Full-automatic shoulder, neck and head massage method capable of accurately searching acupuncture points

Abstract

The invention provides a full-automatic shoulder, neck and head massage method capable of accurately searching acupuncture points, which comprises the following steps of: fixing the head of a user on a shoulder, neck and head massage device, and acquiring a shoulder, neck and head image of the user through a visual detection device of the shoulder, neck and head massage device; generating a shoulder-neck-head three-dimensional model according to the shoulder-neck-head image; inputting the shoulder, neck and head images into a trained acupuncture point recognition model to recognize the positions of preset acupuncture points, and determining the positions of other acupuncture points according to the positions of the preset acupuncture points and a shoulder, neck and head three-dimensional model; controlling the massage head to massage the target acupuncture point according to the massage mode, position and force selected by the user; user feedback is monitored in real time through a plurality of sensors, and the position and force of the massage head are dynamically adjusted. Through the combination of a deep learning algorithm and a three-dimensional mechanical structure, the positioning precision of the shoulder, neck and head acupuncture points is remarkably improved, and the problem that a traditional massage instrument cannot perform precise positioning is solved.

Description

Technical Field

[0001] The present invention relates to the field of massage, and particularly discloses a shoulder, neck and head massage method for automatically and accurately finding acupoints. Background Art

[0002] The existing massage and physiotherapy devices on the market can relax muscles, soothe nerves, relieve fatigue, significantly reduce muscle soreness, and relax the body to reduce stress, but they cannot accurately locate acupoints like traditional Chinese medicine to achieve a therapeutic effect.

[0003] There are a series of problems in the massage process of the existing massage devices in the prior art. For example, mechanical massage is inevitably restricted by factors such as the patient's body shape, massage time length, and massage strength. At the same time, it has high requirements for the accurate positioning of acupoints and the effect of massage heads. Moreover, the control of massage strength during the massage by the massage device cannot be timely feedback for variable control, and it is entirely controlled by the parameters set by the massage device. It is impossible to accurately control the comfort of the massage recipient. For users, they need to adjust their positions and massage intensities by themselves, and the experience effect is not satisfactory, which will result in the failure to achieve the massage effect due to frequent changes in massage positions and intensities, and the low massage work efficiency. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a shoulder, neck and head massage method for automatically and accurately finding acupoints to improve the above problems.

[0005] The embodiment of the present invention provides a shoulder, neck and head massage method for automatically and accurately finding acupoints, which includes the following steps:

[0006] S101, fixing the user's head on the shoulder, neck and head massage device, and collecting the shoulder, neck and head image of the user through the visual detection device on the shoulder, neck and head massage device;

[0007] S102, generating a three-dimensional model of the shoulder, neck and head according to the shoulder, neck and head image;

[0008] S103, inputting the shoulder, neck and head image into the trained acupoint recognition model to identify the positions of predetermined acupoints, and then determining the positions of other acupoints according to the positions of the predetermined acupoints and the three-dimensional model of the shoulder, neck and head;

[0009] S104, controlling the massage head to massage the target acupoints according to the massage mode, position and strength selected by the user;

[0010] S105, real-time monitoring of the user's feedback through a plurality of sensors provided, and dynamically adjusting the position and strength of the massage head.

[0011] Preferably, the acupoint recognition model is the YOLOv8 model, which includes a Backbone module for extracting image features, a Neck module for feature fusion, and an output layer Head module. The Backbone module includes a CBS structure, a C3-1 structure, a C3-2 structure, and an SPPF structure; the Neck module includes an FPN structure and a PAN structure, and it is trained through the following steps:

[0012] Collect human shoulder, neck, and head images in different body shapes and postures, unify the size of the human shoulder, neck, and head images to a fixed pixel, use a labeling tool to label the acupoints on the human shoulder, neck, and head images to generate a labeling file, and divide the labeling file into a training set and a test set according to a ratio;

[0013] Build the YOLOv8 model, use the training set to train the YOLOv8 model, and use the test set to test and verify the trained YOLOv8 model.

[0014] Preferably, in step S103, use the trained YOLOv8 model to detect the positions of the two ears from the shoulder, neck, and head images, calculate the specific position of the Fengchi acupoint according to the horizontal height of the two earlobes being flush, and further locate the positions of other shoulder, neck, and head acupoints in combination with the three-dimensional model of the shoulder, neck, and head.

[0015] Preferably, the shoulder, neck, and head massage device further includes:

[0016] An adjustable support frame capable of adjusting the height by telescoping;

[0017] A human body positioning pad provided on the adjustable support frame and having a positioning seat suitable for the size of the human head;

[0018] A three-dimensional massage device fixed on the adjustable support frame, which has massage heads and can move freely in three directions: front and back, left and right, and up and down;

[0019] A horizontal support frame fixed on the adjustable support frame and located below the human body positioning pad for fixing the visual detection device; the visual detection device is connected to the horizontal support frame through a connecting arm, and its data acquisition end is higher than the human body positioning pad;

[0020] A visualization operation platform fixed on the adjustable support frame and located below the horizontal support frame.

[0021] Preferably, the three-dimensional massage device includes a support cross bar, a longitudinal lead screw mechanism, a transverse lead screw mechanism, a vertical lead screw mechanism, and a massage head mechanism. The longitudinal lead screw mechanism is installed above the support cross bar to control the front and back movement of the massage head mechanism; the transverse lead screw mechanism and the vertical lead screw mechanism are installed on the support cross bar to realize the left and right movement and the up and down movement of the massage head mechanism.

[0022] Preferably, the massage head mechanism includes a massage head fixing seat, a pressure sensor, a spring, and a massage head. The massage head is fixed on the massage head fixing seat through the spring, and the pressure sensor is arranged on the massage head.

[0023] Preferably, the height of the human body positioning pad on the horizontal support frame is adjusted by a tapered sleeve, a pressing piece, and a fine-tuning bolt.

[0024] Preferably, the visualization operation platform includes a touch screen, and the user selects a massage mode, position, and intensity through the touch screen.

[0025] Preferably, during the movement of the massage head mechanism, it judges whether the massage position is an accurate acupoint through the data of facial expression recognition and biosensors. If it is determined not to be an acupoint, the massage head is controlled to search near the massage area until an accurate acupoint is found.

[0026] Through the combination of deep learning algorithms and three-dimensional mechanical structures in the present invention, the positioning accuracy of acupoints on the shoulders, necks, and heads is significantly improved, solving the problem that traditional massage devices cannot accurately locate. Further, through pressure sensors, biosensors, and facial expression recognition technology, it can monitor user feedback in real time and dynamically adjust the massage intensity, significantly improving the user's comfort and massage effect. In particular, the device supports height adjustment and multi-dimensional movement, is suitable for users of different body types and heights, and meets diverse massage needs. In addition, the whole process from acupoint positioning to massage execution is automated, reducing the cumbersome manual adjustment operations of users and improving massage efficiency. Description of the Drawings

[0027] Figure 1 It is a schematic flow chart of the full-automatic accurate acupoint-finding shoulder, neck, and head massage method according to an embodiment of the present invention;

[0028] Figure 2 It is a schematic structural diagram of a shoulder, neck, and head massage device according to an embodiment of the present invention;

[0029] Figure 3 It is another schematic structural diagram of a shoulder, neck, and head massage device according to an embodiment of the present invention;

[0030] Figure 4 It is a schematic structural diagram of an adjustable support frame according to an embodiment of the present invention;

[0031] Figure 5 It is a schematic structural diagram of a massage head mechanism according to an embodiment of the present invention;

[0032] The reference numerals are as follows:

[0033] 1. Three-dimensional massage device; 2. Human body positioning pad; 3. Horizontal support frame; 4. Visual operation platform; 5. Adjustable support frame; 6. Visual detection device; 11. Support cross bar; 12. Longitudinal lead screw mechanism; 13. Transverse lead screw mechanism; 14. Vertical lead screw mechanism; 15. Massage head mechanism; 151. Massage head fixing seat; 152. Pressure sensor; 153. Spring; 154. Massage head; 52. Taper sleeve; 53. Compression piece; 54. Fine adjustment bolt; 61. Camera support frame; 62. Binocular structured light camera. Detailed implementation manners

[0034] The present invention provides a shoulder, neck and head massage method for automatically and accurately finding acupoints, which realizes the automatic recognition and precise positioning of acupoints on the shoulder, neck and head by combining deep learning algorithms and three-dimensional mechanical structure design, and can dynamically adjust the massage intensity according to user feedback.

[0035] Please refer to Figure 1 , an embodiment of the present invention provides a shoulder, neck and head massage method for automatically and accurately finding acupoints, which includes the following steps:

[0036] S101, fix the user's head on the shoulder, neck and head massage device, and collect the user's shoulder, neck and head image through the visual detection device on the shoulder, neck and head massage device;

[0037] S102, generate a three-dimensional model of the shoulder, neck and head according to the shoulder, neck and head image;

[0038] S103, input the shoulder, neck and head image into the trained acupoint recognition model to identify the positions of predetermined acupoints, and then determine the positions of other acupoints according to the positions of the predetermined acupoints and the three-dimensional model of the shoulder, neck and head;

[0039] S104, control the massage head to massage the target acupoints according to the massage mode, position and intensity selected by the user;

[0040] S105, real-time monitor the user feedback through multiple set sensors, and dynamically adjust the position and intensity of the massage head.

[0041] The following will describe in detail the specific implementation manners of the present invention with reference to the accompanying drawings.

[0042] 1. Data acquisition and preprocessing

[0043] In this embodiment, it is first necessary to collect human shoulder, neck, and head acupoint data in different body shapes and postures, including but not limited to Fengchi acupoint, Jianjing acupoint, Tianzhu acupoint, Dazhui acupoint, Baihui acupoint, etc. These data can be obtained through actual shooting and annotation, or from publicly available medical image databases. The collected image data is preprocessed, specifically including uniformly adjusting the image size to 608×608 pixels to meet the input requirements of the YOLOv8 network model. The ears in each picture in the dataset are annotated using the image annotation tool Labelimg to generate a TXT format annotation file, which contains the position information of the ears. The processed human hindbrain dataset is divided into a training set and a test set according to a ratio of 9:1 to ensure that the training and test data of the model have high representativeness.

[0044] 2. Construction and Training of Acupoint Recognition Model

[0045] In this embodiment, the acupoint recognition model can be a model of the YOLO series, such as the YOLOv8 network structure model. The YOLOv8 network model includes a Backbone module for extracting image features, a Neck module for feature fusion, and an output layer Head module. The Backbone module contains CBS structures (Convolution + BatchNorm + SiLU), C3-1 structures, C3-2 structures, and SPPF structures (Spatial Pyramid Pooling - Fast). The Neck module contains FPN structures (Feature Pyramid Network) and PAN structures (Path Aggregation Network). The combination of these modules enables the model to effectively extract and fuse image features, improving the accuracy of acupoint localization.

[0046] Then, the YOLOv8 network structure model is trained using the training set. During the training process, set epochs to 100, the batch size to 10, the input image size to 608×608, and the initial weights to the official pre-trained weights, referring to YOLOv8s.pt. Through a large amount of training data, the model can learn the characteristics of human shoulder, neck, and head acupoints in different body shapes and postures, so as to accurately identify the acupoint positions in actual applications. During the training process, the performance of the model can be further optimized by adjusting hyperparameters such as the learning rate and optimizer. Of course, it should be noted that in other embodiments of the present invention, the parameters during the training process can be set according to actual needs, and the present invention will not elaborate here.

[0047] 3. Acupoint Localization and Testing

[0048] In this embodiment, the trained YOLOv8 model is used to test the test set.

[0049] 4. Operation Process of the Massage Device

[0050] In this embodiment, the user starts to use the neck, shoulder and head massage device of the present invention. First, according to their own height and body shape, the user adjusts the height of the human body positioning pad 2 through the tapered sleeve 52, the pressing piece 53 and the horizontal support frame 3 so that it can be exactly spliced with the edge of the bed. If further fine-tuning is required, the fine-tuning bolt 54 on the adjustable support frame 5 can be used to fine-tune the overall height. After the adjustment is completed, the user lies prone on the human body positioning pad 2, with the face placed downward in the positioning seat to ensure that the head position is fixed.

[0051] At this time, the binocular structured light camera 62 starts to capture images of the user's body. During the capture process, the camera is fixed above the three-dimensional massage device 1 through the camera support frame 61, and the capture field of view can cover the entire human body positioning pad 2. The images captured by the camera are used to obtain the image feature information of the user's head to shoulder through an image segmentation algorithm, and then three-dimensional reconstruction is performed to generate a three-dimensional model of the neck, shoulder and head. The generated three-dimensional model is displayed on the touch screen 42 through the visualization operation platform 4, and the user can clearly see their own neck, shoulder and head model and mark the acupoint positions that need to be massaged.

[0052] As Figure 2 and Figure 3 shown, the binocular structured light camera 62 captures images of the user's neck, shoulder and head, obtains the image feature information of the user's head to shoulder through an image segmentation algorithm, and then performs three-dimensional reconstruction to generate a three-dimensional model of the neck, shoulder and head. The YOLOv8 model combines the generated three-dimensional model to identify the acupoint positions on the user's neck, shoulder and head. In particular, for the positioning of the Fengchi acupoint, the model will detect the positions of the two ears and calculate the specific position of the Fengchi acupoint according to the horizontal height at which the two earlobes are flush, that is, 2.5 cm to the left and 2.5 cm to the right of the center point between the two earlobes. In this way, the model can accurately locate the Fengchi acupoint, providing an accurate reference point for subsequent massage.

[0053] In addition, the user can also select the massage mode, position and intensity through the touch screen 42. The three-dimensional model of the user's neck, shoulder and head will be displayed on the touch screen 42, and the acupoints that can be massaged will be marked. The user can select different massage modes according to their own needs, such as gentle mode, medium mode and strong mode, and can also select different acupoints for massage. After the selection is completed, the system will start the massage process, and the three-dimensional massage device 1 will control the massage head mechanism 15 to move to the corresponding acupoint positions according to the user's selection and the recognition result of the model.

[0054] 5. Dynamic Adjustment during Massage

[0055] In this embodiment, after the massage starts, the system monitors the user's feedback information in real time through a variety of biosensors and facial expression recognition technologies. Specifically, the massage head mechanism 15 includes a massage head fixing seat 151, a pressure sensor 152, a spring 153, and a massage head 154. The pressure sensor 152 is installed on the massage head fixing seat 151, and the massage head 154 is installed on the massage head fixing seat 151 in contact with the pressure sensor 152 through the spring 153. During the massage process, the pressure sensor 152 detects the massage force in real time and adjusts the massage intensity through the spring 153 to ensure that the massage force always remains within the comfortable range of the user.

[0056] At the same time, the camera on the touch screen 42 will take pictures of the user's facial expressions in real time. Through the facial expression recognition algorithm, the system can judge the user's comfort level. For example, if the user shows a painful expression, the system will automatically reduce the massage force; if the user shows a relaxed expression, the system will appropriately increase the massage force. In addition, the system is also equipped with biosensors such as a blood pressure sensor, a heart rate sensor, and a voice control sensor. These sensors can monitor the user's physiological parameters, such as blood pressure and heart rate, in real time. According to the changes in these physiological parameters, the system can further adjust the massage force and position to ensure that the user gets the best massage experience.

[0057] 6. Precise positioning and search of acupoint positions

[0058] In this embodiment, when the system initially identifies the acupoint positions of the user, it will start the massage head mechanism 15 for preliminary massage. During the massage process, the system will judge whether the massage position is an accurate acupoint through the data of facial expression recognition and biosensors. If it is determined that it is not an acupoint, the system will search near the massage area until an accurate acupoint is found. During the search process, the system will control the movement of the massage head mechanism 15 through a three-dimensional transmission mechanism (including a longitudinal lead screw mechanism 12, a transverse lead screw mechanism 13, and a vertical lead screw mechanism 14) to achieve precise adjustment of the acupoint position.

[0059] Specifically, the longitudinal lead screw mechanism 12 is installed above the support cross bar 11 and can control the front and back movement of the massage head mechanism 15; the transverse lead screw mechanism 13 and the vertical lead screw mechanism 14 are installed on the support cross bar 11 and can achieve the left and right movement and up and down movement of the massage head mechanism 15. Through the free movement in these three directions, the massage head mechanism 15 can flexibly adjust its position to ensure the precise positioning of the acupoints.

[0060] 7. User feedback and optimization of massage effects

[0061] In this embodiment, the system continuously monitors the user's feedback information, including facial expressions, blood pressure, heart rate, etc. According to this feedback information, the system dynamically adjusts the massage intensity and position to provide the user with the best massage experience. For example, if the user's blood pressure and heart rate show abnormal fluctuations during the massage, the system will automatically reduce the massage intensity to avoid discomfort to the user. If the user's facial expression shows relaxation, the system will appropriately increase the massage intensity to further improve the massage effect.

[0062] In addition, the user can also manually set the massage duration or manually end the massage through the touch screen 42. When the set massage duration ends or the user manually ends the massage, the system controls the three-dimensional massage device 1 to return to the initial position to prepare for the next massage.

[0063] 8. Maintenance and Upgrade of the Massage Device

[0064] To ensure the long-term stable operation of the massage device, the present invention also provides a set of maintenance and upgrade solutions. The user can regularly clean and check the device to ensure the normal operation of each mechanical component. In addition, the system supports online updates. The user can download the latest YOLO model weight file and software update through the visualization operation platform 4 to further improve the accuracy of acupoint positioning and the massage effect. In this way, the neck, shoulder and head massage device of the present invention can adapt to the changing user needs and technological progress.

[0065] The following further illustrates the usage process and effects of the present invention through a specific embodiment.

[0066] Suppose user A needs a neck, shoulder and head massage. User A first adjusts the height of the human body positioning pad 2 according to his own height and body shape through the tapered sleeve 52, the pressing piece 53 and the horizontal support frame 3 so that it can just be spliced with the edge of the bed. After the adjustment is completed, user A lies prone on the human body positioning pad 2 with his face down in the positioning seat to ensure that the head position is fixed. At this time, the binocular structured light camera 62 starts to capture the body image of user A. During the capture process, the camera obtains the image feature information from the user A's head to the shoulder through the image segmentation algorithm, and then performs three-dimensional reconstruction to generate a three-dimensional model of user A's neck, shoulder and head.

[0067] The generated three-dimensional model is displayed on the touch screen 42 through the visualization operation platform 4. User A can select different massage modes and acupoints. Suppose user A selects the gentle mode and massages the Fengchi acupoint. The system will start the massage head mechanism 15 and control the massage head 154 to move to the position of the Fengchi acupoint through the longitudinal lead screw mechanism 12, the transverse lead screw mechanism 13 and the vertical lead screw mechanism 14. During the massage process, the pressure sensor 152 continuously detects the massage intensity and adjusts the massage intensity through the spring 153 to ensure that the massage intensity always remains within the comfortable range of user A.

[0068] Meanwhile, the camera on the touch screen 42 will capture the facial expressions of user A in real time. Through the facial expression recognition algorithm, the system judges the comfort level of user A. Suppose user A shows a painful expression, the system will automatically reduce the massage intensity; if user A shows a relaxed expression, the system will appropriately increase the massage intensity. In addition, the system also monitors the physiological parameters of user A in real time through blood pressure sensors, heart rate sensors and voice control sensors, and further adjusts the massage intensity and position according to the changes of these parameters.

[0069] During the massage process, user A can manually set the massage duration or end the massage through the touch screen 42. Suppose the massage duration set by user A is 15 minutes. When 15 minutes are over, the system will control the three-dimensional massage device 1 to return to the initial position to prepare for the next massage.

[0070] In the embodiment of the present invention, through the deep learning algorithm and the three-dimensional mechanical structure design, the automatic recognition and precise positioning of the acupoints on the shoulder, neck and head are realized, and the massage intensity can be dynamically adjusted according to the user's feedback. In the specific implementation process, the user can adjust the height of the device according to his own height and body type to ensure that the device can adapt to different massage scenarios. The binocular structured light camera 62 is used to capture the image of the user's body, generate a three-dimensional model of the shoulder, neck and head, and combine with the YOLOv8 model to identify the acupoint positions. During the massage process, the system monitors the feedback information of the user in real time through facial expression recognition and biosensors, and dynamically adjusts the massage intensity and position to ensure that the user gets the best massage experience. The beneficial effect of the present invention is that it significantly improves the accuracy of acupoint positioning, solves the problem that traditional massage devices cannot accurately position, and at the same time improves the user's comfort and massage effect.

[0071] The above is the preferred implementation mode of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A shoulder, neck and head massage method for fully automatic and accurate acupuncture point search, characterized in that: The following steps are involved: S101, fixing the user's head on the shoulder, neck and head massage device, and collecting the user's shoulder, neck and head image through a visual detection device on the shoulder, neck and head massage device; S102, generating a three-dimensional model of the shoulder, neck and head according to the shoulder, neck and head image; S103, inputting the shoulder, neck and head image into a trained acupoint recognition model to identify the location of the predetermined acupoint, and then determining the location of other acupoints according to the location of the predetermined acupoint and the shoulder, neck and head three-dimensional model; S104, controlling the massage head to massage the target acupuncture points according to the massage mode, position and strength selected by the user; S105, monitor user feedback in real time through multiple sensors, and dynamically adjust the position and strength of the massage head.

2. The shoulder, neck and head massage method for fully automatic and accurate acupuncture point search according to claim 1 is characterized in that: The acupoint recognition model is a YOLOv8 model, which includes a Backbone module for extracting image features, a Neck module for feature fusion, and an output layer Head module. The Backbone module includes a CBS structure, a C3-1 structure, a C3-2 structure, and an SPPF structure; the Neck module includes an FPN structure and a PAN structure, and is trained by the following steps: Collect human shoulder, neck and head images in different body shapes and postures, unify the size of the human shoulder, neck and head images to fixed pixels, use annotation tools to annotate the acupoints on the human shoulder, neck and head images to generate annotation files, and divide the annotation files into training sets and test sets in proportion; Build a YOLOv8 model, use the training set to train the YOLOv8 model, and use the test set to test and verify the trained YOLOv8 model.

3. The shoulder, neck and head massage method for fully automatic and accurate acupuncture point search according to claim 1 is characterized in that: In step S103, the trained YOLOv8 model is used to detect the position of both ears from the shoulder, neck and head image, and the specific position of the Fengchi acupoint is calculated according to the horizontal height of the two earlobes, and the positions of other shoulder, neck and head acupoints are further located in combination with the shoulder, neck and head three-dimensional model.

4. The shoulder, neck and head massage method for fully automatic and accurate acupuncture point search according to claim 1 is characterized in that: The shoulder, neck and head massage device also includes: Adjustable support frame, which can be adjusted in height by telescoping; A human body positioning cushion is arranged on the adjustable support frame and is provided with a positioning seat suitable for the size of a human head; A three-dimensional massage device, fixed on the adjustable support frame, has a massage head and can move freely in three directions: front and back, left and right, and up and down; A horizontal support frame is fixed on the adjustable support frame and is located below the human body positioning pad, and is used to fix the visual detection device; the visual detection device is connected to the horizontal support frame through a connecting arm, and its data collection end is higher than the human body positioning pad; The visual operation platform is fixed on the adjustable support frame and is located below the horizontal support frame.

5. The shoulder, neck and head massage method for fully automatic and accurate acupuncture point search according to claim 4, characterized in that: The three-dimensional massage device includes a supporting cross bar, a longitudinal screw mechanism, a transverse screw mechanism, a vertical screw mechanism and a massage head mechanism. The longitudinal screw mechanism is installed above the supporting cross bar to control the forward and backward movement of the massage head mechanism; the transverse screw mechanism and the vertical screw mechanism are installed on the supporting cross bar to realize the left and right movement and up and down movement of the massage head mechanism.

6. The shoulder, neck and head massage method for fully automatic and accurate acupuncture point search according to claim 5, characterized in that: The massage head mechanism comprises a massage head fixing seat, a pressure sensor, a spring and a massage head. The massage head is fixed on the massage head fixing seat through the spring, and the pressure sensor is arranged on the massage head.

7. The shoulder, neck and head massage method for fully automatic and accurate acupuncture point search according to claim 4, characterized in that: The height of the human body positioning pad on the horizontal support frame is adjusted by a cone sleeve, a pressing sheet and a fine-tuning bolt.

8. The shoulder, neck and head massage method for fully automatic and accurate acupuncture point search according to claim 4, characterized in that: The visual operation platform includes a touch screen and a camera. The user selects the massage mode, position and strength through the touch screen, and the camera is used to capture the user's facial expression.

9. The shoulder, neck and head massage method for fully automatic and accurate acupuncture point search according to claim 8, characterized in that: During the movement, the massage head mechanism determines whether the massage position is an accurate acupuncture point through facial expression recognition and data from the biosensor. If it is determined to be a non-acupuncture point, the massage head is controlled to search near the massage point until the accurate acupuncture point is found.

Images