Neck Acupoint Recognition Method, Device, Massage Robot and Digital Human

Abstract

The present invention provides a method and device for identifying neck acupoints, a massage robot and a digital human, belonging to the technical field of acupoint identification. The method includes: determining key points related to horizontal and vertical cun and feature points related to meridians of the neck image to be identified, and determining the unit vertical cun and unit horizontal cun of the neck image to be identified based on the key points, and determining the Conception Vessel and Governor Vessel of the neck image to be identified based on the feature points; the key points include the Yintang acupoint; determining the laryngeal prominence of the neck image to be identified; determining the positional relationships among the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel and the first type of neck acupoints, and determining the first type of neck acupoints based on the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationships, the unit vertical cun and the unit horizontal cun; determining the association relationships between the first type of neck acupoints and the second type of neck acupoints, and determining the second type of neck acupoints based on the association relationships and the first type of neck acupoints. The present invention does not require prior calibration of neck acupoints, improving the accuracy and efficiency of neck acupoint identification.

Description

Technical Field

[0001] The present invention relates to the technical field of acupoint recognition, and specifically relates to a method and device for recognizing neck acupoints, a massage robot, and a digital human. Background Technique

[0002] With the extensive use of a series of electronic products such as mobile phones and tablets by people, the cervical vertebrae remain stationary for a long time, seriously threatening the health of the cervical vertebrae. Traditional Chinese medicine massage can solve 95% of cervical vertebra problems. By performing massage manipulations such as pressing, kneading, dotting, pressing, and rubbing on acupoints, it can effectively eliminate the fatigue of neck muscles, play the role of dredging meridians and promoting qi and blood circulation, thereby improving blood circulation in the head and neck and effectively relieving the discomfort caused by cervical spondylosis. Acupoint massage has the characteristics of being convenient and easy to operate, and is especially suitable for daily health care massage. At present, acupoint positioning is mostly manually performed by professional physicians and requires a large amount of professional training, which is not conducive to non-professionals for daily health care of acupoint massage and is also not conducive to the automation of massage medical devices.

[0003] The existing patent CN111598949A discloses an automatic positioning method for acupoints on the human head and neck, which needs to train a model based on an image set marked with acupoint points to directly obtain neck acupoints through the trained model. That is: in the prior art, a model with the input of an image to be recognized and the output of neck acupoints is trained, and the training process of the model depends on the accurate marking of neck acupoints. It has the following technical problems: it is necessary to obtain a second image set through manual marking to realize neck acupoint recognition, relying on the accuracy of manual marking, with a relatively low recognition accuracy, and moreover, the marking process is complex and time-consuming, resulting in a low recognition efficiency of neck acupoints.

[0004] Therefore, there is an urgent need to provide a method and device for recognizing neck acupoints, a massage robot, and a digital human to improve the recognition accuracy and efficiency of neck acupoints. Summary of the Invention

[0005] In view of this, it is necessary to provide a method and device for recognizing neck acupoints, a massage robot, and a digital human to solve the technical problem in the prior art that depends on manual annotation, resulting in relatively low recognition accuracy and recognition efficiency of neck acupoints.

[0006] In a first aspect, the present invention provides a method for recognizing neck acupoints, including:

[0007] Determine the key points related to horizontal and vertical cun and the characteristic points related to meridians of the neck image to be recognized, and determine the unit vertical cun and unit horizontal cun of the neck image to be recognized based on the key points, and determine the Conception Vessel and Governor Vessel of the neck image to be recognized based on the characteristic points; the key points include the Yintang acupoint;

[0008] Determine the laryngeal prominence of the neck image to be recognized;

[0009] Determine the positional relationship among the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel, and the first type of neck acupoints, and determine the first type of neck acupoints based on the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationship, the unit vertical cun, and the unit horizontal cun.

[0010] Determine the association relationship between the first type of neck acupoints and the second type of neck acupoints, and determine the second type of neck acupoints based on the association relationship and the first type of neck acupoints.

[0011] In some possible implementation manners, the key points further include the left anterior hairline corner point and the right anterior hairline corner point related to the horizontal cun, and the anterior hairline related to the vertical cun; the feature points include the left earlobe point, the right earlobe point, the left shoulder peak point, and the right shoulder peak point; then the determination of the key points related to the horizontal and vertical cuns and the feature points related to the meridians of the neck image to be recognized includes:

[0012] Detect the neck image to be recognized to obtain a plurality of candidate key points, and determine the left anterior hairline corner point, the right anterior hairline corner point, the first target point and the second target point related to the Yintang acupoint, and the left earlobe point, the right earlobe point, the left shoulder peak point, and the right shoulder peak point among the plurality of candidate key points based on the mapping relationship between the candidate key point numbers and the candidate key points.

[0013] Determine the midpoint of the first target point and the second target point as the Yintang acupoint.

[0014] Determine the face rectangular frame of the neck image to be recognized, and use the frame line near the top of the head in the face rectangular frame as the anterior hairline.

[0015] In some possible implementation manners, determine the unit vertical cun and the unit horizontal cun of the neck image to be recognized based on the key points, and determine the Conception Vessel and the Governor Vessel of the neck image to be recognized based on the feature points, including:

[0016] Determine the longitudinal distance and longitudinal bone cun between the Yintang acupoint and the anterior hairline, and the transverse distance and transverse bone cun between the left anterior hairline corner point and the right anterior hairline corner point.

[0017] Determine the unit vertical cun based on the longitudinal distance and the longitudinal bone cun, and determine the unit horizontal cun based on the transverse distance and the transverse bone cun.

[0018] Determine the first connection line and the first midpoint of the left earlobe point and the right earlobe point, and take the first straight line on the front of the human body that is perpendicular to the first connection line and passes through the first midpoint as the Conception Vessel, and determine the second connection line and the second midpoint of the left acromion point and the right acromion point, and take the second straight line on the back of the human body that is perpendicular to the second connection line and passes through the second midpoint as the Governor Vessel.

[0019] In some possible implementation manners, determining the Adam's apple of the neck image to be recognized includes:

[0020] Determine a plurality of candidate reference points and a plurality of candidate reference point numbers of the neck image to be recognized;

[0021] Determine the Adam's apple based on the Adam's apple number of the Adam's apple and the pre-constructed association relationship between the candidate reference point numbers and the candidate reference point positions.

[0022] In some possible implementation manners, the positional relationship is the osteopathic positional relationship; then determining the first type of neck acupoints based on the Adam's apple, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationship, the unit vertical cun, and the unit horizontal cun includes:

[0023] Determine the osteopathic positions of the first type of neck acupoints based on the Adam's apple, the Yintang acupoint, the Conception Vessel, the Governor Vessel, and the positional relationship;

[0024] Convert the osteopathic positions into image positions based on the unit vertical cun and the unit horizontal cun, and take the image positions as the positions where the first type of neck acupoints are located.

[0025] In some possible implementation manners, the first type of neck acupoints include the Renying acupoint, the Futu acupoint, the Tiantu acupoint, the Tianchuang acupoint located in the front neck, and the Yamen acupoint and the Dazhui acupoint located in the back neck; the positional relationships include:

[0026] The Renying acupoint is 1.5 horizontal cuns away from the Adam's apple;

[0027] The Futu acupoint is 3 horizontal cuns away from the Adam's apple;

[0028] The Tiantu acupoint is located on the Conception Vessel, below the Adam's apple and 2 vertical cuns away from the Adam's apple;

[0029] The Tianchuang acupoint is 3.5 horizontal cuns away from the Adam's apple;

[0030] The Yamen acupoint is located on the Governor Vessel, 14.5 vertical cuns away from the Yintang acupoint;

[0031] The Dazhui acupoint is located on the Governor Vessel, 18 vertical cuns away from the Yintang acupoint.

[0032] In some possible implementation manners, the second type of neck acupoints include Lianquan acupoint, Qishe acupoint, Shuitu acupoint, Quepen acupoint located in the front neck, and Fengfu acupoint, Fengchi acupoint, Jingbailao acupoint, Tianzhu acupoint located in the back neck; the association relationship is as follows:

[0033] The Lianquan acupoint is located at the midpoint of the line connecting the chin and the laryngeal prominence; the chin is the lower frame line of the rectangular face frame of the human face;

[0034] The Qishe acupoint is 1.5 transverse cun away from the Tiantu acupoint;

[0035] The Shuitu acupoint is located on the Conception Vessel and at the midpoint of the line connecting the Qishe acupoint and the Renying acupoint;

[0036] The Quepen acupoint is the intersection point of the extension line of the line connecting the Tiantu acupoint and the Qishe acupoint and the target line segment, and the target line segment is a line segment perpendicular to the line connecting the left nipple and the right nipple and passing through the left nipple or the right nipple;

[0037] The Tianding acupoint is located at the midpoint of the line connecting the Futu acupoint and the Quepen acupoint;

[0038] The Fengfu acupoint is located above the Yamen acupoint and is 0.5 vertical cun away from the Yamen acupoint;

[0039] The Fengchi acupoint is 1.3 transverse cun away from the Fengfu acupoint;

[0040] The Jingbailao acupoint is located above the Dazhui acupoint and is 2 vertical cun and 1 transverse cun away from the Dazhui acupoint;

[0041] The Tianzhu acupoint is located below the Yamen acupoint and is 0.3 vertical cun and 1.3 transverse cun away from the Yamen acupoint.

[0042] On the other hand, the present invention also provides a neck acupoint recognition device, including:

[0043] A reference information determination unit, configured to determine key points related to horizontal and vertical cun and feature points related to meridians of the neck image to be recognized, and determine the unit vertical cun and unit horizontal cun of the neck image to be recognized based on the key points, and determine the Conception Vessel and Governor Vessel of the neck image to be recognized based on the feature points; the key points include the Yintang acupoint;

[0044] A laryngeal prominence determination unit, configured to determine the laryngeal prominence of the neck image to be recognized;

[0045] A first type of neck acupoint determination unit, configured to determine the positional relationship between the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel and the first type of neck acupoints, and determine the first type of neck acupoints based on the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationship, the unit vertical cun and the unit horizontal cun;

[0046] The second type of neck acupoint determination unit is used to determine the association relationship between the first type of neck acupoints and the second type of neck acupoints, and determine the second type of neck acupoints based on the association relationship and the first type of neck acupoints.

[0047] On the other hand, the present invention also provides a massage robot, including an acupoint recognition subsystem for determining neck acupoints and a massage subsystem for massaging based on the neck acupoints. The acupoint recognition subsystem includes a memory and a processor, wherein,

[0048] The memory is used to store programs;

[0049] The processor is coupled to the memory and is used to execute the programs stored in the memory to implement the steps in the neck acupoint recognition method in any of the above possible implementation manners.

[0050] On the other hand, the present invention also provides a digital human, including a virtual neck model and neck acupoints marked in the virtual neck model. The neck acupoints are determined based on the neck acupoint recognition method, and the neck acupoint recognition method is the neck acupoint recognition method in any of the above possible implementation manners.

[0051] The beneficial effects of adopting the above embodiments are as follows: The neck acupoint recognition method provided by the present invention realizes the recognition of the first type of neck acupoints only by determining the positions of the Adam's apple, Yintang acupoint, Conception Vessel, Governor Vessel, and the first type of neck acupoints, and then determines the second type of neck acupoints based on the association relationship. Thus, all neck acupoints are obtained. Compared with the marking of neck acupoints, the Adam's apple is a relatively obvious feature of the human body, and the key points and feature points for determining the Yintang acupoint, Conception Vessel, and Governor Vessel are also relatively obvious features. The calibration process is simpler, and there are various algorithms in the prior art that can directly determine the Adam's apple, key points, and feature points without sample marking and training processes, that is, the recognition process of neck acupoints does not rely on manual marking, ensuring the recognition accuracy and efficiency of neck acupoints. In addition, the present invention constructs unit vertical cun and unit horizontal cun through the determined key points, and eliminates the influence of individual differences on neck acupoints through unit vertical cun and unit horizontal cun, further ensuring the recognition accuracy of neck acupoints. Description of the Drawings

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative efforts.

[0053] Figure 1 Schematic flow chart of an embodiment of the neck acupoint recognition method provided by the present invention;

[0054] Figure 2 Schematic flow chart of an embodiment of determining the unit vertical cun and unit horizontal cun provided by the present invention;

[0055] Figure 3 For the present invention Figure 1 Schematic flow chart of an embodiment of step S102 in the present invention;

[0056] Figure 4 Schematic diagram of an embodiment of the detection result of the second detection algorithm;

[0057] Figure 5 For the present invention Figure 1 Schematic flow chart of an embodiment of determining the neck acupoints in step S103 in the present invention;

[0058] Figure 6 Schematic structural diagram of an embodiment of the neck acupoint recognition device provided by the present invention;

[0059] Figure 7 Schematic structural diagram of an embodiment of the massage robot provided by the present invention. Detailed implementation manners

[0060] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0061] It should be understood that the schematic drawings are not drawn to actual scale. The flowcharts used in the present invention illustrate the operations implemented according to some embodiments of the present invention. It should be understood that the operations in the flowchart may not be implemented in sequence, and the steps without logical context may be reversed or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart or remove one or more operations from the flowchart under the guidance of the content of the present invention. Some of the block diagrams shown in the drawings are functional entities, which do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor systems and / or microcontroller systems.

[0062] In the embodiments of the present invention, the descriptions such as "first" and "second" are only for descriptive purposes, and cannot be construed as indicating or implying their relative importance or implicitly specifying the quantity of the indicated technical features. Therefore, the technical features defined with "first" and "second" may explicitly or implicitly include at least one such feature.

[0063] Referring to "embodiment" herein means that the specific features, structures, or characteristics described in connection with the embodiment may be included in at least one embodiment of the present invention. The phrase appearing at various positions in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein may be combined with other embodiments.

[0064] The present invention provides a neck acupoint recognition method, device, massage robot, and digital human, which will be described separately below.

[0065] Figure 1 It is a schematic flowchart of an embodiment of the neck acupoint recognition method proposed in the embodiments of the present invention. As Figure 1 shown, the neck acupoint recognition method includes:

[0066] S101. Determine the key points related to the horizontal and vertical inches and the feature points related to the meridians of the neck image to be recognized, and determine the unit vertical inch and unit horizontal inch of the neck image to be recognized based on the key points, and determine the Conception Vessel and Governor Vessel of the neck image to be recognized based on the feature points; the key points include the Yintang acupoint.

[0067] Among them, the acquisition method of the neck image to be recognized is: based on a depth camera to collect the three-dimensional point cloud data of the human head and neck, and use the three-dimensional point cloud data as the neck image to be recognized.

[0068] In the specific embodiments of the present invention, the key points related to the horizontal and vertical inches and the feature points related to the veins of the neck image to be recognized are determined based on the first detection algorithm. The first detection algorithm is a mature algorithm in the prior art. Specifically, the first detection algorithm is an open-source algorithm that does not require training and can be directly used. Through this setting, the recognition efficiency of neck acupoints can be further improved.

[0069] Among them, the unit vertical inch refers to the conversion relationship between the image size in the first direction and the bone degree size, the first direction refers to the direction from head to foot, and the unit horizontal inch refers to the conversion relationship between the image size in the second direction and the bone degree size, and the second direction is perpendicular to the first direction.

[0070] It should be noted that: the "inch" in the embodiments of the present invention is not the commonly used measurement unit in life, but the equal division unit on the body. For different people, the length of the equal division unit is different.

[0071] It should also be noted that the selection of key points needs to be based on the bone measurement method. For example, the bone measurement method stipulates the bone measurement distance between the anterior hairline and the posterior hairline, so the key points need to be the anterior hairline and the posterior hairline. In other words, the key points refer to the points corresponding to the bone measurement in the bone measurement method.

[0072] S102. Determine the Adam's apple of the neck image to be recognized.

[0073] Specifically, determine the Adam's apple of the neck image to be recognized based on the second detection algorithm. Among them, the second detection algorithm is also a mature algorithm in the prior art. Specifically, the second detection algorithm is an open-source algorithm that does not require training and can be directly used. Through this setting, the recognition efficiency of neck acupoints can be further improved.

[0074] S103. Determine the positional relationship between the Adam's apple, the Yintang acupoint, the Conception Vessel, the Governor Vessel, and the first type of neck acupoints, and determine the first type of neck acupoints based on the Adam's apple, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationship, the unit vertical inch, and the unit horizontal inch.

[0075] Among them, the first type of neck acupoints refers to the neck acupoints directly related to the Adam's apple, the Yintang acupoint, the Conception Vessel, and the Governor Vessel.

[0076] S104. Determine the association relationship between the first type of neck acupoints and the second type of neck acupoints, and determine the second type of neck acupoints based on the association relationship and the first type of neck acupoints.

[0077] Among them, the second type of neck acupoints refers to the neck acupoints that are not directly related to the Adam's apple, the Yintang acupoint, the Conception Vessel, and the Governor Vessel and have a geometric positional relationship with the first type of neck acupoints.

[0078] It should be understood that the neck acupoint recognition method in the embodiments of the present invention can be implemented in any device based on neck acupoint recognition, such as acupuncture robots, massage robots, etc. Specifically, the neck acupoint recognition method is stored in the above-mentioned device in the form of a prepared program. When the device is started, the program is called, and the neck acupoint recognition method is implemented.

[0079] Compared with the prior art, the neck acupoint recognition method provided by the embodiments of the present invention realizes the recognition of the first type of neck acupoints only by determining the positions of the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel and the first type of neck acupoints through determining the positional relationship among them. Then, based on the association relationship, the second type of neck acupoints can be determined. Thus, all neck acupoints are obtained. Compared with the annotation of neck acupoints, the laryngeal prominence is a relatively obvious feature of the human body, and the key points and feature points for determining the Yintang acupoint, the Conception Vessel and the Governor Vessel are also relatively obvious features. The calibration process is simpler, and there are various algorithms in the prior art that can directly determine the laryngeal prominence, the key points and the feature points without sample annotation and training process, that is: the recognition process of neck acupoints does not rely on manual annotation, ensuring the recognition accuracy and recognition efficiency of neck acupoints. Moreover, the embodiments of the present invention construct the unit vertical cun and the unit horizontal cun through the determined key points, and eliminate the influence of individual differences on neck acupoints through the unit vertical cun and the unit horizontal cun, further ensuring the recognition accuracy of neck acupoints.

[0080] In some embodiments of the present invention, the bone measurement method selects the bone measurement as the horizontal bone measurement between the left anterior hair corner point and the right anterior hair corner point is 9 cun, and the vertical bone measurement between the anterior hairline and the Yintang acupoint is 3 cun. Correspondingly, the key points include the left anterior hair corner point and the right anterior hair corner point related to the horizontal cun, and the anterior hairline related to the vertical cun.

[0081] It should be understood that the bone measurement is not limited to the above specific embodiments. For example, the vertical bone measurement can also be the bone measurement between the anterior hairline and the posterior hairline, which will not be elaborated one by one here.

[0082] The Conception Vessel and the Governor Vessel are respectively the midlines on the front and back of the human body, and the human body is a symmetrical structure. Therefore, in some embodiments of the present invention, the feature points should be symmetrical points. Specifically, the feature points include the left earlobe point, the right earlobe point, the left acromion point and the right acromion point.

[0083] Based on the above key points and feature points, in some embodiments of the present invention, the first detection algorithm includes a key point detection algorithm for determining the left anterior hair corner point, the right anterior hair corner point, the left earlobe point, the right earlobe point, the left acromion point, the right acromion point and the Yintang acupoint, and a face detection algorithm for determining the anterior hairline. Then, the determination of the key points related to the horizontal and vertical cun and the feature points related to the meridians in the neck image to be recognized in step S101 includes:

[0084] Detect the neck image to be recognized to obtain a plurality of candidate key points, and determine the left anterior hair corner point, the right anterior hair corner point, the first target point and the second target point related to the Yintang acupoint, and the left earlobe point, the right earlobe point, the left acromion point and the right acromion point among the plurality of candidate key points based on the mapping relationship between the candidate key point numbers and the candidate key points;

[0085] Determine the midpoint of the first target point and the second target point as the Yintang acupoint;

[0086] Determine the face rectangle of the neck image to be recognized, and use the frame line close to the top of the head in the face rectangle as the anterior hairline.

[0087] Among them, the neck image to be recognized can be detected based on the key point detection algorithm, and the face rectangle of the neck image to be recognized is determined based on the face detection algorithm.

[0088] It should be noted that: the detection result of the key point detection algorithm includes the numbers of candidate key points, that is, when determining the left front hairline point, right front hairline point, left earlobe point, right earlobe point, left shoulder peak point, right shoulder peak point, first target point and second target point, only need to determine the left front hairline point, right front hairline point, left earlobe point, right earlobe point, left shoulder peak point, right shoulder peak point, first target point and second target point from multiple candidate key points according to the pre-determined numbers of the left front hairline point, right front hairline point, left earlobe point, right earlobe point, left shoulder peak point, right shoulder peak point, first target point and second target point, further improving the efficiency of neck acupoint recognition.

[0089] In a specific embodiment of the present invention, the key point detection algorithm is Mediapipe or Dlib.

[0090] Mediapipe can extract 478 key points of the face. Among them, key points 8 and 9 are the first target point and the second target point respectively. The midpoint of the first target point and the second target point is the Yintang acupoint, and key points 103 and 332 are the left front hairline point and the right front hairline point respectively.

[0091] Dlib can extract 68 key points of the face, including facial features such as eyes, eyebrows, nose, and mouth. Among them, key points 21 and 22 are the first target point and the second target point respectively. The midpoint of the first target point and the second target point is the Yintang acupoint, and key points 18 and 25 are the left front hairline point and the right front hairline point respectively.

[0092] The specific numbers of the left earlobe point, right earlobe point, left shoulder peak point, and right shoulder peak point are not elaborated here.

[0093] In a specific embodiment of the present invention, the face detection algorithm is the RetinaFace algorithm or the MogFace algorithm. The detection results of both the RetinaFace algorithm and the MogFace algorithm include a detection frame, and the upper frame line of the detection frame is flush with the anterior hairline, and the lower frame line is flush with the chin.

[0094] The embodiment of the present invention can quickly determine the anterior hairline by using the face rectangle obtained by the face detection algorithm, further improving the efficiency of neck acupoint recognition.

[0095] In a specific embodiment of the present invention, as Figure 2 shown, determining the unit vertical dimension and the unit horizontal dimension of the neck image to be recognized based on key points, and determining the Conception Vessel and the Governor Vessel of the neck image to be recognized based on the feature points includes:

[0096] S201. Determine the vertical distance and the vertical bone dimension between the Yintang acupoint and the anterior hairline, and the horizontal distance and the horizontal bone dimension between the left anterior hair corner point and the right anterior hair corner point.

[0097] Among them, the vertical distance and the horizontal distance refer to the distances in the image coordinate system where the neck image to be recognized is located, and the horizontal bone dimension and the vertical bone dimension refer to the distances in the bone measurement method.

[0098] S202. Determine the unit vertical dimension based on the vertical distance and the vertical bone dimension, and determine the unit horizontal dimension based on the horizontal distance and the horizontal bone dimension.

[0099] Specifically, the unit vertical dimension is the ratio of the vertical bone dimension to the vertical distance, and the unit horizontal dimension is the ratio of the horizontal bone dimension to the horizontal distance.

[0100] S203. Determine the first connection line and the first midpoint of the left earlobe point and the right earlobe point, use the first straight line on the front part of the human body that is perpendicular to the first connection line and passes through the first midpoint as the Conception Vessel, determine the second connection line and the second midpoint of the left acromion point and the right acromion point, and use the second straight line on the back part of the human body that is perpendicular to the second connection line and passes through the second midpoint as the Governor Vessel.

[0101] By determining the unit horizontal dimension and the unit vertical dimension in the embodiment of the present invention, the adverse effects on the acupoint recognition result caused by different shooting angles, individual differences, etc. of the neck image to be recognized, resulting in different horizontal distances and vertical distances, are eliminated, and the neck acupoints are recognized with normalized dimensions, further improving the recognition accuracy.

[0102] In some embodiments of the present invention, as Figure 3 shown, step S102 includes:

[0103] S301. Determine multiple candidate reference points of the neck image to be recognized.

[0104] Among them, the candidate reference points include the Adam's apple.

[0105] It should be noted that since the Adam's apple is crucial for the accuracy of subsequent neck acupoint recognition, therefore, to ensure the recognition accuracy of the Adam's apple, before step S301, it is necessary to perform three-dimensional reconstruction on the three-dimensional point cloud data of the head and neck obtained by the depth camera using the Poisson reconstruction method or KinectFusion or DynamicFusion technology.

[0106] S302. Determine the Adam's apple based on the Adam's apple number and the pre - constructed association relationship between the candidate reference point numbers and the candidate reference point positions.

[0107] In the embodiment of the present invention, the Adam's apple is determined based on the pre - constructed association relationship between the candidate reference point numbers and the candidate reference point positions. There is no need to locate the Adam's apple through parameters such as coordinates. Only based on the numbers of the candidate reference points can the Adam's apple be located, which reduces the complexity of Adam's apple positioning and further improves the positioning efficiency of neck acupoints.

[0108] In a specific embodiment of the present invention, since the accurate identification of the Adam's apple is crucial for the accurate identification of neck acupoints, in a specific embodiment of the present invention, multiple candidate reference points of the neck image to be identified are determined based on HRNet (High - Resolution Network). HRNet generates reliable high - resolution representations by parallelly connecting sub - networks with different resolutions and performing multi - scale fusion between these sub - networks. The design of HRNet makes it have high accuracy and efficiency in key - point detection tasks.

[0109] Specifically, when the second detection algorithm is HRNet, its detection result is as Figure 4 shown. It can be seen from Figure 4 that the candidate reference point numbered 1 is the Adam's apple.

[0110] In a specific embodiment of the present invention, the positional relationship is the bone - measurement positional relationship, that is: the positional relationship refers to the horizontal - inch and vertical - inch relationships between each neck acupoint and the Adam's apple or between neck acupoints.

[0111] And the identification of neck acupoints refers to the positions in the actual neck image to be identified. Therefore, in some embodiments of the present invention, as Figure 5 shown, step S103 of determining the first - type neck acupoints based on the Adam's apple, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationship, the unit vertical inch, and the unit horizontal inch includes:

[0112] S501. Determine the bone - measurement position of the first - type neck acupoints based on the Adam's apple, the Yintang acupoint, the Conception Vessel, the Governor Vessel, and the positional relationship.

[0113] Among them, the bone - measurement position refers to the position measured in horizontal inches and vertical inches.

[0114] S502. Convert the bone - measurement position into an image position based on the unit vertical inch and the unit horizontal inch, and use the image position as the position where the first - type neck acupoints are located.

[0115] In the embodiment of the present invention, after determining the bone - measurement position, converting the bone - measurement position into an image position based on the unit vertical inch and the unit horizontal inch can improve the adaptability between the neck acupoints and the neck image to be identified, which is beneficial to subsequent post - processing operations on the identified neck acupoints.

[0116] It should be noted that the second type of neck acupoints are also bone positions. After the second type of neck acupoints are determined, they should also be converted into image positions based on the unit vertical inch and unit horizontal inch.

[0117] In some embodiments of the present invention, the first type of cervical acupoints include Renying, Futu, Tiantu, and Tianchuang located on the front of the neck, and Yamen and Dazhui located on the back of the neck; the position relationship includes:

[0118] Renying acupoint is 1.5 horizontal inches away from Adam's apple;

[0119] Futu point is 3 horizontal inches away from Adam's apple;

[0120] Tiantu acupoint is located on the Ren channel, below the Adam's apple and 2 inches away from the Adam's apple;

[0121] Tianchuang acupoint is 3.5 horizontal inches away from Adam's apple;

[0122] Yamen acupoint is located on the Du channel, 14.5 straight inches away from Yintang acupoint;

[0123] The Dazhui point is located on the Du meridian, 18 straight inches away from the Yintang point.

[0124] In some embodiments of the present invention, the second type of neck acupoints include Lianquan, Qishe, Shuitu, Quepen located on the front neck, and Fengfu, Fengchi, Jingbailao, and Tianzhu located on the back neck; the association relationship is:

[0125] Lianquan acupoint is located at the midpoint of the line connecting the chin and Adam's apple; the chin is the lower frame line of the rectangular frame of the human face;

[0126] Qishe point is 1.5 horizontal inches away from Tiantu point;

[0127] Shuitu acupoint is located on the Ren meridian and at the midpoint of the line connecting Qishe acupoint and Renying acupoint;

[0128] The Quepen point is the intersection of the extended line of the connection between the Tiantu point and the Qishe point and the target line segment. The target line segment is a line segment that is perpendicular to the line connecting the left nipple and the right nipple and passes through the left nipple or the right nipple. Among them, the left nipple and the right nipple can be determined based on open source models such as Faster-CNN, SSD and YOLO.

[0129] Tianding acupoint is located at the midpoint of the line connecting Futu acupoint and Quepen acupoint;

[0130] Fengfu acupoint is located above Yamen acupoint, 0.5 cun away from Yamen acupoint;

[0131] Fengchi point is 1.3 horizontal inches away from Fengfu point;

[0132] The Bailao point on the neck is located above the Dazhui point, 2 vertical inches and 1 horizontal inch away from the Dazhui point;

[0133] Tianzhu acupoint is located below Yamen acupoint, 0.3 vertical cun away from Yamen acupoint and 1.3 horizontal cun away.

[0134] It should be noted that when determining the position of Tianzhu acupoint, the positional relationship between cervical vertebrae is utilized. Specifically, assuming that each cervical vertebra has the same size, there are 6 cervical vertebrae between Yamen acupoint and Dazhui acupoint. From the above positional relationship, it can be known that the distance between Yamen acupoint and Dazhui acupoint is 3.5 vertical cun, so each cervical vertebra = 3.5 vertical cun / 6 = 0.6 vertical cun. Tianzhu acupoint is separated from Yamen acupoint by half a cervical vertebra, so the distance between Tianzhu acupoint and Yamen acupoint can be obtained as 0.3 vertical cun.

[0135] In the process of determining the positional relationship in the embodiment of the present invention, the position of the human cervical vertebra is referred to, which further ensures the accuracy of the positional relationship.

[0136] Since the bone measurement used in the foregoing embodiment is the bone measurement between Yintang acupoint and the anterior hairline as the reference for identifying neck acupoints. And the bone measurement can also be the bone measurement between the anterior hairline and the posterior hairline. Therefore, in order to further verify the accuracy of the identified neck acupoints, in some embodiments of the present invention, the neck acupoint identification method further includes:

[0137] Determine the posterior hairline based on the anterior hairline, and the posterior hairline is 12 vertical cun away from the anterior hairline;

[0138] Verify the positional relationship based on the posterior hairline and Yamen acupoint, Dazhui acupoint and Fengfu acupoint.

[0139] Specifically, determine Yamen acupoint, Dazhui acupoint and Fengfu acupoint by verifying the positional relationship based on the posterior hairline and Yamen acupoint, Dazhui acupoint and Fengfu acupoint, and judge whether they coincide / are consistent with Yamen acupoint, Dazhui acupoint and Fengfu acupoint determined based on Yintang acupoint. If they coincide / are consistent, it is determined that the positional relationship is accurate; otherwise, it is inaccurate.

[0140] In the specific embodiment of the present invention, Yamen acupoint is located above the posterior hairline, 0.5 vertical cun away from the posterior hairline, Dazhui acupoint is located below the posterior hairline, 3 vertical cun away from the posterior hairline, and Fengfu acupoint is located below the posterior hairline, 1 vertical cun away from the posterior hairline.

[0141] In summary, the embodiment of the present invention further verifies the accuracy of neck acupoints through the posterior hairline.

[0142] In order to better implement the neck acupoint identification method in the embodiment of the present invention, correspondingly, on the basis of the neck acupoint identification method, the embodiment of the present invention also provides a neck acupoint identification device, as Figure 6 shown, the neck acupoint identification device 600 includes:

[0143] A reference information determination unit 601 is configured to determine key points related to the horizontal and vertical dimensions and feature points related to the meridians of the neck image to be recognized, determine the unit vertical dimension and unit horizontal dimension of the neck image to be recognized based on the key points, and determine the Conception Vessel and Governor Vessel of the neck image to be recognized based on the feature points; the key points include the Yintang acupoint.

[0144] A laryngeal prominence determination unit 602 is configured to determine the laryngeal prominence of the neck image to be recognized.

[0145] A first type of neck acupoint determination unit 603 is configured to determine the positional relationship between the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel, and the first type of neck acupoints, and determine the first type of neck acupoints based on the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationship, the unit vertical dimension, and the unit horizontal dimension.

[0146] A second type of neck acupoint determination unit 604 is configured to determine the association relationship between the first type of neck acupoints and the second type of neck acupoints, and determine the second type of neck acupoints based on the association relationship and the first type of neck acupoints.

[0147] The neck acupoint recognition device 600 provided in the above embodiment can implement the technical solutions described in the above embodiment of the neck acupoint recognition method. The specific implementation principles of the above modules or units can be referred to the corresponding content in the above embodiment of the neck acupoint recognition method, which will not be elaborated here.

[0148] As Figure 7 shown, the present invention also correspondingly provides a massage robot 700. The massage robot 700 includes an acupoint recognition subsystem 710 for determining neck acupoints and a massage subsystem 720 for massaging based on the neck acupoints. The acupoint recognition subsystem 710 includes a processor 711, a memory 712, and a display 713. Figure 7 Only some components of the massage robot 700 are shown, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be alternatively implemented.

[0149] The memory 712 may be an internal storage unit of the acupoint recognition subsystem 710 in some embodiments, such as the hard disk or memory of the acupoint recognition subsystem 710. The memory 712 may also be an external storage device of the acupoint recognition subsystem 710 in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the acupoint recognition subsystem 710.

[0150] Furthermore, the memory 712 may also include both the internal storage unit of the acupoint recognition subsystem 710 and the external storage device. The memory 712 is used to store the application software for installing the acupoint recognition subsystem 710 and various types of data.

[0151] In some embodiments, the processor 711 may be a central processing unit (CPU), a microprocessor, or other data processing chips, which are used to run the program code stored in the memory 712 or process data, such as the neck acupoint recognition method in the present invention.

[0152] In some embodiments, the display 713 may be an LED display, a liquid crystal display, a touch liquid crystal display, and an OLED (Organic Light-Emitting Diode) toucher, etc. The display 713 is used to display the information of the acupoint recognition subsystem 710 and to display a visual user interface. The components 711-713 of the acupoint recognition subsystem 710 communicate with each other through the system bus.

[0153] In some embodiments of the present invention, when the processor 711 executes the neck acupoint recognition program in the memory 712, the following steps may be implemented:

[0154] Determine the key points related to the horizontal and vertical inches and the characteristic points related to the meridians of the neck image to be recognized, and determine the unit vertical inch and unit horizontal inch of the neck image to be recognized based on the key points, and determine the Conception Vessel and Governor Vessel of the neck image to be recognized based on the characteristic points; the key points include the Yintang acupoint;

[0155] Determine the laryngeal prominence of the neck image to be recognized;

[0156] Determine the positional relationship between the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel and the first type of neck acupoints, and determine the first type of neck acupoints based on the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationship, the unit vertical inch and the unit horizontal inch;

[0157] Determine the association relationship between the first type of neck acupoints and the second type of neck acupoints, and determine the second type of neck acupoints based on the association relationship and the first type of neck acupoints.

[0158] It should be understood that when the processor 711 executes the neck acupoint recognition program in the memory 712, in addition to the above functions, other functions may also be implemented. For specific details, please refer to the description of the relevant method embodiments above.

[0159] Among them, as Figure 7 shown, the massage subsystem 720 includes a path planning unit 721 and a massage unit 722. The path planning unit 721 is used to perform path planning according to the moxibustion needle position and acupoint coordinates in the massage unit 722 to generate a planned path, and the massage unit 722 moves the moxibustion needle to the position where the acupoint is located according to the planned path.

[0160] In an embodiment of the present invention, by setting up a massage robot 700, it can replace doctors to complete long-term massage treatment work, reduce the burden on doctors, and lower medical consumption.

[0161] On the other hand, an embodiment of the present invention also provides a digital human, including a virtual neck model and neck acupoints marked in the virtual neck model, wherein the neck acupoints are determined based on a neck acupoint recognition method, and the neck acupoint recognition method is the neck acupoint recognition method in any of the above embodiments.

[0162] In an embodiment of the present invention, by setting up a digital human, training purposes can be achieved through the digital human, such as training on the recognition of neck acupoints.

[0163] Those skilled in the art can understand that all or part of the processes of implementing the above embodiment methods can be completed by instructing relevant hardware (such as a processor, a controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a magnetic disk, an optical disk, a read-only memory, or a random access memory, etc.

[0164] The above has introduced in detail the neck acupoint recognition method, device, massage robot, and digital human provided by the present invention. Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A neck acupoint recognition method, characterized in that, Including: Determine the key points related to the horizontal dimension and vertical dimension and the feature points related to the meridians of the neck image to be recognized, determine the unit vertical dimension and unit horizontal dimension of the neck image to be recognized based on the key points, and determine the Conception Vessel and Governor Vessel of the neck image to be recognized based on the feature points; the key points include the Yintang acupoint; Determine the laryngeal prominence of the neck image to be recognized; Determine the positional relationships among the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel and the first type of neck acupoints, and determine the first type of neck acupoints based on the laryngeal prominence, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationships, the unit vertical dimension and the unit horizontal dimension; Determine the association relationship between the first type of neck acupoints and the second type of neck acupoints, and determine the second type of neck acupoints based on the association relationship and the first type of neck acupoints; The first type of neck acupoints include the Renying acupoint, Futu acupoint, Tiantu acupoint, Tianchuang acupoint located in the front neck and the Yamen acupoint, Dazhui acupoint located in the back neck; the positional relationships include: the Renying acupoint is 1.5 horizontal inches away from the laryngeal prominence; the Futu acupoint is 3 horizontal inches away from the laryngeal prominence; The Tiantu acupoint is located on the Conception Vessel, below the laryngeal prominence and 2 vertical inches away from the laryngeal prominence; the Tianchuang acupoint is 3.5 horizontal inches away from the laryngeal prominence; the Yamen acupoint is located on the Governor Vessel, 14.5 vertical inches away from the Yintang acupoint; the Dazhui acupoint is located on the Governor Vessel, 18 vertical inches away from the Yintang acupoint; The second type of neck acupoints include the Lianquan acupoint, Qishe acupoint, Shuitu acupoint, Quepen acupoint, Tianding acupoint located in the front neck and the Fengfu acupoint, Fengchi acupoint, Jingbailao acupoint, Tianzhu acupoint located in the back neck; the association relationship is: the Lianquan acupoint is located at the midpoint of the line connecting the chin and the laryngeal prominence; the Qishe acupoint is 1.5 horizontal inches away from the Tiantu acupoint; the Shuitu acupoint is located on the Conception Vessel and at the midpoint of the line connecting the Qishe acupoint and the Renying acupoint; the Quepen acupoint is the intersection point of the extension line of the line connecting the Tiantu acupoint and the Qishe acupoint and the target line segment, and the target line segment is the line segment perpendicular to the line connecting the left nipple and the right nipple and passing through the left nipple or the right nipple; the Tianding acupoint is located at the midpoint of the line connecting the Futu acupoint and the Quepen acupoint; the Fengfu acupoint is above the Yamen acupoint, 0.5 vertical inches away from the Yamen acupoint; the Fengchi acupoint is 1.3 horizontal inches away from the Fengfu acupoint; the Jingbailao acupoint is above the Dazhui acupoint, 2 vertical inches and 1 horizontal inch away from the Dazhui acupoint; the Tianzhu acupoint is below the Yamen acupoint, 0.3 vertical inches and 1.3 horizontal inches away from the Yamen acupoint.

2. The neck acupoint recognition method according to claim 1, characterized in that The key points further include the left front hair corner point and the right front hair corner point related to the horizontal dimension and the anterior hairline related to the vertical dimension; the feature points include the left earlobe point, the right earlobe point, the left shoulder peak point and the right shoulder peak point; then the determination of the key points related to the horizontal dimension and vertical dimension and the feature points related to the meridians of the neck image to be recognized includes: Detect the neck image to be recognized to obtain multiple candidate key points, and determine the left front hair corner point, the right front hair corner point, the first target point and the second target point related to the Yintang acupoint, the left earlobe point, the right earlobe point, the left shoulder peak point and the right shoulder peak point among the multiple candidate key points based on the mapping relationship between the candidate key point numbers and the candidate key points; Determine the midpoint of the first target point and the second target point as the Yintang acupoint; Determine the face rectangular frame of the neck image to be recognized, and use the frame line close to the top of the head in the face rectangular frame as the anterior hairline, and the chin as the lower frame line of the face rectangular frame of the neck image to be recognized.

3. The neck acupoint recognition method according to claim 2, wherein Determine the unit vertical cun and unit horizontal cun of the neck image to be recognized based on the key points, and determine the Conception Vessel and Governor Vessel of the neck image to be recognized based on the feature points, including: Determine the longitudinal distance, longitudinal bone cun between the Yintang acupoint and the anterior hairline, and the transverse distance, transverse bone cun between the left front hair corner point and the right front hair corner point; Determine the unit vertical cun based on the longitudinal distance and the longitudinal bone cun, and determine the unit horizontal cun based on the transverse distance and the transverse bone cun; Determine the first connection line and the first midpoint of the left earlobe point and the right earlobe point, use the first straight line in the front part of the human body perpendicular to the first connection line and passing through the first midpoint as the Conception Vessel, determine the second connection line and the second midpoint of the left shoulder peak point and the right shoulder peak point, and use the second straight line in the back part of the human body perpendicular to the second connection line and passing through the second midpoint as the Governor Vessel.

4. The neck acupoint recognition method according to claim 1, characterized in that, The determination of the Adam's apple of the neck image to be recognized includes: Determine multiple candidate reference points and multiple candidate reference point numbers of the neck image to be recognized; Determine the Adam's apple based on the Adam's apple number of the Adam's apple and the pre-constructed association relationship between the candidate reference point numbers and the candidate reference point positions.

5. The neck acupoint recognition method according to claim 1, characterized in that, The positional relationship is the bone measurement positional relationship; then determine the first type of neck acupoints based on the Adam's apple, the Yintang acupoint, the Conception Vessel, the Governor Vessel, the positional relationship, the unit vertical cun and the unit horizontal cun, including: Determine the bone measurement position of the first type of neck acupoints based on the Adam's apple, the Yintang acupoint, the Conception Vessel, the Governor Vessel and the positional relationship; Convert the bone measurement position into an image position based on the unit vertical cun and the unit horizontal cun, and use the image position as the position where the first type of neck acupoints are located.

6. A neck acupoint recognition device, characterized in that, Including: A reference information determination unit for determining key points related to horizontal cun and vertical cun and feature points related to meridians of the neck image to be recognized, and determining the unit vertical cun and unit horizontal cun of the neck image to be recognized based on the key points, and determining the Conception Vessel and Governor Vessel of the neck image to be recognized based on the feature points; the key points include the Yintang acupoint; An Adam's apple determination unit for determining the Adam's apple of the neck image to be recognized; A first-category neck acupoint determination unit is used to determine the positional relationship between the Adam's apple, the Yintang acupoint, the Ren meridian, the Governor meridian, and the first-category neck acupoint, and determine the first-category neck acupoint based on the Adam's apple, the Yintang acupoint, the Ren meridian, the Governor meridian, the positional relationship, the unit vertical inch, and the unit horizontal inch; A second-category neck acupoint determination unit, configured to determine an association relationship between the first-category neck acupoint and the second-category neck acupoint, and determine the second-category neck acupoint based on the association relationship and the first-category neck acupoint; The first type of neck acupoints include Renying, Futu, Tiantu and Tianchuang located on the front neck, and Yamen and Dazhui located on the back neck; the position relationship includes: the Renying point is 1.5 horizontal inches away from the Adam's apple; the Futu point is 3 horizontal inches away from the Adam's apple; The Tiantu point is located on the Ren channel, below the Adam's apple and 2 inches away from the Adam's apple; the Tianchuang point is 3.5 inches away from the Adam's apple; the Yamen point is located on the Governor Vessel, 14.5 inches away from the Yintang point; the Dazhui point is located on the Governor Vessel, 18 inches away from the Yintang point; The second type of neck acupoints include Lianquan, Qishe, Shuitu, Quepen and Tianding located on the front neck, and Fengfu, Fengchi, Jingbailao and Tianzhu located on the back neck; the association relationship is: Lianquan is located at the midpoint of the line connecting the chin and Adam's apple; Qishe is 1.5 horizontal inches away from Tiantu; Shuitu is located on the Renmai and at the midpoint of the line connecting Qishe and Renying; Quepen is the intersection of the extended line connecting Tiantu and Qishe and the target line segment. point, the target line segment is a line segment perpendicular to the line connecting the left nipple and the right nipple and passing through the left nipple or the right nipple; the Tianding point is located at the midpoint of the line connecting the Futu point and the Quepen point; the Fengfu point is located above the Yamen point, 0.5 vertical inches away from the Yamen point; the Fengchi point is 1.3 horizontal inches away from the Fengfu point; the Jingbailao point is located above the Dazhui point, 2 vertical inches and 1 horizontal inch away from the Dazhui point; the Tianzhu point is located below the Yamen point, 0.3 vertical inches and 1.3 horizontal inches away from the Yamen point.

7. A massage robot, characterized in that, The invention comprises an acupoint identification subsystem for determining neck acupoints and a massage subsystem for performing massage based on the neck acupoints, wherein the acupoint identification subsystem comprises a memory and a processor, wherein: The memory is used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the neck acupoint identification method described in any one of claims 1 to 5 above.

8. A digital human, characterized in that, It includes a virtual neck model and neck acupoints marked in the virtual neck model, wherein the neck acupoints are determined based on a neck acupoint recognition method, and the neck acupoint recognition method is the neck acupoint recognition method described in any one of claims 1 to 5 above.

Images