Face region positioning method and device, wearable device, and beauty care system

Abstract

The application discloses a face region positioning method and device, a wearable device, and a beauty care system. The face region positioning method comprises the following steps: when a user wears a wearable device and initializes a beauty care device based on preset face key points, acquiring electromagnetic induction data corresponding to the preset face key points; when the user uses the beauty care device, acquiring real-time electromagnetic induction data between the wearable device and the beauty care device; and determining a face region currently acted on by the beauty care device by using a preset face region positioning model according to the electromagnetic induction data corresponding to the preset face key points and the real-time electromagnetic induction data. The application considers the differences in face region distribution characteristics of different users, pre-trains a positioning model capable of distinguishing the face regions of different users, and combines the electromagnetic induction data of the face key points of the user determined in the initialization stage and the real-time electromagnetic induction data generated at present, so that the accurate positioning of the beauty care device on the face region of the user can be realized.

Description

Technical Field

[0001] This application relates to the field of beauty care technology, and in particular to a face region positioning method, device, wearable device, and beauty care system. Background Technology

[0002] Traditional medical beauty devices mainly switch between different beauty modes by setting buttons for different areas of the face. This method is relatively reliable but the user experience is poor.

[0003] To address the issue that traditional medical beauty devices cannot automatically switch between beauty modes, the automatic localization of the facial area affected by the device becomes the starting point. By automatically localizing the facial area affected by the device, precise skincare in different zones can be achieved, and personalized care suggestions can be automatically generated.

[0004] Currently, the facial region positioning scheme of beauty care devices based on magnetic induction positioning technology mainly uses magnetic field strength distribution models for positioning. However, this scheme does not take into account the differences in facial region distribution characteristics among different users. For example, for the same magnetic induction reading, it may correspond to the eye area of ​​user A and the cheek area of ​​user B at the same time. Therefore, simply using the magnetic field strength distribution model to process the real-time read magnetic induction data cannot achieve accurate positioning of the facial region. Summary of the Invention

[0005] This application provides a face region localization method, device, wearable device, and beauty care system to improve the accuracy of face region localization for different users.

[0006] The embodiments of this application adopt the following technical solutions:

[0007] In a first aspect, embodiments of this application provide a face region localization method, the face region localization method comprising:

[0008] When a user wears a wearable device and initializes the beauty care device based on preset facial key points, the electromagnetic induction data corresponding to the preset facial key points is obtained;

[0009] When a user is using a beauty device, real-time electromagnetic induction data between the wearable device and the beauty device is acquired.

[0010] Based on the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data, the preset facial region positioning model is used to locate the facial region currently being treated by the beauty care device.

[0011] Optionally, the preset face region localization model includes a face key point recognition sub-model and a partition localization sub-model. The step of using the preset face region localization model to locate the face region currently being used by the beauty care device, based on the electromagnetic induction data corresponding to the preset face key points and real-time electromagnetic induction data, includes:

[0012] Based on the electromagnetic induction data corresponding to the preset facial key points, the user identity is identified using the facial key point recognition sub-model to obtain the user identity corresponding to the preset facial key points.

[0013] Obtain the partition location sub-model corresponding to the user identity;

[0014] Based on the real-time electromagnetic induction data, the facial region currently being treated by the beauty care device is obtained by using the partitioned positioning sub-model for partitioned positioning.

[0015] Optionally, the step of using a preset face region positioning model to locate the face region currently being used by the beauty care device based on the electromagnetic induction data corresponding to the preset facial key points and real-time electromagnetic induction data includes:

[0016] The electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data are directly input into the preset facial region positioning model to obtain the facial region currently being used by the beauty care device.

[0017] Optionally, the preset face region localization model includes a face key point recognition sub-model and a region localization sub-model, and the preset face region localization model is trained in the following manner:

[0018] Acquire magnetic induction data of preset facial key points from multiple users;

[0019] A facial landmark recognition sub-model was trained using magnetic induction data of preset facial landmarks from multiple users.

[0020] Acquire magnetic induction data of the facial regions of multiple users;

[0021] The magnetic induction data of the facial regions of multiple users are used to train the corresponding partition localization sub-model for each user.

[0022] Optionally, training the facial landmark recognition sub-model using magnetic induction data of preset facial landmarks from multiple users includes:

[0023] Based on the magnetic induction data of preset facial key points of multiple users, construct the facial key point feature vector corresponding to each user;

[0024] The facial landmark recognition sub-model is trained using the facial landmark feature vector corresponding to each user.

[0025] Optionally, training a localization sub-model for each user using magnetic induction data of the facial regions of multiple users includes:

[0026] Based on the magnetic induction data of the facial regions of multiple users, construct the first region feature vector for each partition of each user;

[0027] Each user's corresponding partition localization sub-model is trained using the first region feature vector corresponding to each partition.

[0028] Optionally, the magnetic induction data of the face region includes magnetic induction data of the left face region and magnetic induction data of the right face region, and the partition localization sub-model includes a left face partition localization sub-model and a right face partition localization sub-model. The step of training the partition localization sub-model corresponding to each user using the magnetic induction data of multiple users' face regions includes:

[0029] Each user's left face region localization sub-model is trained using the magnetic induction data of each user's left face region.

[0030] Each user's right face region localization sub-model is trained using the magnetic induction data of the right face region.

[0031] Optionally, the preset face region localization model is trained in the following manner:

[0032] Acquire magnetic induction data of preset facial key points and magnetic induction data of facial regions from multiple users;

[0033] Based on the magnetic induction data of preset facial key points and facial regions of multiple users, a second region feature vector is constructed for each user.

[0034] The preset face region localization model is trained using the second region feature vector corresponding to each user.

[0035] Secondly, embodiments of this application also provide a face region positioning device, the face region positioning device comprising:

[0036] The first acquisition unit is used to acquire electromagnetic induction data corresponding to the preset facial key points when the user is wearing a wearable device and the beauty care device is initialized based on preset facial key points.

[0037] The second acquisition unit is used to acquire real-time electromagnetic induction data between the wearable device and the beauty device when the user is using the beauty device.

[0038] The positioning unit is used to locate the face area currently being treated by the beauty care device by using a preset face area positioning model based on the electromagnetic induction data corresponding to the preset facial key points and real-time electromagnetic induction data.

[0039] Thirdly, embodiments of this application also provide a wearable device, including:

[0040] Processor; and

[0041] A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform any of the aforementioned face region localization methods.

[0042] Fourthly, embodiments of this application also provide a beauty care system, which includes beauty care devices and wearable devices as described above.

[0043] Fifthly, embodiments of this application also provide a computer-readable storage medium storing one or more programs that, when executed by a wearable device including multiple applications, cause the wearable device to perform any of the methods described above.

[0044] The face region localization method adopted in this application embodiment can achieve the following beneficial effects: First, when the user wears a wearable device and initializes the beauty device based on preset facial key points, electromagnetic induction data corresponding to the preset facial key points is obtained. Then, when the user uses the beauty device, real-time electromagnetic induction data between the wearable device and the beauty device is obtained. Finally, based on the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data, a preset face region localization model is used to locate the face region currently being used by the beauty device. The face region localization method in this application embodiment considers the different facial region distribution characteristics of different users. A face region localization model capable of distinguishing different users is pre-trained. Using this face region localization model, combined with the electromagnetic induction data of the user's facial key points determined during the initialization phase and the currently generated real-time electromagnetic induction data, accurate localization of the beauty device in the user's face region can be achieved, improving the accuracy of face region localization. Attached Figure Description

[0045] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0046] Figure 1 This is a flowchart illustrating a face region localization method according to an embodiment of this application;

[0047] Figure 2 This is a schematic diagram of face region division in an embodiment of this application;

[0048] Figure 3 This is a schematic diagram of a face region localization process in an embodiment of this application;

[0049] Figure 4 This is a schematic diagram of another face region localization process in an embodiment of this application;

[0050] Figure 5 This is a schematic diagram of the structure of a face region positioning device according to an embodiment of this application;

[0051] Figure 6 This is a schematic diagram of the structure of a wearable device according to an embodiment of this application;

[0052] Figure 7 This is a schematic diagram of the architecture of a beauty care system according to an embodiment of this application. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0054] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0055] This application provides a method for locating facial regions, such as... Figure 1 The diagram provided illustrates a flowchart of a face region localization method according to an embodiment of this application. The face region localization method includes at least the following steps S110 to S130:

[0056] Step S110: When the user is wearing a wearable device and the beauty care device is initialized based on preset facial key points, the electromagnetic induction data corresponding to the preset facial key points is obtained.

[0057] The face region localization method in this application embodiment is mainly used to determine the face region currently being used by the beauty care device. Here, the "face region" can be a facial skin area divided according to the functional requirements of the beauty care device, such as... Figure 2The diagram illustrates a facial region division method according to an embodiment of this application. For example, it may include a forehead region, an eye region, and a cheek region. This division method specifically corresponds to five regions: the forehead region, the left eye region, the right eye region, the left cheek region, and the right cheek region. Of course, the facial regions can be further divided into more detailed regions. For example, the cheek region can be further divided into a jawline region and an apple cheek region, resulting in a seven-region division. Naturally, the specific method of facial region division can be flexibly defined by those skilled in the art according to actual needs, and no specific limitations are made here.

[0058] The more detailed the division of the face region, the higher the accuracy requirement for region positioning. The beauty care device positioning scheme of this application embodiment can support sufficiently fine-grained face region division, providing strong support for users' precise skin care by region.

[0059] The face region localization method of this application embodiment can be divided into two stages: initialization of the beauty care device and real-time localization. In the initialization stage, the user needs to wear the wearable device first, and then complete the initialization operation of the beauty care device according to the guidance or known predefined facial key points.

[0060] The aforementioned wearable devices can be considered as auxiliary positioning devices independent of beauty and personal care devices. They can be worn on the user's head, neck, or wrist in any form, such as headbands, hair clips, necklaces, chokers, bracelets, or wristbands. One of the wearable or beauty / personal care devices contains a magnetic source (such as an electromagnet) that generates a magnetic field, while the other contains a magnetic sensor, also known as a Hall effect sensor (or simply "Hall sensor"), which is a transducer that converts changes in the magnetic field into changes in the output voltage. Through this structure, real-time electromagnetic induction data can be generated when the beauty / personal care device enters the facial area covered by the wearable device.

[0061] The aforementioned preset facial landmarks can be flexibly defined according to actual needs, and may include facial features such as eyebrows, eyelids, lips, and nose. Compared to feature points in other areas of the face, the differences in features at facial landmarks are more pronounced among different users. By using multiple predefined facial landmarks, different users can be distinguished relatively accurately. During the initialization phase, the user needs to hold the beauty device and apply it sequentially to the positions of each preset facial landmark. At this time, electromagnetic induction data corresponding to each preset facial landmark can be obtained. Since the spatial pose relationship between the user's facial landmarks and the head-mounted device is basically fixed after the user wears the wearable device, the electromagnetic induction data corresponding to each preset facial landmark obtained during the initialization phase can be regarded as data that can characterize the user's facial features.

[0062] Step S120: When the user is using the beauty care device, acquire real-time electromagnetic induction data between the wearable device and the beauty care device.

[0063] After the user completes the initialization of the beauty device, they can begin to use the device for skincare. At this point, the device enters the stage of locating the facial area where it is applied, and the device can obtain real-time electromagnetic induction data between the wearable device and the beauty device.

[0064] Step S130: Based on the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data, the preset facial region positioning model is used to locate the facial region currently being used by the beauty care device.

[0065] This application embodiment takes into account the differences in the distribution of facial region features among different users, and pre-trains a facial region localization model. This facial region localization model can be trained based on electromagnetic induction data corresponding to the facial key points of different users and real-time electromagnetic induction data. The electromagnetic induction data corresponding to the facial key points can characterize the user's facial features, and therefore can be used to distinguish the user's identity. The real-time electromagnetic induction data can further determine which region of the user's face it corresponds to based on the user's identity, so that the model can ensure the accuracy of facial region localization for different users while distinguishing different user identities.

[0066] Therefore, based on the aforementioned preset face region positioning model, the electromagnetic induction data corresponding to the preset face key points obtained in the previous steps and the real-time electromagnetic induction data can be input into the preset face region positioning model to obtain the face region currently being acted upon by the beauty care device.

[0067] The face region localization method of this application takes into account the different face region distribution characteristics of different users. A face region localization model that can distinguish different users is trained in advance. By using this face region localization model, combined with the electromagnetic induction data of the user's facial key points determined in the initialization stage and the electromagnetic induction data generated in real time, the beauty care device can be accurately located in the user's face region, thereby improving the accuracy of face region localization.

[0068] In some embodiments of this application, the preset face region localization model includes a face key point recognition sub-model and a partition localization sub-model. The step of using the preset face region localization model to locate the face region currently being used by the beauty device based on the electromagnetic induction data corresponding to the preset face key points and real-time electromagnetic induction data includes: using the face key point recognition sub-model to identify the user identity based on the electromagnetic induction data corresponding to the preset face key points, obtaining the user identity corresponding to the preset face key points; obtaining the partition localization sub-model corresponding to the user identity; and using the partition localization sub-model to perform partition localization based on the real-time electromagnetic induction data, obtaining the face region currently being used by the beauty device.

[0069] The pre-trained preset face region localization model in this application embodiment can be divided into a face key point recognition sub-model and a region localization sub-model. The face key point recognition sub-model and the region localization sub-model can be regarded as two independent models that can be trained in parallel.

[0070] The facial landmark recognition model is primarily trained based on electromagnetic induction data of facial landmarks from different users. The aim is to train the model to recognize and differentiate the facial landmarks of different users, thus identifying the user's identity. The regional localization sub-model, on the other hand, is trained separately based on electromagnetic induction data for each user's facial region. This means each user has their own corresponding regional localization sub-model, thereby avoiding the impact of differences in feature distribution across different users' facial regions on the accuracy of regional localization.

[0071] During the face region localization stage, the electromagnetic induction data corresponding to the currently acquired preset facial key points can be input into the aforementioned facial key point recognition model to identify the facial key points and thus determine the user identity corresponding to those key points. After determining the user identity, the corresponding partition localization sub-model can be further invoked, and the real-time acquired electromagnetic induction data can be input into the partition localization sub-model to determine which area of ​​the user's face the beauty care device is currently acting on.

[0072] In some embodiments of this application, the step of using a preset face region positioning model to locate the face region currently being used by the beauty care device based on the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data includes: directly inputting the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data into the preset face region positioning model to obtain the face region currently being used by the beauty care device.

[0073] In addition to training the facial landmark recognition sub-model and the regional localization sub-model respectively based on the aforementioned embodiments to achieve the functions of user identification and regional localization, the embodiments of this application can also pre-train a comprehensive facial region localization model. This facial region localization model can be trained based on the electromagnetic induction data of facial landmarks of different users and the real-time electromagnetic induction data of different regions, thereby having the ability to accurately locate the facial regions of different users.

[0074] Based on the above face region localization model, the electromagnetic induction data corresponding to the currently acquired preset face key points and the real-time electromagnetic induction data can be directly input into the model, and the face region currently being acted upon by the beauty care device can be directly obtained from the model output.

[0075] It should be noted that both the above-described implementation methods of training the facial key point recognition sub-model and the regional localization sub-model separately, and the implementation method of training a comprehensive facial region localization model, can achieve accurate localization of the facial regions of different users. In practical applications, those skilled in the art can flexibly choose according to actual needs, and no specific limitations are made here.

[0076] The former approach has a more significant advantage in the model training phase because the training of the two sub-models is conducted independently, requiring less data acquisition compared to training a single model, and resulting in higher training efficiency. The latter approach has a more significant advantage in the model application phase. After acquiring the electromagnetic induction data corresponding to the preset facial key points and real-time electromagnetic induction data, both can be input into the comprehensive face region localization model to obtain the final face region localization result. This eliminates the need for prior user identification and subsequent region localization, and also avoids concern with the model's internal processing logic, simplifying the localization process and improving face region localization efficiency.

[0077] In some embodiments of this application, the preset face region localization model includes a face key point recognition sub-model and a partition localization sub-model. The preset face region localization model is trained in the following manner: acquiring magnetic induction data of preset face key points of multiple users; training the face key point recognition sub-model using the magnetic induction data of preset face key points of multiple users; acquiring magnetic induction data of face regions of multiple users; and training the partition localization sub-model corresponding to each user using the magnetic induction data of face regions of multiple users respectively.

[0078] As mentioned above, the embodiments of this application can train the facial landmark recognition sub-model and the regional localization sub-model separately, and the training processes of the two are independent of each other. For training the facial landmark recognition sub-model, a certain number of magnetic induction data corresponding to each preset facial landmark location of a user can be obtained first. These magnetic induction data corresponding to each preset facial landmark location of a user can be used as training sample data to train the facial landmark recognition sub-model's ability to recognize user identity.

[0079] For training the partitioned localization sub-model, since the facial region feature distributions of different users are different, the partitioned localization sub-models corresponding to different users can be trained separately. In order to reduce training costs and improve training efficiency, the embodiments of this application can also divide users into different user groups. The facial region feature distributions of users in different user groups are significantly different, while the facial region feature distributions of users within the same user group are less different. Therefore, the corresponding partitioned localization sub-models can be trained separately for different user groups. This method does not require that the user set data collected when training the facial key point recognition sub-model correspond completely or be consistent with the user set data collected when training the partitioned localization sub-model.

[0080] Of course, those skilled in the art can flexibly set the specific methods for dividing user groups according to actual needs. For example, they can cluster the magnetic induction data of different facial regions of a large number of users to group similar users into one user group. Alternatively, they can divide users into more detailed dimensions such as region, age, and gender. We will not go into further detail here.

[0081] The users in the training phase described above can be real users or simulated users based on deformable 3D facial models. The number of users can be flexibly adjusted according to the training effect of the model, and no specific limitation is made here. The facial key point recognition sub-model and the partition localization sub-model described above can both be regarded as multi-classification models, and both can be trained based on existing convolutional neural network structures. The specific form of convolutional neural network used can be flexibly selected by those skilled in the art according to actual needs, and no specific limitation is made here.

[0082] In some embodiments of this application, training a facial landmark recognition sub-model using magnetic induction data of preset facial landmarks of multiple users includes: constructing a facial landmark feature vector corresponding to each user based on the magnetic induction data of preset facial landmarks of multiple users; and training the facial landmark recognition sub-model using the facial landmark feature vector corresponding to each user.

[0083] In this embodiment of the application, when training a facial landmark recognition sub-model using magnetic induction data of preset facial landmarks from multiple users, the feature vector of each user's facial landmark can be constructed based on the magnetic induction data of each user's preset facial landmarks, and used as the input for model training.

[0084] Taking a triaxial transmitting coil as an example of a beauty device, the triaxial transmitting coil is composed of three mutually perpendicular circular electromagnetic coils, corresponding to the X, Y, and Z coordinate axes. Under the action of a sinusoidal voltage value, the triaxial transmitting coil generates a high-frequency oscillating magnetic field. This magnetic field is received in space by the magnetic sensor, i.e., the magnetic receiver, set in the wearable device. The magnetic receiver is also composed of three mutually perpendicular circular electromagnetic coils. The receiving coil and the transmitting coil have the same frequency. At an instant, each axis of the receiving coil can receive the magnetic field strength from the X, Y, and Z axis coils of the transmitting coil, respectively. This generates vector values ​​of the magnetic field strength of the transmitting coil at the X, Y, and Z axes of the receiving coil, respectively. This generates a 9-dimensional feature vector data, which can be regarded as a unique scalar of the spatial relative positional relationship between the receiving coil and the transmitting coil at this instant.

[0085] Based on this, when the beauty care device is applied to each preset facial key point, it will generate a corresponding 9-dimensional feature vector data. By fusing the 9-dimensional feature vector data of multiple preset facial key points corresponding to each user, unique feature vector data that can represent the facial features of each user can be obtained.

[0086] In some embodiments of this application, the step of training a sub-model for each user's corresponding partition using magnetic induction data of the facial regions of multiple users includes: constructing a first region feature vector for each partition of each user based on the magnetic induction data of the facial regions of multiple users; and training a sub-model for each user's corresponding partition using the first region feature vector for each partition of each user.

[0087] In this embodiment of the application, when training the corresponding partition localization sub-model for each user using the magnetic induction data of the face regions of multiple users, the magnetic induction data of each user in each face region can be collected first. The first region feature vector of each partition is constructed with the partition as the dimension. The first region feature vector can be regarded as a unique scalar representing the features of the face region.

[0088] For example, electromagnetic induction data can be collected at 100 points on a user's forehead. Each point's electromagnetic induction data consists of a 9-dimensional feature vector. These 100 9-dimensional feature vectors can be merged together as a unique scalar for the forehead region. The same principle applies to other regions. This yields the first region feature vector for each user's facial region. This vector can then be used to train a corresponding regional localization sub-model for each user, enabling the model to distinguish between different facial regions.

[0089] In some embodiments of this application, the magnetic induction data of the face region includes magnetic induction data of the left face region and magnetic induction data of the right face region. The partitioned localization sub-model includes a left face partitioned localization sub-model and a right face partitioned localization sub-model. The step of training the partitioned localization sub-model corresponding to each user using the magnetic induction data of the face regions of multiple users includes: training the left face partitioned localization sub-model corresponding to each user using the magnetic induction data of the left face region of each user; and training the right face partitioned localization sub-model corresponding to each user using the magnetic induction data of the right face region of each user.

[0090] To further improve the training effect and positioning accuracy of the partitioned localization sub-model, this embodiment of the application can further divide the user's face region into a left face region and a right face region, and train corresponding partitioned localization sub-models for the left face region and the right face region respectively.

[0091] For example, magnetic induction data of the left facial regions, such as the left forehead, left eye area, left cheek, and left jawline, can be collected first. Then, feature vectors corresponding to each left facial region can be constructed. These feature vectors are used as training samples for the left facial region localization sub-model, enabling it to distinguish between different left facial regions. The training logic for the right facial region localization sub-model is similar and will not be elaborated here.

[0092] In some embodiments of this application, the preset face region localization model is trained in the following manner: acquiring magnetic induction data of preset face key points and magnetic induction data of face regions of multiple users; constructing a second region feature vector corresponding to each user based on the magnetic induction data of preset face key points and magnetic induction data of face regions of multiple users; and training the preset face region localization model using the second region feature vector corresponding to each user.

[0093] As mentioned above, the preset face region localization model in this application embodiment can also be a comprehensively trained model, that is, it has the ability to distinguish different face regions of different users. When training this face region localization model, it is also necessary to obtain a certain number of magnetic induction data corresponding to each preset face key point for each user. In addition, it is also necessary to obtain the magnetic induction data of each user in each face region. Based on the magnetic induction data of each user corresponding to each preset face key point and the magnetic induction data in each face region, a second region feature vector corresponding to each user is constructed. That is, it is equivalent to fusing the feature vector of the input face key point recognition sub-model and the feature vector of the input partition localization sub-model in the aforementioned embodiment.

[0094] The aforementioned second region feature vector contains both the feature information of facial key points that can represent the user's identity and the feature information of different facial regions of the user. The feature vector obtained by fusing the two can serve as a unique scalar for different facial regions of different users.

[0095] Similarly, the users here can be real users or users simulated based on deformable 3D facial models. The number of users can be flexibly adjusted according to the training effect of the model, and no specific limit is made here.

[0096] To facilitate understanding of the various embodiments of this application, the face region localization process of the embodiments of this application can be divided into two stages: the face region localization model training stage and the face region localization model application stage.

[0097] like Figure 3 The diagram illustrates a face region localization process according to an embodiment of this application. During the face region localization model training phase, a face key point recognition sub-model and a partitioned localization sub-model can be trained separately. The face key point recognition sub-model is trained by collecting magnetic induction data of preset face key points from multiple users, and the partitioned localization sub-model for each user is trained by collecting magnetic induction data of the face regions from multiple users.

[0098] In the face region localization model application stage, electromagnetic induction data corresponding to the preset face key points collected in the initialization stage are first obtained. Then, real-time electromagnetic induction data between the wearable device and the beauty device are obtained. The electromagnetic induction data corresponding to the preset face key points collected in the initialization stage is converted into feature vectors and input into the face key point recognition sub-model to obtain the user corresponding to the face key points. Then, the partition localization sub-model corresponding to the user is called. The real-time electromagnetic induction data between the wearable device and the beauty device is input into the partition localization sub-model to obtain the localization result of the beauty device in the face region of the user.

[0099] like Figure 4As shown, another face region localization process is illustrated in this embodiment of the application. During the face region localization model training phase, a face region localization model is trained by comprehensively collecting magnetic induction data of preset facial key points and magnetic induction data of the face region from multiple users.

[0100] In the application stage of the face region localization model, the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data between the wearable device and the beauty device can be directly converted into feature vectors and then fused into the face region localization model, thereby directly obtaining the localization result of the beauty device in the current user's face region.

[0101] This application embodiment also provides a face region positioning device 500, such as Figure 5 The diagram shows a structural schematic of a face region positioning device according to an embodiment of this application. The face region positioning device 500 includes: a first acquisition unit 510, a second acquisition unit 520, and a positioning unit 530, wherein:

[0102] The first acquisition unit 510 is used to acquire electromagnetic induction data corresponding to the preset facial key points when the user wears a wearable device and initializes the beauty care device based on preset facial key points.

[0103] The second acquisition unit 520 is used to acquire real-time electromagnetic induction data between the wearable device and the beauty device when the user is using the beauty device.

[0104] The positioning unit 530 is used to locate the face area currently being used by the beauty care device by using a preset face area positioning model based on the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data.

[0105] In some embodiments of this application, the preset face region positioning model includes a face key point recognition sub-model and a partition positioning sub-model. The positioning unit 530 is specifically used for: using the face key point recognition sub-model to identify the user identity based on the electromagnetic induction data corresponding to the preset face key points, and obtaining the user identity corresponding to the preset face key points; obtaining the partition positioning sub-model corresponding to the user identity; and using the partition positioning sub-model to perform partition positioning based on the real-time electromagnetic induction data, and obtaining the face region currently being used by the beauty care device.

[0106] In some embodiments of this application, the positioning unit 530 is specifically used to: directly input the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data into the preset facial region positioning model to obtain the facial region currently being used by the beauty care device.

[0107] In some embodiments of this application, the preset face region localization model includes a face key point recognition sub-model and a partition localization sub-model. The preset face region localization model is trained in the following manner: acquiring magnetic induction data of preset face key points of multiple users; training the face key point recognition sub-model using the magnetic induction data of preset face key points of multiple users; acquiring magnetic induction data of face regions of multiple users; and training the partition localization sub-model corresponding to each user using the magnetic induction data of face regions of multiple users respectively.

[0108] In some embodiments of this application, training a facial landmark recognition sub-model using magnetic induction data of preset facial landmarks of multiple users includes: constructing a facial landmark feature vector corresponding to each user based on the magnetic induction data of preset facial landmarks of multiple users; and training the facial landmark recognition sub-model using the facial landmark feature vector corresponding to each user.

[0109] In some embodiments of this application, the step of training a sub-model for each user's corresponding partition using magnetic induction data of the facial regions of multiple users includes: constructing a first region feature vector for each partition of each user based on the magnetic induction data of the facial regions of multiple users; and training a sub-model for each user's corresponding partition using the first region feature vector for each partition of each user.

[0110] In some embodiments of this application, the magnetic induction data of the face region includes magnetic induction data of the left face region and magnetic induction data of the right face region. The partitioned localization sub-model includes a left face partitioned localization sub-model and a right face partitioned localization sub-model. The step of training the partitioned localization sub-model corresponding to each user using the magnetic induction data of the face regions of multiple users includes: training the left face partitioned localization sub-model corresponding to each user using the magnetic induction data of the left face region of each user; and training the right face partitioned localization sub-model corresponding to each user using the magnetic induction data of the right face region of each user.

[0111] In some embodiments of this application, the preset face region localization model is trained in the following manner: acquiring magnetic induction data of preset face key points and magnetic induction data of face regions of multiple users; constructing a second region feature vector corresponding to each user based on the magnetic induction data of preset face key points and magnetic induction data of face regions of multiple users; and training the preset face region localization model using the second region feature vector corresponding to each user.

[0112] It is understood that the above-mentioned face region positioning device can realize all the steps of the face region positioning method provided in the foregoing embodiments. The relevant explanations of the face region positioning method are applicable to the face region positioning device, and will not be repeated here.

[0113] This application also provides a wearable device, including: a processor; and a memory arranged to store computer-executable instructions, which, when executed, cause the processor to perform any of the aforementioned face region localization methods.

[0114] Figure 6 This is a schematic diagram of the structure of a wearable device according to an embodiment of this application. Please refer to it. Figure 6 At the hardware level, the wearable device includes a processor, and optionally also an internal bus, network interface, and memory. The memory may include RAM, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk drive. Of course, the wearable device may also include other hardware required for its business operations.

[0115] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 6 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0116] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0117] The processor reads the corresponding computer program from non-volatile memory into main memory and then runs it, forming a face region localization device at the logical level. The processor executes the program stored in memory and specifically performs the following operations:

[0118] When a user wears a wearable device and initializes the beauty care device based on preset facial key points, the electromagnetic induction data corresponding to the preset facial key points is obtained;

[0119] When a user is using a beauty device, real-time electromagnetic induction data between the wearable device and the beauty device is acquired.

[0120] Based on the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data, the preset facial region positioning model is used to locate the facial region currently being treated by the beauty care device.

[0121] The above is as stated in this application. Figure 1 The face region localization device method disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this application can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0122] The wearable device can also perform Figure 1 A method for implementing a face region localization device, and how the face region localization device is used in... Figure 1 The functions of the embodiments shown are not described in detail here.

[0123] This application also proposes a computer-readable storage medium that stores one or more programs, the programs including instructions that, when executed by a wearable device comprising multiple applications, enable the wearable device to perform... Figure 1 The method executed by the face region positioning device in the illustrated embodiment is specifically used to perform:

[0124] When a user wears a wearable device and initializes the beauty care device based on preset facial key points, the electromagnetic induction data corresponding to the preset facial key points is obtained;

[0125] When a user is using a beauty device, real-time electromagnetic induction data between the wearable device and the beauty device is acquired.

[0126] Based on the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data, the preset facial region positioning model is used to locate the facial region currently being treated by the beauty care device.

[0127] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0129] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0130] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0131] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0132] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0133] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0134] This application also provides a beauty care system, such as... Figure 7 As shown, a schematic diagram of the architecture of a beauty care system according to an embodiment of this application is provided. The beauty care system includes beauty care devices and the aforementioned wearable devices.

[0135] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0136] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0137] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for locating a face region, characterized in that, The face region localization method includes: When a user wears a wearable device and initializes the beauty care device based on preset facial key points, the electromagnetic induction data corresponding to the preset facial key points is obtained; When a user is using a beauty device, real-time electromagnetic induction data between the wearable device and the beauty device is acquired. Based on the electromagnetic induction data corresponding to the preset facial key points and the real-time electromagnetic induction data, the preset facial region positioning model is used to locate and obtain the facial region currently being acted upon by the beauty care device. The preset face region localization model includes a face key point recognition sub-model and a partition localization sub-model. The localization process, based on the electromagnetic induction data corresponding to the preset face key points and real-time electromagnetic induction data, utilizes the preset face region localization model to obtain the face region currently being used by the beauty care device, including: Based on the electromagnetic induction data corresponding to the preset facial key points, the user identity is identified using the facial key point recognition sub-model to obtain the user identity corresponding to the preset facial key points. Obtain the partition location sub-model corresponding to the user identity; Based on the real-time electromagnetic induction data, the facial region currently being treated by the beauty care device is obtained by using the partitioned positioning sub-model for partitioned positioning.

2. The face region localization method according to claim 1, characterized in that, The preset face region localization model is trained in the following way: Acquire magnetic induction data of preset facial key points from multiple users; A facial landmark recognition sub-model was trained using magnetic induction data of preset facial landmarks from multiple users. Acquire magnetic induction data of the facial regions of multiple users; The magnetic induction data of the facial regions of multiple users are used to train the corresponding partition localization sub-model for each user.

3. The face region localization method according to claim 2, characterized in that, The step of training a facial landmark recognition sub-model using magnetic induction data of preset facial landmarks from multiple users includes: Based on the magnetic induction data of preset facial key points of multiple users, construct the facial key point feature vector corresponding to each user; The facial landmark recognition sub-model is trained using the facial landmark feature vector corresponding to each user.

4. The face region localization method according to claim 2, characterized in that, The step of training a localization sub-model for each user using magnetic induction data of the facial regions of multiple users includes: Based on the magnetic induction data of the facial regions of multiple users, construct the first region feature vector for each partition of each user; Each user's corresponding partition localization sub-model is trained using the first region feature vector corresponding to each partition.

5. The face region localization method according to claim 2, characterized in that, The magnetic induction data of the face region includes magnetic induction data of the left face region and magnetic induction data of the right face region. The partition localization sub-model includes a left face partition localization sub-model and a right face partition localization sub-model. Training the partition localization sub-model corresponding to each user using the magnetic induction data of the face regions of multiple users includes: Each user's left face region localization sub-model is trained using the magnetic induction data of each user's left face region. Each user's right face region localization sub-model is trained using the magnetic induction data of the right face region.

6. The face region localization method according to claim 1, characterized in that, The preset face region localization model is trained in the following way: Acquire magnetic induction data of preset facial key points and magnetic induction data of facial regions from multiple users; Based on the magnetic induction data of preset facial key points and facial regions of multiple users, a second region feature vector is constructed for each user. The preset face region localization model is trained using the second region feature vector corresponding to each user.

7. A face region positioning device, characterized in that, The face region positioning device includes: The first acquisition unit is used to acquire electromagnetic induction data corresponding to the preset facial key points when the user is wearing a wearable device and the beauty care device is initialized based on preset facial key points. The second acquisition unit is used to acquire real-time electromagnetic induction data between the wearable device and the beauty device when the user is using the beauty device. The positioning unit is used to locate the face area currently being treated by the beauty care device by using a preset face area positioning model based on the electromagnetic induction data corresponding to the preset facial key points and real-time electromagnetic induction data. The preset face region localization model includes a face key point recognition sub-model and a region localization sub-model, and the localization unit is specifically used for: Based on the electromagnetic induction data corresponding to the preset facial key points, the user identity is identified using the facial key point recognition sub-model to obtain the user identity corresponding to the preset facial key points. Obtain the partition location sub-model corresponding to the user identity; Based on the real-time electromagnetic induction data, the facial region currently being treated by the beauty care device is obtained by using the partitioned positioning sub-model for partitioned positioning.

8. A wearable device, comprising: processor; as well as A memory configured to store computer-executable instructions, which, when executed, cause the processor to perform any of the face region localization methods according to claims 1 to 6.

9. A beauty care system, characterized in that, The beauty care system includes beauty care devices and wearable devices as described in claim 8.

Images