the second feature line and the third feature line are extracted by hough transform (straight line detection threshold p = 1, theta = 1 degree), according to the spatial distribution of the three main lines, the palm area is macroscopically divided into "thenar eminence area", "palm center area", "hypothenar area" and so on, among which the "palm center area" where laogong acupoint is located outputs the boundary coordinates: left (1200, 900), right (1800, 1000), up (1100, 800), down (1900, 1100).

[0066] ②, analysis of the texture direction: the palm muscle texture analysis is carried out on the "palm center area", the texture direction is extracted by hough transform (mainly longitudinal and radial), and the number of texture intersection points per square millimeter is about 8-12, which provides the basis for subsequent fine segmentation.

[0067] Global reference establishment: 1, coordinate system definition: selecting the midpoint O (1500, 2000) of the wrist transverse line as the global origin (0, 0), defining the horizontal coordinate axis (X axis: pointing to the thumb direction, horizontal to the right) and the vertical coordinate axis (Y axis: pointing to the fingertip direction, vertical upward), and the unit conversion is millimeter (1 pixel = 0.04mm);

[0068] 2, fine area segmentation: combining the texture direction of the palm center area (mainly longitudinal) and the texture intersection point density per square millimeter (10 / mm²), the "palm center area" is divided into laogong acupoint candidate fine area, and the boundary coordinates (global coordinate system) are: left up (1300, 920), right up (1700, 980), left down (1350, 1050), right down (1650, 1100), and the boundary coordinates are stored in the mapping library.

[0069] Local adaptation: 1, local coordinate system establishment, calculate the geometric center of the fine area (local origin P):

[0070] Horizontal coordinate: (1300+1700+1350+1650) / 4 =1500;

[0071] Vertical coordinate: (920+980+1050+1100) / 4 =1012.5;

[0072] Global coordinates of local origin P: (1500, 1012.5);

[0073] 2, offset calculation: the offset of local origin P and global origin O is pixel (0mm), pixel (-39.5mm), the mapping of local positioning result and global space is realized by coordinate conversion formula (global coordinates = local coordinates + offset).

[0074] Candidate point screening: 1, initial conditions: extract the spatial relationship parameters of the fine region and the second feature line (distance 3-5mm perpendicular to the second feature line, located below the second feature line);

[0075] 2, multi-feature fusion screening:

[0076] ①, cross feature of lines: identify 6 intersection points of palmar muscle lines in the region, calculate the Euclidean distance with the region boundary (retain 4 candidate points within 2-4mm range);

[0077] ②, gray gradient feature: use Sobel operator to extract 3x3 pixel gray gradient around the candidate point, retain 3 candidate points with ring gradient feature (center gray value is lower than the surrounding 10-15);

[0078] ③, association verification of acupoint group: combined with the spatial topological relationship of Laogong acupoint and adjacent Daling acupoint and Zhongchong acupoint (distance 8-10mm from Daling acupoint, 60° distribution with Zhongchong acupoint), eliminate 1 deviation point.

[0079] 3, depth verification: collect the depth information of the candidate point through 3D depth camera, Laogong acupoint is a concave acupoint, select the candidate point with "center depth value lower than the average depth value of the surrounding 0.2-0.4mm and ring distribution of depth mutation point", finally retain 1 optimal candidate point.

[0080] 4, result output: the global coordinates of the optimal candidate point are (1520, 1030), converted to physical coordinates (0.8mm, -38.8mm), positioning error ±0.2mm, consistent with the traditional Chinese medicine positioning standard of "between the 2nd and 3rd metacarpal bones in the palm".

[0081] Global coordinate system: take the midpoint O of the wrist horizontal line as the origin, X axis to the right (thumb direction), Y axis upward (finger tip direction), feature marks: wrist horizontal line, finger root horizontal line (divide finger area / palm area), first feature line, second feature line and third feature line, fine region: rectangular region between 2nd and 3rd metacarpal bones in the center of palm (Laogong acupoint candidate area), candidate point and target point: mark 4 initial candidate points and the final Laogong acupoint, and mark the ring gray gradient range.

[0082] In addition, such as Figure 4As shown, an automatic human meridian and acupoint detection system based on AI machine vision is designed, which includes a hand direction judgment module, a hand feature extraction module, a global reference establishment module, a local adaptation module and a candidate point screening module. Data interaction between each module is realized through hardware interface and software protocol to ensure the stability and accuracy of system operation. In the actual application scene, the user only needs to place the hand in the designated position, and the system can automatically complete the whole process from image acquisition to acupoint confirmation, ensuring the full automation from image acquisition to acupoint confirmation. The user can realize efficient and accurate acupoint detection without manual intervention. Through multi-level processing and multi-feature fusion, the system significantly improves the automation degree and positioning accuracy of acupoint detection, meeting the needs of efficient and accurate acupoint detection for traditional Chinese medicine diagnosis and treatment.

[0083] The above is only a preferred embodiment of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments. Any technical solutions falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that, for ordinary skilled persons in the art, some improvements and refinements without departing from the principles of the present application shall also be considered as falling within the protection scope of the present application.

Claims

1. An automatic detection method for acupuncture points on the human hand based on AI machine vision, characterized by: The steps include: A hand face direction determination step is performed to determine whether a continuous distribution curve can be extracted from the user's hand image captured by the visual camera, and output the palm face or the back of the hand according to the extraction result; a hand surface feature extraction step, dividing the hand image into a finger area and a palm area with the finger root crease as the boundary, and selecting a corresponding feature extraction strategy according to the surface of the hand image to extract characteristic landmarks in the hand image; A global reference establishment step, selecting the midpoint of the wrist crease as the global origin, defining the horizontal and vertical coordinate axes, segmenting the hand image according to the characteristic marks to obtain a refined region, extracting the boundary coordinates of the refined region in the global coordinate system and forming a mapping library; In the local adaptation step, a local coordinate system is established for each refined area with its geometric center as the local origin. The offset between the local origin and the global origin is calculated and a coordinate transformation is performed to enable the local positioning result to be directly mapped to the global space. In the candidate point screening step, the spatial relationship parameters between the refined area and the feature marks are extracted as the initial conditions for subsequent candidate point screening. The proportionate candidate points are screened out from the refined area in combination with the texture intersection features, grayscale gradient features and acupoint group correlation verification. The optimal candidate point is retained as the target point by performing concave-convex feature matching verification on the candidate points.

2. The method for automatically detecting acupuncture points on the human hand based on AI machine vision according to claim 1, characterized in that: When the hand image is of the palm surface, the feature extraction strategy includes using an edge detection algorithm to extract the palm main line, which reflects three feature lines. The palm area is divided into several macro areas according to the spatial distribution of the palm main line, and the boundary coordinates are output. The macro areas are then subjected to palm muscle texture analysis, and the texture direction is extracted through Hough transform. The texture direction includes longitudinal, transverse and radial directions. The preset palm acupoints are mapped one by one to each macro area to obtain acupoint location information.

3. The method for automatically detecting acupuncture points on the human hand based on AI machine vision according to claim 2, characterized in that: When the hand image is of the back of the hand, the feature extraction strategy includes using a contour detection algorithm to identify the metacarpal bone vertices and the dorsal wrist creases, analyzing the local gradient direction and intensity changes of the tendon through a texture tracking algorithm to extract the texture direction of the extensor tendon, and mapping the preset dorsal hand acupoints one by one to each macro region to obtain acupoint location information.

4. The method for automatically detecting acupuncture points on the human hand based on AI machine vision according to claim 3 is characterized by: The global benchmark establishment step includes a refined region segmentation strategy, and the refined region segmentation strategy includes a palm face segmentation step and a back hand segmentation step. The palm surface segmentation step further divides the macro region into fine regions by combining the texture arrangement and the number of texture intersections per square millimeter; The step of segmenting the back of the hand divides the metacarpal bone apex and the dorsal wrist crease into fine regions by combining tendon directions and bony landmarks.

5. The method for automatically detecting acupuncture points on the human hand based on AI machine vision according to claim 4, characterized in that: The global benchmark establishment step includes a benchmark establishment strategy, which includes a coordinate axis construction step and a boundary coordinate determination step. In the coordinate axis construction step, the midpoint of the wrist transverse crease pointing toward the thumb is defined as the horizontal coordinate axis, and the midpoint of the wrist transverse crease pointing toward the fingertips is defined as the vertical coordinate axis; The boundary coordinate determination step is to identify the contour edges of each refined area through an edge detection algorithm, and obtain the coordinate values ​​of at least four vertices as boundary coordinates.

6. The method for automatically detecting acupuncture points on the human hand based on AI machine vision according to claim 5, characterized in that: In the local adaptation step, the average of the horizontal coordinates of all boundary vertices in the refined area is used as the horizontal coordinate of the local origin, and the average of the vertical coordinates of all boundary vertices in the refined area is used as the vertical coordinate of the local origin. The offset includes a horizontal offset and a vertical offset.

7. The method for automatically detecting acupuncture points on the human hand based on AI machine vision according to any one of claims 1 to 6, characterized in that: The spatial relationship parameters include distance parameters and angle parameters. The candidate point screening step also includes a multi-feature fusion screening strategy, which includes a texture cross-matching step and a grayscale gradient analysis step. The pattern cross-matching step identifies the intersection points, turning points and dense texture clusters of the palm print texture in the refined area, calculates the Euclidean distance between the candidate point and the area boundary and the angle with the feature line, and selects the top 5 candidate points that meet the preset ratio. The grayscale gradient analysis step extracts gradient changes around candidate points in the hand image and retains candidate points with annular gradient features.

8. The method for automatically detecting acupuncture points on the human hand based on AI machine vision according to claim 7, characterized in that: The multi-feature fusion screening strategy also includes an acupoint group correlation verification step; The acupoint group correlation verification step verifies the spatial topological relationship of the candidate points based on the distribution position correlation between the acupoints in the palm area, and eliminates the candidate points whose relative position deviation from the adjacent acupoints exceeds a preset threshold.

9. The method for automatically detecting acupuncture points on the human hand based on AI machine vision according to claim 8, characterized in that: The multi-feature fusion screening strategy also includes a depth verification step; The depth verification step samples the depth values ​​of the screened candidate points, extracts their three-dimensional coordinates and analyzes the depth value distribution, and retains the optimal candidate points that meet the conditions through concave-convex feature recognition and feature matching verification.

10. The automatic detection system for human hand meridian acupoints based on AI machine vision is characterized by: include: The hand direction determination module determines whether a continuous distribution curve can be extracted from the user's hand image captured by the visual camera, and outputs the palm surface or back of the hand based on the extraction result; A hand face feature extraction module, which divides the hand image into a finger region and a palm region based on the finger root creases, and selects a corresponding feature extraction strategy according to the face of the hand image to extract characteristic landmarks in the hand image; A global reference establishment module selects the midpoint of the wrist crease as the global origin, defines the horizontal and vertical coordinate axes, segments the hand image according to the characteristic landmarks to obtain a refined region, extracts the boundary coordinates of the refined region in the global coordinate system and forms a mapping library; The local adaptation module establishes a local coordinate system for each refined area with its geometric center as the local origin, calculates the offset between the local origin and the global origin, and forms a coordinate transformation so that the local positioning results can be directly mapped to the global space; The candidate point screening module extracts the spatial relationship parameters between the refined area and the feature marks as the initial conditions for subsequent candidate point screening. It also combines the texture intersection features, grayscale gradient features and acupoint group correlation verification to screen out proportional candidate points from the refined area. The optimal candidate point is retained as the target point by performing concave-convex feature matching verification on the candidate points.

Images