Facial recognition methods, devices and electronic equipment

Abstract

This invention provides a facial recognition method, apparatus, and electronic device. The method involves acquiring a target image and the age parameter of a face to be recognized within the target image; extracting target facial features from the target image; and determining whether the face to be recognized matches the target identity based on the age parameter, the target facial features, and the feature classification results corresponding to the target identity. The feature classification results include standard facial features for multiple age groups corresponding to the target identity. This approach takes into account that a person's appearance changes with age. Therefore, during facial recognition, the facial features of the face to be recognized are matched with standard facial features that match its age parameter, thereby improving the accuracy of facial recognition.

Description

Technical Field

[0001] This invention relates to the field of facial recognition technology, and more specifically, to a facial identity recognition method, device, and electronic device. Background Technology

[0002] In related technologies, when performing facial recognition on photos in an album, face detection methods are typically used to detect the coordinates of facial key points in the photo. Deep learning algorithms are then used to extract and adjust facial recognition vectors that identify the individual's identity. Finally, based on the similarity between the extracted recognition vectors calculated using Euclidean distance and cosine similarity and the target user's pre-determined recognition vectors, it is determined whether the two faces corresponding to the two recognition vectors belong to the same person. However, as people age, their appearance often changes significantly. When it is necessary to identify whether two faces with a large age gap belong to the same person, the accuracy of the above facial recognition methods is relatively low. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a facial recognition method, device and electronic device to improve the accuracy of facial recognition.

[0004] In a first aspect, embodiments of the present invention provide a face recognition method, the method comprising: acquiring a target image and an age parameter of a face to be recognized in the target image; extracting target face features of the face to be recognized from the target image; determining whether the face to be recognized matches the target identity based on the age parameter, the target face features and the feature classification result corresponding to the target identity; wherein the feature classification result includes: standard face features of multiple age groups corresponding to the target identity.

[0005] The steps described above for obtaining the target image and the age parameters of the face to be identified in the target image include: obtaining the target image and the shooting time of the target image; and calculating the age parameters of the face to be identified in the target image based on the pre-acquired birthday information of the target identity and the shooting time of the target image.

[0006] The aforementioned standard facial features are determined as follows: For each age group, multiple images containing faces corresponding to the target identity are acquired; the number of images matches a preset number; for each image, initial facial features are extracted; the initial facial features of multiple images are clustered to obtain clustering results; the clustering results include at least one set of facial feature combinations; the facial feature combination includes a central cluster feature and at least one initial facial feature; the central cluster feature is determined as the standard facial feature.

[0007] The steps described above for acquiring multiple images containing faces corresponding to the target identity and age groups include: acquiring images containing faces corresponding to the target identity and the image capture time; determining the age parameter corresponding to the image based on the pre-acquired birthday information of the target identity and the image capture time; determining whether the age parameter corresponding to the image is within the age group; if so, determining the acquired image as an image containing faces corresponding to the target identity; determining whether the number of acquired images containing faces corresponding to the target identity and age groups is equal to a preset number; if not, continuing to execute the steps of acquiring images containing faces corresponding to the target identity and the image capture time until the number of acquired images containing faces corresponding to the target identity is equal to the preset number.

[0008] The steps described above for extracting target facial features of a face to be identified from a target image include: performing face recognition processing on the target image to obtain key points of the face to be identified; performing pose adjustment processing on the face to be identified based on the key points of the face to be identified; and performing feature extraction processing on the adjusted face to be identified using a pre-trained face recognition model to obtain the target facial features of the face to be identified.

[0009] The steps described above, which determine whether the face to be identified matches the target identity based on the age parameter, the target facial features, and the feature classification results corresponding to the target identity, include: determining the standard facial features of the age group that matches the age parameter from the standard facial features of multiple age groups corresponding to the target identity as reference standard facial features; matching the target facial features with the reference standard facial features to obtain a matching result; and determining whether the face to be identified matches the target identity based on the matching result.

[0010] The aforementioned reference standard face features are obtained by clustering initial face features from multiple images, including the face corresponding to the target identity; the faces in the images correspond to age groups that match the age parameter; the reference standard face features include the central cluster feature of a set of face feature combinations obtained through clustering; the face feature combination also includes multiple initial face features; the matching result includes whether the target face feature matches the reference standard face feature, or whether the target face feature does not match the reference standard face feature; the step of matching the target face feature with the reference standard face feature to obtain the matching result includes: calculating a first distance between the target face feature and the reference standard face feature; determining whether the first distance is greater than the cluster distance of the face feature combination in which the reference standard face feature is located; the cluster distance includes the farthest distance among the distances between the central cluster feature of the face feature combination and multiple initial face features in the face feature combination; if not, determining that the target face feature matches the reference standard face feature.

[0011] The aforementioned reference standard face features are obtained by clustering initial face features from multiple images, including the face corresponding to the target identity; the faces in the images correspond to age groups that match the age parameter; the reference standard face features include the central cluster feature of multiple face feature combinations obtained by clustering; the face feature combination also includes at least one initial face feature; among the multiple face feature combinations, the face feature combination that includes the most initial face features is the target feature combination for the target identity in the age group that matches the age parameter; the matching result includes matching or not matching; the step of matching the target face feature with the reference standard face features to obtain the matching result includes: for each reference standard face feature, calculating the second distance between the target face feature and the reference standard face feature; taking the shortest second distance among multiple second distances as the target distance of the target face feature; determining whether the reference standard face feature corresponding to the target distance is located in the target feature combination; if so, determining that the target face feature matches the reference standard face feature.

[0012] The matching results include whether the target facial features match the reference standard facial features or not. Based on the matching results, the step of determining whether the face to be identified matches the target identity includes: if the target facial features match the reference standard facial features, determining that the face to be identified matches the target identity; if the target facial features do not match the reference standard facial features, determining that the face to be identified does not match the target identity.

[0013] Secondly, embodiments of the present invention provide a face recognition device, which includes: an image acquisition module for acquiring a target image and an age parameter of a face to be recognized in the target image; a feature extraction module for extracting target face features of the face to be recognized from the target image; and a matching module for determining whether the face to be recognized matches the target identity based on the age parameter, the target face features, and the feature classification result corresponding to the target identity; wherein the feature classification result includes standard face features of multiple age groups corresponding to the target identity.

[0014] Thirdly, embodiments of the present invention provide an electronic device, including a processor and a memory, wherein the memory stores machine-executable instructions that can be executed by the processor, and the processor executes the machine-executable instructions to implement the above-described face recognition method.

[0015] Fourthly, embodiments of the present invention provide a machine-readable storage medium storing machine-executable instructions. When the machine-executable instructions are invoked and executed by a processor, the machine-executable instructions cause the processor to implement the aforementioned face recognition method.

[0016] The embodiments of the present invention bring the following beneficial effects:

[0017] The aforementioned facial recognition method, apparatus, and electronic device acquire a target image and the age parameter of the face to be recognized in the target image; extract target facial features of the face to be recognized from the target image; and determine whether the face to be recognized matches the target identity based on the age parameter, the target facial features, and the feature classification result corresponding to the target identity. The feature classification result includes standard facial features for multiple age groups corresponding to the target identity. This method takes into account that a person's appearance changes with age; therefore, during facial recognition, it matches the facial features of the face to be recognized with standard facial features that match its age parameter, thereby improving the accuracy of facial recognition.

[0018] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 A flowchart of a face recognition method provided in an embodiment of the present invention;

[0022] Figure 2 A flowchart of a face recognition method provided in an embodiment of the present invention;

[0023] Figure 3 A flowchart illustrating a digital image preprocessing method provided in an embodiment of the present invention;

[0024] Figure 4 A flowchart of a digital image face recognition method provided in an embodiment of the present invention;

[0025] Figure 5 This is a schematic diagram of the structure of a facial recognition device provided in an embodiment of the present invention;

[0026] Figure 6This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Facial recognition technology generally refers to biometric identification technology that identifies or matches individuals based on given facial images. It has wide applications in many fields such as criminal investigation and identity authentication. The main problem that facial recognition technology solves is the automatic association between facial image data and the identity of the person.

[0029] With the deepening research in related fields in recent years, facial recognition technology has gradually formed an implementation scheme with data collection, face detection, pose adjustment, feature extraction, and identity determination as the main steps. It has achieved high recognition accuracy on public datasets such as LFW (Labled Faces in the Wild) and CFP. The application scenarios of general facial recognition technology are mainly divided into two types: one is face verification: comparing two input faces to determine whether they belong to the same person; the other is face retrieval: given a face and a certain size face database, matching the identity of the given face from the database.

[0030] The aforementioned mature facial recognition technologies primarily rely on a certain amount of labeled data or require the detected face to be in a relatively fixed pose. In the application scenario of facial recognition based on album photos, the labeled information often only exists at the album and photo level, while a single photo may contain multiple unlabeled faces in relatively free poses. Existing facial recognition technologies can only identify individuals by comparing scattered facial samples, resulting in relatively low accuracy in this application scenario. Furthermore, currently available facial recognition datasets lack labeled infant facial samples, making some machine learning-based facial recognition technologies prone to misjudgments in this age group.

[0031] In facial recognition applications using photo albums, for unlabeled facial data samples with diverse poses and ages, the main technology used is deep learning-based facial verification. For example... Figure 1As shown, firstly, a face detection method is used to detect the position of the face and the coordinates of some facial key points in the photo, so as to extract the face to be identified from the photo. Then, based on the detected facial key point coordinates, the face in the photo is transformed to a fixed position, thereby achieving the effect of pose adjustment. Next, the adjusted face is input into a feature extractor, and a deep learning algorithm is used to convert the face image into a recognition vector that identifies the identity features. Finally, the extracted recognition vectors are used to obtain the similarity between the two recognition vectors through methods such as Euclidean distance and cosine similarity, thereby determining whether the two faces belong to the same person.

[0032] However, when the time span of the album photos is large, the appearance of the same person can change significantly. Relying solely on facial recognition algorithms and facial image data makes it difficult to establish cross-age associations for the same person. Existing publicly available facial recognition datasets lack labeled infant facial samples, making it difficult for deep learning-based feature extraction methods to distinguish between infants of different ages, thus affecting the accuracy of facial recognition. Furthermore, the aforementioned methods have high requirements for external factors such as facial pose and lighting conditions, which are often difficult to meet in the context of album photos. Using only distance-based facial verification methods can easily lead to misidentification. Therefore, this invention provides a facial identity recognition method, device, and electronic device that can be applied to various album scenarios or facial recognition scenarios with a large age span.

[0033] This invention provides a facial recognition method, such as... Figure 2 As shown, the method includes the following steps:

[0034] Step S202: Obtain the target image and the age parameter of the face to be identified in the target image.

[0035] The target image mentioned above can be an image captured by image acquisition devices such as mobile phones, cameras, and monitoring devices. Typically, when a target image is captured, the capture time is also stored as a parameter in the device that captured the image.

[0036] The target image displays a face to be identified. The age parameter of the face to be identified is usually related to the time the target image was captured and the date of birth of the user corresponding to the face. Therefore, to determine whether the face to be identified matches the target identity, it is usually necessary to obtain the target identity's birthdate information, which can be provided by the user corresponding to the target identity.

[0037] When the birthday information corresponding to the target identity cannot be obtained, feature extraction can be performed on the face to be identified, and the age parameters of the face to be identified can be predicted based on the extracted features. Specifically, a model based on neural network or deep learning theory can be trained first on a large number of sample images including faces and their corresponding age parameters to obtain an age prediction model. Then, the age prediction model is used to predict the age parameters of the face to be identified to obtain the age parameters of the face to be identified.

[0038] Step S204: Extract the target face features of the face to be identified from the target image.

[0039] During feature extraction, the target region containing the face to be identified in the target image can be detected first. Then, feature extraction processing is performed on the face to be identified in the target region. Specifically, a pre-trained face recognition model can be used as input to the target region, and the target region can be processed by this model to obtain the target face features.

[0040] Since the face to be identified in the target image may be at angles unfavorable for feature extraction, such as side view, top view, or bottom view, the pose of the face in the target region can be adjusted before being input into the face recognition model. For example, the positions of facial key points in the face to be identified can be detected. These key points typically include the facial contour, eyes, eyebrows, lips, and nose contour. Then, based on the positions of the facial key points, the current pose of the face to be identified is determined, and further, the face to be identified is adjusted from the current pose to the target pose, such as a level view, using perspective relationships. The above-mentioned face pose adjustment process can also be implemented by adding corresponding modules to the face recognition model and training it.

[0041] Step S206: Based on the age parameter, target facial features, and feature classification results corresponding to the target identity, determine whether the face to be identified matches the target identity; wherein, the feature classification results include: standard facial features of multiple age groups corresponding to the target identity.

[0042] The standard facial features corresponding to the age group of the aforementioned target identity can be obtained by extracting features from multiple images including the face corresponding to the target identity, and then processing the extracted features. The aforementioned target identity can be the user who took the target image, or the target person who needs to be matched with the face to be identified, and can be specifically represented by a corresponding user identifier.

[0043] The process of feature processing on multiple images, including the face corresponding to the target identity, can usually be achieved in the same way as the method described above for extracting the target face features from the target image, and will not be elaborated here. The face features obtained through feature extraction can be called the initial face features.

[0044] Age parameters for multiple images of the face corresponding to the target identity need to be provided in advance. This can be done by pre-labeling the faces in these images for the user, or by calculating the time difference between the target identity's birthdate and the image's capture time. Since a person's appearance typically changes significantly with age, especially during infancy when changes are rapid and slowing down with age, age groups can be defined based on this characteristic, such as 0-6 months, 6-12 months, 12-36 months, 3-6 years, and 6-10 years. Based on the age group of the images corresponding to each initial facial feature, the initial facial features are assigned to the corresponding age-group facial vector set.

[0045] To ensure the accuracy of the standard facial features obtained from the set of facial vectors for each age group, if the number of initial facial features in the set is not less than a preset threshold, the initial facial features can be processed to determine one or more standard facial features. Each facial feature is typically represented by a feature vector. For example, the feature vectors corresponding to all initial facial features in the set can be averaged, and the averaged feature vector can be used as the feature vector corresponding to the standard facial features.

[0046] In the process of obtaining standard facial features, the initial facial features in the set can be clustered to obtain clustering results. Clustering typically involves automatically summarizing and classifying large amounts of unlabeled data based on a defined similarity metric and clustering process. The clustering results usually include a cluster center and one or more data points belonging to that cluster center used for clustering. The cluster center usually shares the same attributes as the data used for clustering. When clustering initial facial features, the resulting cluster center is also considered a facial feature. Since the cluster center best matches the category corresponding to the clustering result, i.e., the face corresponding to the target identity, it can be identified as a standard facial feature. Clustering multiple initial facial features may yield one or more clustering results; therefore, an age group may include one or more standard facial features.

[0047] After obtaining the target facial features of the person to be identified in the target image, the first step is to determine the age range of the target image based on the age parameter corresponding to the face to be identified, thereby determining the standard facial features to be matched. When there is only one standard facial feature, it can be determined whether the target facial feature falls within the clustering region generated by the clustering results corresponding to the standard facial feature. If the target facial feature falls within this region, the face to be identified can be considered to match the target identity; otherwise, there is no match. When there are multiple standard facial features, the feature distance between the target facial feature and each standard facial feature can be calculated. Then, the feature distance is used to determine whether the face to be identified matches the target identity. For example, if the sum of the feature distances is less than a preset distance threshold, the face to be identified is considered to match the target identity; otherwise, there is no match. The specific judgment method can be determined according to the requirements, or it can be set through multiple experiments based on the accuracy of the experimental results.

[0048] The aforementioned facial recognition method acquires a target image and the age parameter of the face to be recognized within the target image; extracts target facial features from the target image; and determines whether the face to be recognized matches the target identity based on the age parameter, target facial features, and feature classification results corresponding to the target identity. The feature classification results include standard facial features for multiple age groups corresponding to the target identity. This method takes into account that a person's appearance changes with age; therefore, during facial recognition, it matches the facial features of the face to be recognized with standard facial features that match its age parameter, thereby improving the accuracy of facial recognition.

[0049] The following embodiments provide an implementation method for acquiring a target image and the age parameter of the face to be identified in the target image.

[0050] The age parameter mentioned above can be manually labeled. For example, if a user takes a picture of the target image on their 20th birthday, they can add the tag "20 years old" when uploading the image; this tag serves as the age parameter for the face to be identified. Alternatively, the age parameter can be determined based on the shooting time of the target image and the target's birthday. Specifically, the shooting time of the target image can be obtained simultaneously with its acquisition. Then, based on the pre-acquired birthday information of the target's identity and the shooting time of the target image, the age parameter of the face to be identified in the target image can be calculated. The birthday information of the target's identity can be provided by the user; the time difference between the shooting time of the target image and the birthday information can be used as the age parameter for the face to be identified in the target image. For example, if the user was born on June 15, 2000, and the target image was taken on June 15, 2020, then the age parameter of the face to be identified would be 20 years old.

[0051] The following embodiments provide an implementation method for determining standard facial features.

[0052] The above-mentioned standard facial features can be determined in the following ways:

[0053] (1) For each age group, acquire multiple images of the face corresponding to the target identity; the number of images matches the preset number.

[0054] Since the images need to be associated with age groups, it's necessary to acquire images containing faces corresponding to the target identity, along with the image capture time. Then, based on the pre-acquired birthday information of the target identity and the image capture time, the age parameter corresponding to the image is determined. It's then determined whether the image's age parameter falls within the age group. For example, if the image's age parameter is 20 years old, and the age group is 15-25 years old, then the image's age parameter is considered to be within that age group, and the acquired image can be identified as containing a face corresponding to the target identity. To obtain a preset number of images, it's necessary to determine whether the number of acquired images containing faces corresponding to the target identity and age groups equals the preset number. If not, continue acquiring images containing faces corresponding to the target identity and determining whether each image corresponds to that age group, until the number of acquired images containing faces corresponding to the target identity equals the preset number. Having enough images and obtaining enough initial facial features is beneficial for subsequently obtaining standard facial features that more closely correspond to the target identity's face.

[0055] (2) For each image, extract the initial facial features of the face in the image.

[0056] Specifically, face recognition processing can be performed on the image to obtain key points of the face in the image; then, based on the key points of the face to be recognized, the pose of the face to be recognized can be adjusted; finally, a pre-trained face recognition model can be used to extract features from the adjusted face to obtain the initial facial features of the face in the image. The above method can also be used when extracting the target facial features of the face to be recognized from the target image.

[0057] (3) Cluster the initial facial features of multiple images to obtain clustering results; the clustering results include at least one set of facial feature combinations; the facial feature combinations include the central cluster feature and at least one initial facial feature.

[0058] The clustering process described above can be achieved using clustering algorithms. Common clustering algorithms include K-Means clustering and hierarchical clustering. Specifically, the appropriate clustering algorithm can be selected to generate clustering results based on the application scenario and clustering requirements.

[0059] (4) The central cluster feature is determined as the standard face feature. Since the central cluster feature is also called the "cluster center", it can represent the face feature that best matches the category corresponding to the clustering result. Therefore, the central cluster feature can be determined as the standard face feature.

[0060] The following embodiments provide an implementation method for determining whether a face to be identified matches a target identity based on age parameters, target facial features, and feature classification results corresponding to the target identity.

[0061] Because a person's appearance changes significantly with age, the standard facial features for each age group also differ considerably. Therefore, the first step is to identify the standard facial features of the age group that match the age parameter from the standard facial features of multiple age groups corresponding to the target identity as reference standard facial features. Then, the target facial features are matched with the reference standard facial features to obtain the matching results. Based on the matching results, it is determined whether the face to be identified matches the target identity, thereby reducing misjudgments in the facial recognition process.

[0062] The reference standard face features obtained by clustering initial face features from multiple images, including faces corresponding to the target identity, may be one or more. During clustering, faces in the images are associated with age groups that match the age parameter.

[0063] When the reference standard facial features include the central cluster feature of a set of facial feature combinations obtained through clustering, and the facial feature combinations also include multiple initial facial features, the target facial features can be matched with the reference standard facial features as follows: First, calculate the first distance between the target facial features and the reference standard facial features; then determine whether the first distance is greater than the cluster distance of the facial feature combination to which the reference standard facial features belong; wherein, the cluster distance can be the farthest distance among the distances between the central cluster feature of the facial feature combination and the multiple initial facial features in the facial feature combination; if not, determine that the target facial features match the reference standard facial features. When the first distance is less than the cluster distance, it can be considered that the target facial features belong to the category of the clustering result, that is, the target facial features correspond to the target identity.

[0064] When the reference standard face features include the central cluster feature of multiple face feature combinations obtained through clustering processing, and the face feature combination also includes at least one initial face feature, the face feature combination with the largest number of initial face features among the multiple face feature combinations can be determined as the target feature combination for the target identity in the age range that matches the age parameter. Specifically, the target face features can be matched with the reference standard face features in the following way: For each reference standard face feature, calculate the second distance between the target face feature and the reference standard face feature; take the shortest second distance among the multiple second distances as the target distance of the target face feature; determine whether the reference standard face feature corresponding to the target distance is located in the target feature combination; if so, determine that the target face feature matches the reference standard face feature.

[0065] If the target facial features match the reference standard facial features, the face to be identified is determined to match the target identity; if the target facial features do not match the reference standard facial features, the face to be identified is determined to not match the target identity.

[0066] This invention also provides another facial recognition method, which in... Figure 2 This method is implemented based on the method shown. The method mainly solves the following two problems: (1) When the time span of the album photos is large, the appearance of the same person changes greatly. It is difficult to establish cross-age association of the same person based solely on face recognition algorithms and face image data; (2) Existing publicly available face recognition datasets lack infant face samples with identity labels, which makes it difficult for deep learning-based feature extraction methods to distinguish infant faces with different identities, affecting the accuracy of face recognition.

[0067] In implementing this technology, users need to provide their own digital images, the time the images were taken, and their personal birthdate. The technology then preprocesses and learns from a batch of real digital images provided by the user before it can recognize and predict subsequent digital images. If the digital image information is not authentic, accurate predictions are impossible.

[0068] This method mainly consists of two parts: one is the preprocessing of the user's digital image, such as... Figure 3 As shown; the second is the facial recognition process of the user's digital image, such as Figure 4 As shown.

[0069] First, the preprocessing workflow for the user's digital images will be described in detail:

[0070] (1) Data processing is performed through module 1 on the client side. The client provides users with a process for setting personal information and records the user's birthday. After setting their birthday, users can submit their digital image. During the upload process, the APP (Application) client extracts the shooting time of the digital image from the associated information of the digital image and uploads it to the server interface along with the digital image.

[0071] (2) User-uploaded digital images are grouped by age using module 2 in the server-side application running on the server. Specifically, after the server obtains the user-uploaded digital images, it performs time-segment clustering based on the user's birthday and the image's shooting date to obtain machine learning samples of digital images for different age groups. For example, 0-6 months is one age group, 7-12 months is another, and so on. The specific age groups can be adjusted according to actual needs. After the digital images are divided by age group, each age group can obtain a set of digital images for the corresponding time period. Here, we take the 0-6 month age group as an example. When the number of digital images in the 0-6 month age group reaches the set threshold, such as 30 digital images (0-6 month digital image group), the digital images are submitted to the face detection module for face detection processing.

[0072] (3) Module 3 performs face detection on a set of digital images for the current age group (also known as "the current batch of digital images"). Each digital image undergoes face recognition processing, marking key facial points and adjusting facial poses, generating a standardized face vector for each face. After the server calculates all face vectors for the current batch (e.g., digital images from 0-6 months), it clusters the faces to obtain the clustering results for all faces in the current batch. The clustering results include multiple face vector groups. The faces corresponding to the face vectors in each face vector group can be considered to correspond to the same person. The faces in the face vector group containing the most face vectors can be labeled as the user's own face. This completes the preprocessing of the user's digital images, enabling learning and clustering of the user's 0-6 month sample database. Similarly, digital images for other age groups of the user can be processed in the same way to obtain clustering results for each age group of the user.

[0073] After preprocessing the user's digital images and completing the learning and clustering of sample digital images of users of different age groups, facial recognition can then be performed on new digital images uploaded by the user, specifically through the following methods:

[0074] (1) Receive a new digital image uploaded by the user through the client. The information of the digital image includes the time when the digital image was captured and the user information to be identified (generally the identifier of the user to be identified, such as ID).

[0075] (2) Obtain the user's information through the server, such as the birthday of the person to be identified recorded in the preprocessing process. Calculate the age range of the digital image based on the birthday of the person to be identified and the time the photo was taken, thereby obtaining the preprocessing clustering results corresponding to the age range of the person to be identified (equivalent to the above "reference standard face features").

[0076] (3) The face vector information of the digital image is calculated by the server. Specifically, the face image data can be processed to determine the key points of the face, and then the face pose can be adjusted based on the key points of the face to finally obtain the standardized face vector (equivalent to the above "target face features").

[0077] (4) Using the face vector information of the new digital image and the clustering results of the preprocessed age group, the face in the digital image is predicted to obtain the prediction result. If the face vector information of the new digital image falls into the category with the most clusters in that age group, the face in the new digital image can be considered to be the person to be identified (also known as the "person to be predicted"); otherwise, it is considered to be someone else.

[0078] For the above method embodiments, see Figure 5 A facial recognition device is shown, the device comprising:

[0079] The image acquisition module 502 is used to acquire the target image and the age parameters of the face to be identified in the target image;

[0080] Feature extraction module 504 is used to extract target facial features of the face to be identified from the target image;

[0081] The matching module 506 is used to determine whether the face to be identified matches the target identity based on the age parameter, the target face features and the feature classification results corresponding to the target identity; wherein, the feature classification results include: standard face features of multiple age groups corresponding to the target identity.

[0082] The aforementioned facial recognition device acquires a target image and the age parameter of the face to be recognized in the target image; extracts target facial features of the face to be recognized from the target image; and determines whether the face to be recognized matches the target identity based on the age parameter, the target facial features, and the feature classification result corresponding to the target identity. The feature classification result includes standard facial features for multiple age groups corresponding to the target identity. This method takes into account that a person's appearance changes with age; therefore, during facial recognition, it matches the facial features of the face to be recognized with standard facial features that match its age parameter, thereby improving the accuracy of facial recognition.

[0083] The image acquisition module described above is also used to: acquire the target image and the shooting time of the target image; and calculate the age parameter of the face to be identified in the target image based on the pre-acquired birthday information of the target identity and the shooting time of the target image.

[0084] The aforementioned device further includes a standard facial feature determination module, used for: acquiring multiple images of faces corresponding to the target identity for each age group; the number of images matching a preset number; extracting initial facial features of the faces in each image; performing clustering processing on the initial facial features of multiple images to obtain clustering results; the clustering results including at least one set of facial feature combinations; the facial feature combination including a central clustering feature and at least one initial facial feature; and determining the central clustering feature as a standard facial feature.

[0085] The aforementioned standard facial feature determination module is also used for: acquiring an image containing a face corresponding to the target identity and the image's capture time; determining the age parameter corresponding to the image based on the pre-acquired birthday information of the target identity and the image's capture time; determining whether the age parameter corresponding to the image is within an age range; if so, determining the acquired image as an image containing a face corresponding to the target identity; determining whether the number of acquired images containing a face corresponding to the target identity and the age range is equal to a preset number; if not, continuing to execute the steps of acquiring an image containing a face corresponding to the target identity and the image's capture time until the number of acquired images containing a face corresponding to the target identity is equal to the preset number.

[0086] The aforementioned feature extraction module is also used for: performing face recognition processing on the target image to obtain key points of the face to be recognized; performing pose adjustment processing on the face to be recognized based on the key points of the face to be recognized; and performing feature extraction processing on the adjusted face to be recognized through a pre-trained face recognition model to obtain the target face features of the face to be recognized.

[0087] The matching module includes: a reference standard face feature determination unit, used to determine the standard face features of the age group that match the age parameter from the standard face features of multiple age groups corresponding to the target identity as the reference standard face features; a feature matching unit, used to match the target face features with the reference standard face features to obtain the matching result; and an identity matching unit, used to determine whether the face to be identified matches the target identity based on the matching result.

[0088] The aforementioned reference standard face features are obtained by clustering initial face features from multiple images, including the face corresponding to the target identity; the faces in the images correspond to age groups that match the age parameter; the reference standard face features include the central cluster feature of a set of face feature combinations obtained through clustering; the face feature combination also includes multiple initial face features; the matching result includes whether the target face feature matches the reference standard face feature, or whether the target face feature does not match the reference standard face feature; the aforementioned feature matching unit is further used to: calculate a first distance between the target face feature and the reference standard face feature; determine whether the first distance is greater than the cluster distance of the face feature combination in which the reference standard face feature is located; the cluster distance includes the farthest distance among the distances between the central cluster feature of the face feature combination and multiple initial face features in the face feature combination; if not, determine that the target face feature matches the reference standard face feature.

[0089] The aforementioned reference standard face features are obtained by clustering initial face features from multiple images, including the face corresponding to the target identity; the faces in the images correspond to age groups that match the age parameter; the reference standard face features include the central cluster feature of multiple face feature combinations obtained by clustering; the face feature combination also includes at least one initial face feature; among the multiple face feature combinations, the face feature combination that includes the most initial face features is the target feature combination for the target identity in the age group that matches the age parameter; the matching result includes matching or not matching; the aforementioned feature matching unit is further used to: calculate a second distance between the target face feature and the reference standard face feature for each reference standard face feature; take the shortest second distance among the multiple second distances as the target distance of the target face feature; determine whether the reference standard face feature corresponding to the target distance is located in the target feature combination; if so, determine that the target face feature matches the reference standard face feature.

[0090] The matching results include whether the target face features match the reference standard face features or whether the target face features do not match the reference standard face features. The identity matching unit is also used to: determine that the face to be identified matches the target identity if the target face features match the reference standard face features; and determine that the face to be identified does not match the target identity if the target face features do not match the reference standard face features.

[0091] This embodiment also provides an electronic device, including a processor and a memory. The memory stores machine-executable instructions that can be executed by the processor. The processor executes the machine-executable instructions to implement the above-described face recognition method, for example:

[0092] Acquire the target image and the age parameter of the face to be identified in the target image; extract the target face features of the face to be identified from the target image; determine whether the face to be identified matches the target identity based on the age parameter, the target face features and the feature classification results corresponding to the target identity; wherein, the feature classification results include: standard face features of multiple age groups corresponding to the target identity.

[0093] The above method takes into account that a person's appearance changes with age. Therefore, during the face recognition process, the facial features of the face to be identified are matched with standard facial features that match its age parameters, thereby improving the accuracy of face recognition.

[0094] Optionally, the steps of obtaining the target image and the age parameter of the face to be identified in the target image include: obtaining the target image and the shooting time of the target image; and calculating the age parameter of the face to be identified in the target image based on the pre-acquired birthday information of the target identity and the shooting time of the target image.

[0095] Optionally, the aforementioned standard facial features are determined as follows: for each age group, multiple images containing faces corresponding to the target identity are acquired; the number of images matches a preset number; for each image, initial facial features are extracted from the face in the image; the initial facial features of multiple images are clustered to obtain clustering results; the clustering results include at least one set of facial feature combinations; the facial feature combination includes a central clustering feature and at least one initial facial feature; the central clustering feature is determined as the standard facial feature.

[0096] Optionally, the above step of acquiring multiple images containing faces corresponding to the target identity and age group includes: acquiring images containing faces corresponding to the target identity and the image capture time; determining the age parameter corresponding to the image based on the pre-acquired birthday information of the target identity and the image capture time; determining whether the age parameter corresponding to the image is within the age group; if so, determining the acquired image as an image containing faces corresponding to the target identity; determining whether the number of acquired images containing faces corresponding to the target identity and age group is equal to a preset number; if not, continuing to execute the steps of acquiring images containing faces corresponding to the target identity and the image capture time until the number of acquired images containing faces corresponding to the target identity is equal to the preset number.

[0097] Optionally, the steps of extracting the target face features of the face to be identified from the target image include: performing face recognition processing on the target image to obtain key points of the face to be identified; performing pose adjustment processing on the face to be identified based on the key points of the face to be identified; and performing feature extraction processing on the adjusted face to be identified through a pre-trained face recognition model to obtain the target face features of the face to be identified.

[0098] Optionally, the step of determining whether the face to be identified matches the target identity based on the age parameter, the target face features, and the feature classification results corresponding to the target identity includes: determining the standard face features of the age group that matches the age parameter from the standard face features of multiple age groups corresponding to the target identity as reference standard face features; matching the target face features with the reference standard face features to obtain a matching result; and determining whether the face to be identified matches the target identity based on the matching result.

[0099] Optionally, the aforementioned reference standard face features are obtained by clustering initial face features from multiple images, including the face corresponding to the target identity; the faces in the images correspond to age groups that match the age parameters; the reference standard face features include the central cluster feature of a set of face feature combinations obtained through clustering; the face feature combination also includes multiple initial face features; the matching result includes whether the target face feature matches the reference standard face feature, or whether the target face feature does not match the reference standard face feature; the step of matching the target face feature with the reference standard face feature to obtain the matching result includes: calculating a first distance between the target face feature and the reference standard face feature; determining whether the first distance is greater than the cluster distance of the face feature combination in which the reference standard face feature is located; the cluster distance includes the farthest distance among the distances between the central cluster feature of the face feature combination and multiple initial face features in the face feature combination; if not, determining that the target face feature matches the reference standard face feature.

[0100] Optionally, the aforementioned reference standard face features are obtained by clustering initial face features from multiple images, including the face corresponding to the target identity; the faces in the images correspond to age groups that match the age parameter; the reference standard face features include the central cluster feature of multiple face feature combinations obtained by clustering; the face feature combination also includes at least one initial face feature; among the multiple face feature combinations, the face feature combination that includes the most initial face features is the target feature combination for the target identity in the age group that matches the age parameter; the matching result includes matching or not matching; the step of matching the target face feature with the reference standard face features to obtain the matching result includes: for each reference standard face feature, calculating the second distance between the target face feature and the reference standard face feature; taking the shortest second distance among the multiple second distances as the target distance of the target face feature; determining whether the reference standard face feature corresponding to the target distance is located in the target feature combination; if so, determining that the target face feature matches the reference standard face feature.

[0101] Optionally, the matching result includes whether the target face feature matches the reference standard face feature or whether the target face feature does not match the reference standard face feature. Based on the matching result, the step of determining whether the face to be identified matches the target identity includes: if the target face feature matches the reference standard face feature, determining that the face to be identified matches the target identity; if the target face feature does not match the reference standard face feature, determining that the face to be identified does not match the target identity.

[0102] See Figure 6 As shown, the electronic device includes a processor 100 and a memory 101. The memory 101 stores machine-executable instructions that can be executed by the processor 100. The processor 100 executes the machine-executable instructions to implement the above-described face recognition method.

[0103] Furthermore, Figure 6 The electronic device shown also includes a bus 102 and a communication interface 103, with the processor 100, the communication interface 103 and the memory 101 connected via the bus 102.

[0104] The memory 101 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 103 (which can be wired or wireless), such as the Internet, wide area network, local area network, or metropolitan area network. The bus 102 may be an ISA bus, PCI bus, or EISA bus, etc. The 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.

[0105] Processor 100 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 100 or by instructions in software form. Processor 100 may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it may 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 invention. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention 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 readily available storage medium in the art, 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 101, and the processor 100 reads the information from memory 101 and, in conjunction with its hardware, completes the steps of the method described in the foregoing embodiments.

[0106] This embodiment also provides a machine-readable storage medium storing machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions cause the processor to implement the above-described face recognition method.

[0107] The facial recognition method, apparatus, and electronic device provided in this invention include a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the preceding method embodiments, for example:

[0108] Acquire the target image and the age parameter of the face to be identified in the target image; extract the target face features of the face to be identified from the target image; determine whether the face to be identified matches the target identity based on the age parameter, the target face features and the feature classification results corresponding to the target identity; wherein, the feature classification results include: standard face features of multiple age groups corresponding to the target identity.

[0109] The above method takes into account that a person's appearance changes with age. Therefore, during the face recognition process, the facial features of the face to be identified are matched with standard facial features that match its age parameters, thereby improving the accuracy of face recognition.

[0110] Optionally, the steps of obtaining the target image and the age parameter of the face to be identified in the target image include: obtaining the target image and the shooting time of the target image; and calculating the age parameter of the face to be identified in the target image based on the pre-acquired birthday information of the target identity and the shooting time of the target image.

[0111] Optionally, the aforementioned standard facial features are determined as follows: for each age group, multiple images containing faces corresponding to the target identity are acquired; the number of images matches a preset number; for each image, initial facial features are extracted from the face in the image; the initial facial features of multiple images are clustered to obtain clustering results; the clustering results include at least one set of facial feature combinations; the facial feature combination includes a central clustering feature and at least one initial facial feature; the central clustering feature is determined as the standard facial feature.

[0112] Optionally, the above step of acquiring multiple images containing faces corresponding to the target identity and age group includes: acquiring images containing faces corresponding to the target identity and the image capture time; determining the age parameter corresponding to the image based on the pre-acquired birthday information of the target identity and the image capture time; determining whether the age parameter corresponding to the image is within the age group; if so, determining the acquired image as an image containing faces corresponding to the target identity; determining whether the number of acquired images containing faces corresponding to the target identity and age group is equal to a preset number; if not, continuing to execute the steps of acquiring images containing faces corresponding to the target identity and the image capture time until the number of acquired images containing faces corresponding to the target identity is equal to the preset number.

[0113] Optionally, the steps of extracting the target face features of the face to be identified from the target image include: performing face recognition processing on the target image to obtain key points of the face to be identified; performing pose adjustment processing on the face to be identified based on the key points of the face to be identified; and performing feature extraction processing on the adjusted face to be identified through a pre-trained face recognition model to obtain the target face features of the face to be identified.

[0114] Optionally, the step of determining whether the face to be identified matches the target identity based on the age parameter, the target face features, and the feature classification results corresponding to the target identity includes: determining the standard face features of the age group that matches the age parameter from the standard face features of multiple age groups corresponding to the target identity as reference standard face features; matching the target face features with the reference standard face features to obtain a matching result; and determining whether the face to be identified matches the target identity based on the matching result.

[0115] Optionally, the aforementioned reference standard face features are obtained by clustering initial face features from multiple images, including the face corresponding to the target identity; the faces in the images correspond to age groups that match the age parameters; the reference standard face features include the central cluster feature of a set of face feature combinations obtained through clustering; the face feature combination also includes multiple initial face features; the matching result includes whether the target face feature matches the reference standard face feature, or whether the target face feature does not match the reference standard face feature; the step of matching the target face feature with the reference standard face feature to obtain the matching result includes: calculating a first distance between the target face feature and the reference standard face feature; determining whether the first distance is greater than the cluster distance of the face feature combination in which the reference standard face feature is located; the cluster distance includes the farthest distance among the distances between the central cluster feature of the face feature combination and multiple initial face features in the face feature combination; if not, determining that the target face feature matches the reference standard face feature.

[0116] Optionally, the aforementioned reference standard face features are obtained by clustering initial face features from multiple images, including the face corresponding to the target identity; the faces in the images correspond to age groups that match the age parameter; the reference standard face features include the central cluster feature of multiple face feature combinations obtained by clustering; the face feature combination also includes at least one initial face feature; among the multiple face feature combinations, the face feature combination that includes the most initial face features is the target feature combination for the target identity in the age group that matches the age parameter; the matching result includes matching or not matching; the step of matching the target face feature with the reference standard face features to obtain the matching result includes: for each reference standard face feature, calculating the second distance between the target face feature and the reference standard face feature; taking the shortest second distance among the multiple second distances as the target distance of the target face feature; determining whether the reference standard face feature corresponding to the target distance is located in the target feature combination; if so, determining that the target face feature matches the reference standard face feature.

[0117] Optionally, the matching result includes whether the target face feature matches the reference standard face feature or whether the target face feature does not match the reference standard face feature. Based on the matching result, the step of determining whether the face to be identified matches the target identity includes: if the target face feature matches the reference standard face feature, determining that the face to be identified matches the target identity; if the target face feature does not match the reference standard face feature, determining that the face to be identified does not match the target identity.

[0118] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and apparatus described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0119] Furthermore, in the description of the embodiments of the present invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention based on the specific circumstances.

[0120] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0121] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0122] Finally, it should be noted that the above embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A face identity recognition method, characterized in that, The method comprises: obtaining a target image and an age parameter of a face to be identified in the target image; extracting a target face feature of the face to be identified from the target image; determining a reference standard face feature from standard face features of multiple age stages corresponding to the target identity, the reference standard face feature being of an age stage matching the age parameter; matching the target face feature with the reference standard face feature to obtain a matching result; determining whether the face to be identified matches the target identity based on the matching result; the reference standard face feature is obtained based on clustering processing of initial face features in multiple images including a face corresponding to the target identity, the face in the images corresponding to an age stage matching the age parameter, the reference standard face feature including a center clustering feature of a set of face feature combinations obtained through clustering processing, the face feature combinations further including multiple initial face features, and the matching result including that the target face feature matches the reference standard face feature or that the target face feature does not match the reference standard face feature; the step of matching the target face feature with the reference standard face feature to obtain a matching result comprises: calculating a first distance between the target face feature and the reference standard face feature; determining whether the first distance is greater than a clustering distance of a face feature combination in which the reference standard face feature is located, the clustering distance including a farthest distance among distances between the center clustering feature of the face feature combination and multiple initial face features in the face feature combination; if not, determining that the target face feature matches the reference standard face feature.

2. The method of claim 1, wherein, The step of obtaining a target image and an age parameter of a face to be identified in the target image comprises: obtaining a target image and a shooting time of the target image; calculating the age parameter of the face to be identified in the target image based on pre-obtained birthday information of the target identity and the shooting time of the target image.

3. The method of claim 2, wherein, The standard face feature is determined by the following method: for each age stage, obtaining multiple images corresponding to the age stage and including a face corresponding to the target identity, the number of the images matching a preset number; for each of the images, extracting an initial face feature of the face in the image; performing clustering processing on the initial face features of the multiple images to obtain a clustering result, the clustering result including at least one face feature combination, the face feature combination including a center clustering feature and at least one initial face feature; determining the center clustering feature as a standard face feature.

4. The method of claim 3, wherein, The step of obtaining multiple images corresponding to the age stage and including a face corresponding to the target identity comprises: obtaining an image including a face corresponding to the target identity and a shooting time of the image; determining an age parameter corresponding to the image based on pre-obtained birthday information of the target identity and the shooting time of the image; determining whether the age parameter corresponding to the image is within the age stage. If yes, the obtained image is determined as an image containing a face corresponding to the target identity; If no, the steps of obtaining an image containing a face corresponding to the target identity and the shooting time of the image are continued until the number of obtained images containing a face corresponding to the target identity is equal to the preset number.

5. The method of claim 1, wherein, The step of extracting the target face feature of the face to be recognized from the target image includes: performing face recognition processing on the target image to obtain key points of the face to be recognized; performing pose adjustment processing on the face to be recognized based on the key points of the face to be recognized; performing feature extraction processing on the adjusted face to be recognized through a pre-trained face recognition model to obtain the target face feature of the face to be recognized.

6. The method of claim 1, wherein, The reference standard face feature is obtained by clustering initial face features in a plurality of images containing a face corresponding to the target identity; the face in the image corresponds to an age range matching the age parameter; The reference standard face feature includes a center clustering feature of a plurality of face feature combinations obtained by clustering; the face feature combination further includes at least one initial face feature; among the plurality of face feature combinations, the face feature combination with the largest number of initial face features is the target feature combination of the target identity in the age range matching the age parameter; The matching result includes matching or not matching; The step of matching the target face feature with the reference standard face feature to obtain a matching result includes: For each reference standard face feature, calculate a second distance between the target face feature and the reference standard face feature; The shortest second distance among a plurality of second distances is taken as a target distance of the target face feature; Determine whether the reference standard face feature corresponding to the target distance is located in the target feature combination; If yes, it is determined that the target face feature matches the reference standard face feature.

7. The method of claim 1, wherein, The matching result includes that the target face feature matches the reference standard face feature, or the target face feature does not match the reference standard face feature; Based on the matching result, the step of determining whether the face to be recognized matches the target identity includes: If the target face feature matches the reference standard face feature, it is determined that the face to be recognized matches the target identity; If the target face feature does not match the reference standard face feature, it is determined that the face to be recognized does not match the target identity.

8. A face identity recognition device, characterized by, The device includes: An image acquisition module for acquiring a target image and an age parameter of a face to be recognized in the target image; A feature extraction module for extracting a target face feature of the face to be recognized from the target image; A matching module for determining a standard face feature of an age range matching the age parameter from a plurality of standard face features of age ranges corresponding to a target identity as a reference standard face feature; match the target face feature with the reference standard face feature to obtain a matching result; determine whether the face to be identified matches the target identity based on the matching result; the reference standard face feature is obtained by clustering initial face features in a plurality of images including a face corresponding to the target identity; the face in the image corresponds to an age range matching the age parameter; the reference standard face feature includes a center clustering feature of a set of face feature combinations obtained by clustering; the face feature combination further includes a plurality of initial face features; the matching result includes that the target face feature matches the reference standard face feature, or the target face feature does not match the reference standard face feature; the matching module is further configured to: calculate a first distance between the target face feature and the reference standard face feature; determine whether the first distance is greater than a clustering distance of a face feature combination in which the reference standard face feature is located; the clustering distance includes the farthest distance among distances between the center clustering feature of the face feature combination and a plurality of initial face features in the face feature combination; if not, determine that the target face feature matches the reference standard face feature.

9. An electronic device, comprising: a processor and a memory, the memory storing machine executable instructions executable by the processor, and the processor executes the machine executable instructions to implement the face identity recognition method of any one of claims 1-7.

10. A machine-readable storage medium, characterized in that, The machine readable storage medium stores machine executable instructions, and when the machine executable instructions are called and executed by the processor, the machine executable instructions cause the processor to implement the face identity recognition method of any one of claims 1-7.

Images