Acupuncture positioning system based on computer vision

Abstract

The invention relates to the technical field of medical informatics, in particular to an acupuncture positioning system based on computer vision. The system comprises a memory and a processor, and the processor executes a computer program stored in the memory to realize the following steps: acquiring a depth image of the back of a patient, performing key point detection on the depth image, correcting the coordinates of the spine and the acupuncture points on the two sides of the patient according to the position difference between the back vertebrae of the patient and the standard back vertebrae of the human body; determining a posture difference factor by combining the position difference between the position of each key point in the depth image and the position of the key point in the depth image of the standard human body back and the influence weight of each key point on each acupuncture point; obtaining an acupoint correction coefficient according to the change probability and the posture difference factor of the acupoint positions in different types of crowds; and correcting coordinates of acupuncture points at other positions except the spine and the two sides of the back of the patient by using the acupuncture point correction coefficient. According to the invention, the accuracy of acupoint positioning is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical informatics, and in particular to an acupuncture positioning system based on computer vision. Background Art

[0002] Acupuncture is a traditional treatment method that uses fine needles to penetrate specific meridian points on the skin to regulate the flow of qi and blood in the body and achieve the purpose of preventing and curing diseases. Effective acupuncture treatment requires accurate positioning of acupoints. However, in addition to acupoints that can be directly identified and located by traditional acupoint-finding methods, the visual features of most acupoints are not obvious and cannot be directly identified and located. Therefore, acupuncture treatment can only rely on the experience and skills of the physician to identify and locate the corresponding feature points of acupoints through observation, touch, pressing, etc.

[0003] Computer vision is a technology that enables computers to understand and interpret images and videos. It identifies and locates objects, features, and patterns through methods such as image processing, machine learning, and deep learning. In recent years, the application of computer vision in the medical field has gradually increased, especially in image analysis, surgical navigation, and telemedicine. Mathematical recognition and positioning based on computer vision can achieve more accurate acupoint positioning and reduce the risk of mispuncture and missed puncture. Since it is difficult to achieve touch, press, and other operations by machine means, this may lead to certain errors in acupoint positioning. In addition, the differences in body shape, body shape, and gender of patients make the body structure of different patients different, resulting in deviations in the location of acupoints. Therefore, the use of standardized models is difficult to apply to all patients, which may lead to inaccurate acupoint positioning for individual patients and affect the treatment effect. Summary of the invention

[0004] In order to solve the problem of inaccurate positioning of acupoints of patients in the existing methods, the purpose of the present invention is to provide an acupuncture positioning system based on computer vision. The technical solutions adopted are as follows: The present invention provides an acupuncture positioning system based on computer vision, comprising a memory and a processor, wherein the processor executes a computer program stored in the memory to implement the following steps: Acquire a depth image of the patient's back, perform key point detection on the depth image to obtain each key point, and obtain the coordinates of each acupuncture point on the patient's back; Correcting the coordinates of the patient's spine and the acupuncture points on both sides according to the position difference between the patient's back vertebrae and the standard human back vertebrae to obtain the corresponding corrected coordinates; determining the posture difference factor by combining the position difference between each key point in the depth image and the key point in the depth image of the standard human back and the influence weight of each key point on each acupuncture point, wherein the influence weight is determined based on the relative position of the key point and the acupuncture point; According to the probability of change of acupoint positions in different categories of people and the posture difference factor, an acupoint correction coefficient is obtained; the coordinates of acupoints at other positions on the patient's back except the spine and both sides are corrected using the acupoint correction coefficient to obtain corresponding corrected coordinates.

[0005] Preferably, the coordinates of the patient's spine and the acupuncture points on both sides are corrected according to the position difference between the patient's back vertebrae and the standard human back vertebrae to obtain the corresponding corrected coordinates, including: For any vertebra on the patient's back: the difference between the coordinates of the center point of the vertebra and the coordinates of the center point of the same vertebra on the back of a standard human body is used as the offset vector of the vertebra; The degree of spinal deviation is determined as the average value of the modulus lengths of the deviation vectors of all vertebrae in the patient's back; The coordinates of each first acupuncture point are corrected respectively using the spinal column deviation degree to obtain the corrected coordinates corresponding to each first acupuncture point; the first acupuncture points are the acupuncture points on the spinal column and both sides.

[0006] Preferably, the method of correcting the coordinates of each first acupuncture point respectively by using the spinal column deviation degree to obtain the corrected coordinates corresponding to each first acupuncture point includes: For any first acupuncture point: calculate the sum of the constant 1 and the degree of spinal column deviation, and take the product of the sum and the coordinates of the first acupuncture point as the corrected coordinates corresponding to the first acupuncture point.

[0007] Preferably, the acquisition of the influence weight of the key point on each acupoint includes: Get the distance between the candidate key point and each acupoint in the area where it is located, and calculate the cumulative sum of the distances between the candidate key point and all acupoints in the area where it is located; The ratio between the distance between the candidate key point and each acupuncture point in the region where it is located and the accumulated sum is used as the influence weight of the candidate key point on the corresponding acupuncture point in the region where it is located; The candidate key point is any key point in the depth image of the patient's back.

[0008] Preferably, the determining of the posture difference factor by combining the position difference between each key point in the depth image and the key point in the depth image of the standard human back and the influence weight of each key point on each acupoint comprises: Calculate the posture difference factor according to the feature vector of the key point in the depth image, the feature vector of the key point corresponding to the key point in the depth image of the standard human back, and the influence weight of the key point on each acupoint; The direction of the feature vector of the key point is the direction from the preset position to the key point, and the size of the feature vector is the distance between the preset position and the key point.

[0009] Preferably, the calculating of the posture difference factor according to the feature vector of the key point in the depth image, the feature vector of the key point corresponding to the key point in the depth image of the standard human back, and the influence weight of the key point on each acupoint includes: The modulus length of the difference between the feature vector of the candidate key point and the feature vector of the key point corresponding to the candidate key point in the depth image of the standard human back is determined as the posture difference corresponding to the candidate key point; Calculate the first product of the influence weight of the candidate key point on each acupuncture point in the area where the candidate key point is located and the posture difference, and determine the sum of all the first products corresponding to the candidate key point as the first eigenvalue of the candidate key point; The average value of the first eigenvalues ​​of all key points in each region is determined as the posture difference factor of the corresponding region.

[0010] Preferably, obtaining the acupoint correction coefficient according to the change probability of the acupoint position in different categories of people and the posture difference factor includes: The ratio between the probability of change of acupoint position in the population of the category to which the patient belongs and the probability of change of acupoint position in the population of all categories is determined as the individual difference factor; The acupoint correction coefficient of each area is determined by combining the individual difference factor and the posture difference factor of each area.

[0011] Preferably, the combining of the individual difference factor and the posture difference factor of each region to determine the acupoint correction coefficient of each region includes: taking the product of the individual difference factor and the posture difference factor of each region as the acupoint correction coefficient of the corresponding region.

[0012] Preferably, the use of the acupoint correction coefficient to correct the coordinates of acupoints at other positions on the patient's back except the spine and both sides to obtain corresponding corrected coordinates includes: For any area except the spine and the areas where the acupuncture points are located on both sides: The acupoint correction coefficient of the region is standardized, and the difference between the standardized acupoint correction coefficient and the constant 1 is used as the comprehensive correction coefficient, wherein the value range of the standardized acupoint correction coefficient is [0, 2]; The coordinates of each acupuncture point in the area are corrected using the comprehensive correction coefficient to obtain the corrected coordinates.

[0013] Preferably, the method of correcting the coordinates of each acupuncture point in the region by using the comprehensive correction coefficient to obtain the corrected coordinates includes: The product of the comprehensive correction coefficient and the coordinates of each acupoint in the area is used as the corrected coordinates of the corresponding acupoint.

[0014] The present invention has at least the following beneficial effects: The present invention first extracts the key points in the depth image of the patient's back and the coordinates of each acupuncture point on the patient's back. Since the acupuncture points on the back of different patients are different and the positions are not exactly the same, there are certain differences between the collected positions and the actual positions. In order to ensure that the acupuncture points on the back can be found quickly and accurately during subsequent acupuncture, the positions of the acupuncture points need to be corrected. Since when the spine changes due to poor posture, exercise or injury, the muscles, ligaments and nerves on both sides of the spine will also be affected, and the spine and the acupuncture points on both sides will change with the change of the spine. Therefore, according to the position difference between the patient's back vertebrae and the standard human back vertebrae, the coordinates of the patient's spine and the acupuncture points on both sides are corrected. Combined with the position difference between the position of each key point in the depth image of the patient's back and the key points in the depth image of the standard human back, the influence weight of each key point on each acupuncture point, and the change probability of the acupuncture point position in different categories of people, the positions of the acupuncture points on the patient's back other than the spine and both sides are corrected. The corrected acupuncture point positions are more accurate. The method provided by the present invention can avoid the problem of deviation in the positioning of the acupuncture points during the subsequent back acupuncture of the patient, and realizes the accurate positioning of the acupuncture points on the patient's back. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] 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.

[0016] Figure 1 The present invention provides a flowchart of a method performed by an acupuncture positioning system based on computer vision according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following is a detailed description of an acupuncture positioning system based on computer vision proposed in accordance with the present invention in conjunction with the accompanying drawings and preferred embodiments.

[0018] 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.

[0019] The specific scheme of the acupuncture positioning system based on computer vision provided by the present invention is described in detail below with reference to the accompanying drawings.

[0020] An embodiment of an acupuncture positioning system based on computer vision: The specific scenario targeted by this embodiment is: when performing acupuncture on the patient's back, in order to overcome the problem of "one prescription for one thousand people" in traditional acupuncture treatment and improve the personalization and accuracy of treatment, it is necessary to consider the posture changes and physical differences of each individual, construct acupoint correction coefficients, correct the standard acupoints according to individual differences, and then obtain the accurate acupoint coordinates of the individual, so as to achieve the purpose of identifying and accurately locating the acupoints.

[0021] This embodiment proposes an acupuncture positioning system based on computer vision, which can achieve the following Figure 1 The specific steps are as follows: Step S1, obtaining a depth image of the patient's back, performing key point detection on the depth image to obtain each key point, and obtaining the coordinates of each acupuncture point on the patient's back.

[0022] First, a depth camera is used to capture the patient's posture and movement to obtain a clear depth image of the patient's back. Based on the depth image, the three-dimensional structural information of the patient's back can be reflected, including depth data of back color, back curve, shoulder width and other body features.

[0023] Then, the deep learning model HRNET is used to perform key point detection on the depth image of the patient's back and extract all key points. The key point extraction process is an existing technology and will not be described in detail here.

[0024] In the study of acupuncture identification and positioning based on computer vision, the establishment of a unified standard coordinate system can enable quantitative analysis of different acupoints, facilitate comparison and analysis of different patient data, and help improve the accuracy and effectiveness of subsequent analysis results.

[0025] In this embodiment, a three-dimensional coordinate system is established with the X-axis representing the head and feet direction of the body, the Y-axis representing the left and right direction of the body, and the Z-axis representing the up and down direction of the body. The implementer can select the position of the origin according to the specific situation.

[0026] Combined with the knowledge of human anatomy, all the acupoints on the back are classified. According to the relative positions of the acupoints on the back, all the acupoints on the back are mainly divided into four categories, namely: acupoints on the spine and both sides, acupoints in the shoulder area, acupoints in the waist area, and acupoints related to meridians. Combined with the results of acupoint division, the depth image of the back is divided into four areas, namely, the spine and both sides, the shoulder area, the waist area, and the area related to meridians. It should be noted that the acupoints on the spine and both sides are located in the spine and both sides, the acupoints in the shoulder area are located in the shoulder area, the acupoints in the waist area are located in the waist area, and the acupoints related to meridians are located in the area related to meridians. The division of acupoints and regions are divided according to the knowledge of human anatomy, and will not be elaborated here.

[0027] According to the coordinate system constructed above and the depth image of the patient's back, the coordinates of each acupuncture point on the patient's back are preliminarily obtained. The coordinates are the initial coordinates, which may be different from the actual coordinate position. In order to ensure the subsequent acupuncture effect, the initial coordinates need to be corrected to achieve accurate positioning of the acupuncture points.

[0028] Step S2, correcting the coordinates of the patient's spine and the acupuncture points on both sides according to the position difference between the patient's back vertebrae and the standard human back vertebrae to obtain the corresponding corrected coordinates; combining the position difference between the key points in the depth image and the key points in the depth image of the standard human back and the influence weight of each key point on each acupuncture point to determine the posture difference factor, wherein the influence weight is determined based on the relative position of the key point and the acupuncture point.

[0029] In the image of the human back, the spine is a distinctive area. The degree of appearance of the spine in the image will vary for people of different body shapes and muscle distributions; when the spine changes due to poor posture, exercise or injury, the muscles, ligaments and nerves on both sides of the spine will also be affected; therefore, the spine and the acupuncture points on both sides will change with the changes in the spine. By calculating the difference between the patient's spine and the standard spine, the accurate three-dimensional coordinates of the patient's spine and the acupuncture points on both sides can be determined.

[0030] For any vertebra on the patient's back: the difference between the coordinates of the center point of the vertebra and the coordinates of the center point of the same vertebra corresponding to the standard human back is used as the offset vector of the vertebra, and the modulus of the offset vector of the vertebra is obtained. Using this method, the modulus of the offset vector of each vertebra on the patient's back can be obtained, and the average value of the modulus of the offset vectors of all vertebrae on the patient's back is determined as the degree of spinal deviation.

[0031] The spine and the acupoints on both sides are recorded as first acupoints. Next, the coordinates of each first acupoint are corrected using the degree of spinal deviation to obtain the corrected coordinates corresponding to each first acupoint. Specifically, for any first acupoint: calculate the sum of the constant 1 and the degree of spinal deviation, and take the product of the sum and the coordinates of the first acupoint as the corrected coordinates corresponding to the first acupoint.

[0032] In this embodiment, the corrected coordinates of the acupuncture points on the spine and both sides are given. The corrected coordinates of the kth acupuncture point on the spine and both sides can be expressed as: in, represents the corrected coordinates of the kth acupuncture point on the spine and both sides, represents the corrected horizontal coordinate of the kth acupoint on the spine and both sides, represents the corrected ordinate of the kth acupuncture point on the spine and both sides, represents the corrected vertical coordinate of the kth acupoint on the spine and both sides, Indicates the degree of spinal deviation. represents the coordinates of the kth acupuncture point on the spine and both sides, represents the horizontal coordinate of the kth acupuncture point on the spine and both sides, represents the ordinate of the kth acupuncture point on the spine and both sides, Represents the vertical coordinate of the kth acupuncture point on the spine and both sides.

[0033] By adopting the above method, the position of each acupuncture point on the spine and both sides can be adjusted to obtain the corrected coordinates.

[0034] Acupoints in the shoulder area, the waist area, and the meridian-related acupoints are those with fuzzy visual features. Due to individual differences, computer vision cannot directly and accurately identify and locate them. By constructing acupoint correction coefficients based on individual differences, the positions of acupoints with fuzzy visual features can be adjusted to more accurately locate and stimulate the acupoints.

[0035] Next, this embodiment is described by taking any key point on the patient's back as an example, and other key points can be processed by the method provided in this embodiment.

[0036] Specifically, any key point in the depth image of the patient's back is recorded as a candidate key point; the distance between the candidate key point and each acupoint in the area where it is located is obtained, and the cumulative sum of the distances between the candidate key point and all acupoints in the area where it is located is calculated; the ratio between the distance between the candidate key point and each acupoint in the area where it is located and the cumulative sum is used as the influence weight of the candidate key point on the corresponding acupoint in the area where it is located.

[0037] In this embodiment, a specific calculation formula for the influence weight of a key point on the acupuncture points in the area where it is located is given, and the influence weight of the i-th key point on the j-th acupuncture point in the area where it is located can be expressed as: in, represents the influence weight of the i-th key point on the j-th acupuncture point in its area, represents the distance between the i-th key point and the j-th acupuncture point in its area, and J represents the number of acupuncture points in the area where the i-th key point is located.

[0038] By using the above method, the influence weight of each key point on each acupuncture point in its area can be obtained.

[0039] The feature vector of each key point is obtained, the direction of the feature vector of the key point is the direction from the preset position to the key point, and the size of the feature vector is the distance between the preset position and the key point. In this embodiment, the preset position is the origin of the coordinates, and in specific applications, the implementer can set it according to specific circumstances.

[0040] The following is still described by taking candidate key points as an example, and other key points can be processed by the method provided in this embodiment.

[0041] Specifically, the modulus of the difference between the feature vector of the candidate key point and the feature vector of the key point corresponding to the candidate key point in the depth image of the standard human back is determined as the posture difference corresponding to the candidate key point. The first product of the influence weight of the candidate key point on each acupuncture point in the area where it is located and the posture difference is calculated respectively, and the sum of all the first products corresponding to the candidate key point is determined as the first eigenvalue of the candidate key point. It should be noted that: when determining the feature vector of the key point in the depth image of the standard human back, the average value of the feature vectors of the key point corresponding to all reference persons in the same category as the patient is used as the feature vector of the key point in the depth image of the standard human back.

[0042] The average value of the first eigenvalues ​​of all key points in each region is determined as the posture difference factor of the corresponding region.

[0043] In this embodiment, a specific calculation formula of the posture difference factor is given. The posture difference factor of any area except the area where the spine and the acupuncture points on both sides are located can be expressed as: Among them, f represents the posture difference factor of the area, is the posture difference corresponding to the sth key point in the area, It represents the influence weight of the sth key point in the area on the rth acupuncture point, represents the number of acupuncture points in the area, and S represents the number of key points in the area.

[0044] It is used to characterize the first product of the influence weight of the sth key point in the region on each acupuncture point in the region and the posture difference, that is, the first product corresponding to the sth key point, Characterizes the first eigenvalue of the sth key point in the region. When the influence weight of the key point in the region on the acupoint is greater and the posture difference corresponding to the key point is greater, the posture difference factor of the region is greater.

[0045] By adopting the above method, the posture difference factors of the shoulder area, the waist area and the area related to the meridians can be obtained.

[0046] Step S3, obtaining an acupoint correction coefficient according to the change probability of the acupoint position in different categories of people and the posture difference factor; using the acupoint correction coefficient to correct the coordinates of the acupoints at other positions on the patient's back except the spine and both sides, to obtain corresponding corrected coordinates.

[0047] There are certain deviations in the acupoints on the back of people of different genders, different age groups, and different physical parameters. There are certain differences in body structure and physiological characteristics between men and women. The physiological changes of different age groups (such as children, teenagers, middle-aged people, and the elderly) are different, and the body fat percentage reflects the physical health status and body shape of different patients. Therefore, the reference personnel are classified into different groups based on gender, age, and body fat percentage, and then the acupoint correction coefficient is constructed according to the influence of different groups on the changes of acupoints and the acupoint location characteristics of the group to which the patient belongs, and then the coordinates of the acupoints at other positions on the patient's back except the spine and both sides are corrected.

[0048] The present invention is divided into male and female according to gender, divided into children (0-12 years old), teenagers (13-19 years old), adults (20-60 years old), and the elderly (over 60 years old) according to age; divided into thin (BMI18.5-24.9), normal weight (BMI18.5-24.9), overweight (BMI25-29.9), and obese (BMI≥30) according to body fat rate; according to this classification standard, all reference persons are divided into multiple different categories. In specific applications, implementers can make classifications according to specific circumstances.

[0049] It should be noted that: there are multiple reference persons in each category, and the depth images of the reference persons' backs and the accurate coordinates of all acupuncture points on the reference persons' backs have been obtained in advance.

[0050] Regression analysis is used to calculate the probability of change of acupoint positions in each category of people and the probability of change of acupoint positions in all categories of people. The method for calculating the change probability is existing technology and will not be described in detail here.

[0051] The ratio between the probability of change of acupoint position in the group of people to which the patient belongs and the probability of change of acupoint position in all groups of people is determined as the individual difference factor. The product of the individual difference factor and the posture difference factor of each area is used as the acupoint correction coefficient of the corresponding area. By using this method, the acupoint correction coefficient of the shoulder area, the acupoint correction coefficient of the waist area and the acupoint correction coefficient of the area related to the meridians can be obtained.

[0052] According to the traditional theory of Chinese medicine on the relationship between acupoints and different parts of the human body, the present invention takes into account the impact of individual differences and posture changes on acupoint positioning and effects, and constructs acupoint correction coefficients based on posture difference factors and the probability of change of acupoint positions to adapt to the physiological and posture characteristics of different individuals.

[0053] Specifically, for any area except the spine and the areas where the acupoints on both sides are located: the acupoint correction coefficients of the area are standardized, and the difference between the standardized acupoint correction coefficients and the constant 1 is used as the comprehensive correction coefficient, and the value range of the standardized acupoint correction coefficients is [0, 2]. The standardization of data is a prior art and will not be described in detail here.

[0054] In the reference group, different individuals were selected for acupoint positioning. Acupuncture was performed on the acupoint correction coefficients of different individuals in different postures. The effectiveness of the corrected acupoints was evaluated and verified using physiological indicators (such as galvanic skin response, pain score, etc.). The weights of the influencing factors and the model structure were optimized based on the experimental results.

[0055] The coordinates of each acupoint in the region are corrected using the comprehensive correction coefficient to obtain the corrected coordinates. Specifically, the product of the comprehensive correction coefficient and the coordinates of each acupoint in the region is used as the corrected coordinates of the corresponding acupoint.

[0056] As a specific implementation, the specific calculation formula of the corrected coordinates is given: in, Indicates the corrected coordinates of the tth acupuncture point in the area, It represents the corrected horizontal coordinate of the tth acupoint in the area. It represents the corrected ordinate of the t-th acupuncture point in the area. Indicates the corrected vertical coordinate of the tth acupuncture point in the area, represents the comprehensive correction coefficient of the area, represents the coordinates of the tth acupuncture point in the area, represents the horizontal coordinate of the tth acupuncture point in the area, represents the ordinate of the tth acupuncture point in the area, Indicates the vertical coordinate of the tth acupuncture point in the area.

[0057] By adopting the above method, the coordinates of each acupoint in each area except the area where the acupoints on the spine and both sides are located can be corrected, and the corrected coordinates of all acupoints on the patient's back can be obtained.

[0058] After the coordinates of all acupoints are corrected, when acupuncture is performed on the patient's back, the corrected coordinates can be directly used to accurately locate the acupoints, thereby achieving real-time identification of the acupoints on the back of the human body and improving treatment effects and personalized services.

[0059] This embodiment first extracts the key points in the depth image of the patient's back and the coordinates of each acupuncture point on the patient's back. Since the acupuncture points on the back of different patients are different and the positions are not exactly the same, there are certain differences between the collected positions and the actual positions. In order to ensure that the acupuncture points on the back can be found quickly and accurately during subsequent acupuncture, the positions of the acupuncture points need to be corrected. Since when the spine changes due to poor posture, exercise or injury, the muscles, ligaments and nerves on both sides of the spine will also be affected, and the spine and the acupuncture points on both sides will change with the changes in the spine. Therefore, according to the position difference between the patient's back vertebrae and the standard human back vertebrae, the coordinates of the patient's spine and the acupuncture points on both sides are corrected. Combined with the position difference between the position of each key point in the depth image of the patient's back and the key points in the depth image of the standard human back, the influence weight of each key point on each acupuncture point, and the probability of change of the acupuncture point position in different categories of people, the positions of the acupuncture points on the patient's back other than the spine and both sides are corrected. The corrected acupuncture point positions are more accurate. The method provided in this embodiment can avoid the problem of deviation in the positioning of acupuncture points during subsequent acupuncture on the patient's back, and achieve accurate positioning of the acupuncture points on the patient's back.

[0060] In other embodiments, a device is also provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes the method executed by the acupuncture positioning system based on computer vision. The device can be specifically a chip, a component or a module, and the chip may include a connected processor and a memory; wherein the memory is used to store instructions, and when the processor calls and executes the instructions, the chip can execute the method executed by the acupuncture positioning system based on computer vision provided in the above embodiment.

[0061] In other embodiments, a computer program product is also provided. When the computer program product is run on a computer, the computer is caused to execute the above-mentioned related steps to implement the method performed by the computer vision-based acupuncture positioning system provided in the above-mentioned embodiment.

[0062] In other embodiments, a computer-readable storage medium is also provided, in which a computer program code is stored. When the computer program code is executed on a computer, the computer executes the above-mentioned related method steps to implement the method performed by the computer vision-based acupuncture positioning system provided in the above embodiment.

[0063] Among them, the provided systems, devices, computer program products, and computer-readable storage media are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods provided above and will not be repeated here.

[0064] It should be noted that 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 acupuncture positioning system based on computer vision, comprising a memory and a processor, characterized in that: The processor executes the computer program stored in the memory to implement the following steps: Acquire a depth image of the patient's back, perform key point detection on the depth image to obtain each key point, and obtain the coordinates of each acupuncture point on the patient's back; Correcting the coordinates of the patient's spine and the acupuncture points on both sides according to the position difference between the patient's back vertebrae and the standard human back vertebrae to obtain the corresponding corrected coordinates; determining the posture difference factor by combining the position difference between each key point in the depth image and the key point in the depth image of the standard human back and the influence weight of each key point on each acupuncture point, wherein the influence weight is determined based on the relative position of the key point and the acupuncture point; According to the probability of change of acupoint positions in different categories of people and the posture difference factor, an acupoint correction coefficient is obtained; the coordinates of acupoints at other positions of the patient's back except the spine and both sides are corrected using the acupoint correction coefficient to obtain corresponding corrected coordinates; Obtaining the influence weight of key points on each acupoint, including: Get the distance between the candidate key point and each acupoint in the area where it is located, and calculate the cumulative sum of the distances between the candidate key point and all acupoints in the area where it is located; The ratio between the distance between the candidate key point and each acupuncture point in the region where it is located and the accumulated sum is used as the influence weight of the candidate key point on the corresponding acupuncture point in the region where it is located; The candidate key point is any key point in the depth image of the patient's back; The determining of the posture difference factor by combining the position difference between each key point in the depth image and the key point in the depth image of the standard human back and the influence weight of each key point on each acupoint comprises: Calculate the posture difference factor according to the feature vector of the key point in the depth image, the feature vector of the key point corresponding to the key point in the depth image of the standard human back, and the influence weight of the key point on each acupoint; The direction of the feature vector of the key point is the direction from the preset position to the key point, and the size of the feature vector is the distance between the preset position and the key point.

2. The computer vision-based acupuncture positioning system according to claim 1, characterized in that: The coordinates of the patient's spine and the acupuncture points on both sides are corrected according to the position difference between the patient's back vertebrae and the standard human back vertebrae to obtain the corresponding corrected coordinates, including: For any vertebra on the patient's back: the difference between the coordinates of the center point of the vertebra and the coordinates of the center point of the same vertebra on the back of a standard human body is used as the offset vector of the vertebra; The degree of spinal deviation is determined as the average value of the modulus lengths of the deviation vectors of all vertebrae in the patient's back; The coordinates of each first acupuncture point are corrected respectively using the spinal column deviation degree to obtain the corrected coordinates corresponding to each first acupuncture point; the first acupuncture points are the acupuncture points on the spinal column and both sides.

3. The computer vision-based acupuncture positioning system according to claim 2, characterized in that: The method of respectively correcting the coordinates of each first acupuncture point by using the spinal column deviation degree to obtain the corrected coordinates corresponding to each first acupuncture point includes: For any first acupuncture point: calculate the sum of the constant 1 and the degree of spinal column deviation, and take the product of the sum and the coordinates of the first acupuncture point as the corrected coordinates corresponding to the first acupuncture point.

4. The acupuncture positioning system based on computer vision according to claim 1, characterized in that: The step of calculating the posture difference factor according to the feature vector of the key point in the depth image, the feature vector of the key point corresponding to the key point in the depth image of the standard human back, and the influence weight of the key point on each acupoint includes: The modulus length of the difference between the feature vector of the candidate key point and the feature vector of the key point corresponding to the candidate key point in the depth image of the standard human back is determined as the posture difference corresponding to the candidate key point; Calculate the first product of the influence weight of the candidate key point on each acupuncture point in the area where the candidate key point is located and the posture difference, and determine the sum of all the first products corresponding to the candidate key point as the first eigenvalue of the candidate key point; The average value of the first eigenvalues ​​of all key points in each region is determined as the posture difference factor of the corresponding region.

5. The computer vision-based acupuncture positioning system according to claim 4, characterized in that: The step of obtaining the acupoint correction coefficient according to the probability of change of the acupoint positions in different categories of people and the posture difference factor comprises: The ratio between the probability of change of acupoint position in the population of the category to which the patient belongs and the probability of change of acupoint position in the population of all categories is determined as the individual difference factor; The acupoint correction coefficient of each area is determined by combining the individual difference factor and the posture difference factor of each area.

6. The computer vision-based acupuncture positioning system according to claim 5, characterized in that: The combining of the individual difference factor and the posture difference factor of each region to determine the acupoint correction coefficient of each region includes: taking the product of the individual difference factor and the posture difference factor of each region as the acupoint correction coefficient of the corresponding region.

7. The computer vision-based acupuncture positioning system according to claim 1, characterized in that: The method of using the acupoint correction coefficient to correct the coordinates of acupoints at other positions on the patient's back except the spine and both sides to obtain corresponding corrected coordinates includes: For any area except the spine and the areas where the acupuncture points are located on both sides: The acupoint correction coefficient of the region is standardized, and the difference between the standardized acupoint correction coefficient and the constant 1 is used as the comprehensive correction coefficient, wherein the value range of the standardized acupoint correction coefficient is [0, 2]; The coordinates of each acupuncture point in the area are corrected using the comprehensive correction coefficient to obtain the corrected coordinates.

8. The computer vision-based acupuncture positioning system according to claim 7, characterized in that: The method of using the comprehensive correction coefficient to correct the coordinates of each acupuncture point in the region to obtain the corrected coordinates includes: The product of the comprehensive correction coefficient and the coordinates of each acupoint in the area is used as the corrected coordinates of the corresponding acupoint.

Images