Identification method and device, equipment, storage medium

Abstract

The application discloses a kind of identification method and device, equipment, storage medium;Wherein, the method comprises: the feature extraction of the object to be identified in the first image collected, and the feature vector of first key feature point is obtained;The difference between the feature vector of any first key feature point and at least one other first key feature point is compared, and the first comparison result corresponding to any first key feature point is obtained;Each first key feature point corresponding first comparison result is used as the feature of the object to be identified, and at least based on the feature vector of each first key feature point and corresponding first comparison result, whether the object to be identified belongs to first service type is identified.Such, increase a kind of consideration factor, so that the identification ability is stronger, and accuracy is higher.

Description

Technical Field

[0001] This application relates to information technology, and to, but is not limited to, an identification method, apparatus, device, and storage medium. Background Technology

[0002] For most stores, once a customer enters the store, the staff cannot accurately determine the customer's identity information (such as whether the customer is a member or not), and therefore cannot provide targeted services based on the customer's identity information. Summary of the Invention

[0003] In view of this, the identification method, apparatus, equipment, and storage medium provided in this application have stronger identification capabilities and higher accuracy in identifying whether the object to be identified belongs to the first service type.

[0004] According to one aspect of the embodiments of this application, an identification method is provided, comprising: extracting features from an object to be identified in a first image to obtain feature vectors of first key feature points; comparing the differences between the feature vectors of any first key feature point and at least one other first key feature point to obtain a first comparison result corresponding to the any first key feature point; using the first comparison results corresponding to each first key feature point as features of the object to be identified, and identifying whether the object to be identified belongs to a first service type based at least on the feature vectors of each first key feature point and the corresponding first comparison results.

[0005] According to one aspect of the embodiments of this application, an identification device is provided, comprising: an extraction module, configured to extract features from an object to be identified in a first image to obtain feature vectors of first key feature points; a comparison module, configured to compare the differences between the feature vectors of any first key feature point and at least one other first key feature point to obtain a first comparison result corresponding to the any first key feature point; and an identification module, configured to use the first comparison results corresponding to each first key feature point as features of the object to be identified, and to identify whether the object to be identified belongs to a first service type based at least on the feature vectors of each first key feature point and the corresponding first comparison results.

[0006] According to one aspect of the present application, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program executable on the processor, and the processor executes the program to implement the method described in the embodiments of the present application.

[0007] According to one aspect of the embodiments of this application, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the methods provided in the embodiments of this application.

[0008] In this embodiment, feature vectors of first key feature points are obtained by extracting features from the object to be identified in the first image; the differences between the feature vectors of any first key feature point and at least one other first key feature point are compared to obtain a first comparison result corresponding to the first key feature point; and whether the object to be identified belongs to a first service type is identified based at least on the feature vectors of each first key feature point and the corresponding first comparison result. This results in stronger identification capability and higher accuracy when identifying whether the object to be identified belongs to a first service type. This is because, in this embodiment, not only are the differences between the feature vectors of the first key feature points of the object to be identified and the feature vectors of the second key feature points of other objects belonging to the first service type compared, but the differences between the feature vectors of the first key feature points of the object to be identified are also considered laterally. Therefore, when identifying whether the object to be identified belongs to a first service type based at least on the feature vectors of each first key feature point and the corresponding first comparison result, an additional factor is added, resulting in stronger identification capability and higher accuracy.

[0009] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0010] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application. Obviously, the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0011] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily need to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0012] Figure 1 A schematic diagram illustrating the implementation process of an identification method provided in an embodiment of this application;

[0013] Figure 2 A schematic diagram illustrating feature extraction processing of the object to be identified, provided in an embodiment of this application;

[0014] Figure 3A schematic diagram illustrating the implementation process of an identification method provided in an embodiment of this application;

[0015] Figure 4 A schematic diagram illustrating the implementation process of a database construction process provided in this application embodiment;

[0016] Figure 5 A schematic diagram of the method flow for a membership service based on a micro base station provided in an embodiment of this application;

[0017] Figure 6 A schematic diagram of the membership management platform provided in an embodiment of this application;

[0018] Figure 7 This is a schematic diagram of the structure of the identification device according to an embodiment of this application;

[0019] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the specific technical solutions of this application will be further described in detail below with reference to the accompanying drawings of the embodiments of this application. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0022] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0023] It should be noted that the terms "first, second, third" used in the embodiments of this application do not represent a specific order of objects. It is understood that "first, second, third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0024] This application provides an identification method applied to an electronic device. This electronic device can be of various types with information processing capabilities, such as a cash register, desktop computer, laptop computer, mini-laptop, tablet computer, mobile phone, or personal digital assistant (PDA). The function implemented by this method can be achieved by a processor in the electronic device calling program code. The program code can be stored in a computer storage medium. Therefore, the electronic device includes at least a processor and a storage medium.

[0025] Figure 1 This is a schematic diagram illustrating the implementation flow of the identification method provided in the embodiments of this application, as shown below. Figure 1 As shown, the method may include the following steps 101 to 103:

[0026] Step 101: Extract features from the object to be identified in the first acquired image to obtain the feature vector of the first key feature point.

[0027] In the embodiments of this application, there may be one or more objects to be identified in the first image, and the number of objects to be identified is not limited.

[0028] Here, a camera can be installed at the entrance of the store to capture images of the object to be identified entering the store, thus obtaining a first image. In some embodiments, after obtaining the first image, the quality of the first image can be detected, and features can be extracted from the first image that meets the quality requirements.

[0029] In some embodiments, the process of feature extraction of the object to be identified in the first image can be as follows: feature vectors of the first key feature points are obtained through face localization, face registration, and face feature extraction. For example... Figure 2 As shown, in the process, facial localization processing represents the face of the object to be identified in the input first image using a facial bounding box 201; facial registration processing transforms the facial coordinate box of the object to be identified into facial key feature points 202. Of course, the number of key feature points is not limited and can be freely set according to the actual situation; facial feature extraction processing transforms the facial key feature points 202 into a feature vector 203 of a fixed length, resulting in the feature vector set V0 = {P1, P2, P3, ..., P...} of the first key feature points. n}, n>0.

[0030] Step 102: Compare the differences between the feature vectors of any first key feature point and at least one other first key feature point to obtain the first comparison result corresponding to the any first key feature point.

[0031] It should be noted that, here, comparing the difference between the feature vector of any first key feature point and at least one other first key feature point requires comparing the feature vector of each first key feature point in set V0 once.

[0032] In this embodiment of the application, the number of other first key feature points is not limited, and can be any first key feature point P. i With another first key feature point P m The difference between the feature vectors; or the comparison of any first key feature point P i The differences between the feature vectors of multiple other first key feature points, such as comparing the first key feature point P. i With the first key feature point P m P j The differences between the feature vectors, or the comparison of the first key feature point P i With the first key feature point P m P j and P d The difference between the feature vectors.

[0033] Of course, there is no restriction on the order in which the differences between the feature vectors of any first key feature point and at least one other first key feature point are compared. i With another first key feature point P m When comparing the differences between feature vectors, one can compare the differences between P1 and P2, or the differences between P1 and P3, P4, P5, etc. Correspondingly, when comparing any first key feature point P... i When comparing the differences between the feature vectors of multiple other first key feature points, one can compare the differences between P1 and P2, P3; or one can compare the differences between P1 and P4, P6.

[0034] In the embodiments of this application, the representation method of the first comparison result is not limited. The first comparison result can be a difference vector between key feature points, or it can be the sum of differences, variance, etc.

[0035] Step 103: Use the first comparison result corresponding to each first key feature point as the feature of the object to be identified, and identify whether the object to be identified belongs to the first service type based at least on the feature vector of each first key feature point and the corresponding first comparison result.

[0036] In some embodiments, the first service type is a membership service type, and identifying whether the object to be identified belongs to the first service type is equivalent to identifying whether the object to be identified is a member.

[0037] In this embodiment of the application, when identifying whether an object to be identified belongs to the first service type, not only is the difference between the feature vectors of the first key feature points of the object to be identified and the feature vectors of the second key feature points of other objects belonging to the first service type compared, but the difference between the feature vectors of the first key feature points of the object to be identified is also considered laterally. Therefore, when identifying whether an object to be identified belongs to the first service type based at least on the feature vectors of each first key feature point and the corresponding first comparison results, the identification ability is stronger and the accuracy is higher.

[0038] Figure 3 This is a schematic diagram illustrating the implementation flow of the identification method provided in the embodiments of this application, as shown below. Figure 3 As shown, the method may include the following steps 301 to 307:

[0039] Step 301: Extract features from the object to be identified in the first image to obtain the feature vector of the first key feature point;

[0040] Step 302: Compare the differences between the feature vectors of any first key feature point and at least one other first key feature point to obtain the first comparison result corresponding to the any first key feature point;

[0041] Step 303: Obtain the second comparison result corresponding to the second key feature point of the first object belonging to the first service type in the database.

[0042] The second comparison result is obtained by comparing the difference between the feature vectors of any second key feature point of the first object and at least one other second key feature point of the first object.

[0043] In this embodiment, the number of first objects belonging to the first service type in the database can be one or more. In some embodiments, the method for determining the second key feature points of the first object can be implemented by performing steps 403 to 404 in the following embodiments, thereby obtaining a set Sv = {V1, V2, V3, ..., V...} of each first object in the database. m}, where m is the number of first objects. For a given first object V... n In this regard, the corresponding set of the second key feature points is V. n ={P1,P2,P3,...,P n}, n>0.

[0044] Step 304: Compare the difference between the feature vector of any first key feature point of the object to be identified and the feature vector of the corresponding second key feature point of the first object to obtain the third comparison result corresponding to any first key feature point.

[0045] Here, the set of the first key feature points of the object to be identified can be determined in a certain order. For example, if the set of the first key feature points V0 = {P1, P2, P3, ..., P...} n Let n > 0 be the set of key facial feature points of the object to be identified. Then, these key feature points can be arranged in the order of the identified eyebrows, nose, and mouth, resulting in set V0. Similarly, when determining the second set of key feature points V for a given object... m At the same time, it is also necessary to obtain the second key feature point set V in the same definite order as in the V0 set. m ={P1,P2,P3,...,P n}, n>0. That is, the number of the first key feature points is the same as the number of the second key feature points, and the first key feature point P i With the second key feature point P i Yes, they correspond, such as the first key feature point P. i Let P be the nasal tip feature point of the object to be identified. i This is also the nasal tip feature point of the first object.

[0046] Therefore, comparing the difference between the feature vectors of any first key feature point of the object to be identified and the corresponding second key feature point of the first object is equivalent to comparing the first key feature point P. i With the corresponding second key feature point P i The difference between the feature vectors.

[0047] Accordingly, in step 302, any first key feature point P in set V0 is compared. i With the first key feature point P m When obtaining the first comparison result based on the differences between the feature vectors of the first object, in step 303, for the second key feature point set V of the first object... m ={P1,P2,P3,...,P n For n>0, it is also done by comparing the second key feature point P. i With the second key feature point P m The difference between the feature vectors is used to obtain the second key feature point P. i The corresponding second comparison result.

[0048] Step 305: Use the second comparison results corresponding to each second key feature point as features of the first object, and determine the matching degree between the object to be identified and the first object based on each first comparison result, each second comparison result and each third comparison result.

[0049] In this embodiment, not only is the difference between the feature vectors of the first key feature point of the object to be identified and the corresponding second key feature point of the first object considered vertically, but the difference between the feature vectors of the first key feature point of the object to be identified and the difference between the feature vectors of the second key feature point of the first object are also considered horizontally. This method of determining the matching degree between the object to be identified and the first object based on multiple factors can make the matching accuracy higher.

[0050] In this embodiment of the application, the method for determining the matching degree between the object to be identified and the first object is not limited. For example, it can be to calculate the difference between the object to be identified and the first object, and use the difference to characterize the matching degree between the object to be identified and the first object; or it can be to calculate the similarity between the object to be identified and the first object, and use the similarity to characterize the matching degree between the object to be identified and the first object.

[0051] In some embodiments, step 305 can be achieved by performing the following steps 3051 to 3053:

[0052] Step 3051: Based on each first comparison result and each second comparison result, determine the first degree of difference between the object to be identified and the first object.

[0053] In some embodiments, the first difference G between the object to be identified and the first object can be determined using Formula 1. m Among them, g i The first key feature point P i The corresponding first comparison result, h i The second key feature point P i The corresponding second comparison result.

[0054]

[0055] Step 3052: Based on each of the third comparison results, determine the second degree of difference between the object to be identified and the first object.

[0056] In some embodiments, the second difference degree D between the object to be identified and the first object can be determined using Formula 2. m Among them, d j The first key feature point P i The corresponding third comparison result.

[0057]

[0058] Step 3053: Based on the first difference degree and the second difference degree, determine the matching degree between the object to be identified and the first object.

[0059] In some embodiments, after obtaining the first difference G m Second difference degree Dm Then, the first difference G can be used to... m Second difference degree D m The product method determines the matching degree between the object to be identified and the first object; alternatively, it can be determined by calculating the first difference G. m Second difference degree D m The summation method is used to determine the matching degree between the object to be identified and the first object, and there are no restrictions on this.

[0060] Step 306: Obtain the second comparison result corresponding to the second key feature point of the next first object, thereby determining the matching degree between the object to be identified and the next first object.

[0061] After determining the matching degree between the object to be identified and a first object in the database, steps 303 to 305 need to be repeated to determine the matching degree between the object to be identified and the next first object in the database, until the matching degree between the object to be identified and each first object in the database is obtained.

[0062] Step 307: Based on each matching degree, determine whether the object to be identified belongs to the first service type.

[0063] In some embodiments, step 307 can be achieved by performing the following steps 3071 to 3074:

[0064] Step 3071: From all the obtained matching degrees, determine the target matching degree that represents the highest similarity between the object to be identified and the first object.

[0065] In some embodiments, a target matching degree that represents the minimum difference between the object to be identified and the first object can be determined from all the obtained matching degrees.

[0066] Step 3072: Determine whether the target matching degree meets specific conditions; if the target matching degree meets specific conditions, proceed to step 3073; otherwise, proceed to step 3074.

[0067] Here, determining whether the target matching degree meets specific conditions can be done by determining whether the target matching degree meets a preset threshold. Of course, the preset threshold can be freely set according to the actual situation.

[0068] When the target matching degree is the highest similarity, it can be determined whether the object to be identified belongs to the first service type by judging whether the target matching degree is less than a preset threshold. For example, if the preset threshold is set to 95%, and the target matching degree with the highest similarity is 96%, which is greater than 95%, it means that there is an object among the first objects with very similar features to the object to be identified, and it can be determined that the object to be identified belongs to the first service type. If the target matching degree with the highest similarity is 50%, which is obviously less than 95%, although the similarity between this first object and the object to be identified is the highest among all first objects, it is still not actually similar to the features of the object to be identified, and it can be determined that the object to be identified does not belong to the first service type.

[0069] When the target matching degree is the minimum difference, it can be determined whether the object to be identified belongs to the first service type by judging whether the target matching degree is greater than a preset threshold. For example, if the preset threshold is set to 5%, if the target matching degree with the minimum difference is 4%, which is less than 5%, it means that there is an object among the first objects that has very similar characteristics to the object to be identified, and it can be determined that the object to be identified belongs to the first service type. If the target matching degree with the minimum difference is 30%, which is obviously greater than 5%, although the difference between this first object and the object to be identified is the smallest among all first objects, it is still not similar to the characteristics of the object to be identified, and it can be determined that the object to be identified does not belong to the first service type.

[0070] Step 3073: Determine that the object to be identified belongs to the first service type.

[0071] In some embodiments, when it is determined that the object to be identified belongs to a first service type, a first prompt message is output; wherein, the first prompt message is used to instruct the first service to be provided to the object to be identified.

[0072] Here, there are no restrictions on how the first prompt message is output. For example, it can be displayed on the screen of an electronic device; or it can be output via voice prompt.

[0073] Step 3074: Determine that the object to be identified belongs to the second service type.

[0074] In some embodiments, when it is determined that the object to be identified belongs to the second service type, a second prompt message is output; wherein the second prompt message is used to indicate that a second service is provided for the object to be identified, the first service type and the second service type are different, and the second service is different from the first service.

[0075] In some embodiments, the first service type is a membership service type, and the second service type can be a non-member service type, or a service type for non-members and members other than those in the first service type. For example, when the first service type is a membership service type (including regular members, premium members, high-quality members, etc.), the second service type is a non-member service type; when the first service type is a regular membership service type, the second service type can be a non-member service type and membership service types such as premium members and high-quality members.

[0076] In some embodiments, such as Figure 4 As shown, the database construction process includes the following steps 401 to 405:

[0077] Step 401: Detect the identity and movement direction of the second object within a specific physical area.

[0078] In some embodiments, a specific physical area is the coverage area of ​​a micro base station, through which the identity and movement direction of a second object within its coverage area can be detected. The second object includes member objects and non-member objects.

[0079] Here, there is no limitation on the type of identification. For example, the identification could be a telephone number used by the second party; another example could be a Personal Identification Number (PIN) used by the second party; yet another example could be a national identity card number used by the second party.

[0080] In some embodiments, the movement direction of the second object can be determined using the temporal location information of the second object. For example, the location of the service provider point is determined to be p0, and the temporal location information of the second object is obtained as Tp. k ={p1,p2,p3,...,p i ,...,p c}, c≥0, c≥i≥0, by calculating the distance between each temporal location information and the location p0 of the service provider point, it is possible to determine whether the second object moves in the direction of the service provider point.

[0081] Step 402: Select the first object from the second objects based on the identity identifier and movement direction of the second object.

[0082] In some embodiments, step 402 is achieved by performing the following steps 4021 to 4022:

[0083] Step 4021: Select the target object belonging to the first service type from the second object based on the identity identifier of the second object.

[0084] Here, the identity identifier of the detected second object is compared with the identity identifier of the first service type object pre-stored in the system of the service provider, so that the target object belonging to the first service type can be selected from the second object.

[0085] In some embodiments, after selecting a target object, the service provider's system can obtain marketing information and send it to the target object. The marketing information can be the same or different. When the marketing information is different, the target object's preferences can be determined through information such as their historical consumption records, allowing for targeted marketing by sending different marketing information. Furthermore, sending marketing information to target objects within a specific physical area avoids frequently disturbing them with marketing messages when they are not near the service provider, i.e., when they have no need for the service.

[0086] Step 4022: Determine whether the target object is gradually approaching the service provision point based on the target object's movement direction; if so, designate the target object as the first object.

[0087] Here, the target object is only identified as the first object when it gradually approaches the service provider point. Thus, when constructing the database using information about the first object, it is not based on all target objects belonging to the first service type within a specific physical area. Instead, it selects objects that are close to the service provider point from among the target objects. This ensures that the database contains first objects that cover the objects to be identified in the first image, while reducing the number of first objects in the database that need to be compared with the objects to be identified. This accelerates the identification speed and improves the accuracy of the identification.

[0088] Step 403: Extract features from the first object to obtain the feature vector of the second key feature point.

[0089] Here, the method for extracting features from the first object is the same as the method for extracting features from the object to be identified in step 101, and will not be repeated here.

[0090] Step 404: Compare the difference between the feature vectors of any second key feature point of the first object and at least one other second key feature point of the first object to obtain a second comparison result corresponding to the any second key feature point.

[0091] Here, the method of comparing the difference between the feature vectors of any second key feature point of the first object and at least one other second key feature point is the same as the method of comparing the difference between the feature vectors of any first key feature point and at least one other first key feature point in step 102, and will not be repeated here.

[0092] Step 405: The identity identifier of the first object and the feature vector of the second key feature point of the first object and the second comparison result corresponding to the second key feature point are taken as a feature group. Based on the feature group of each first object, a database is obtained.

[0093] It should be noted that each feature group of the first object stores the feature vector of each second key feature point of the first object and the corresponding second comparison result.

[0094] In some embodiments, the database can also be updated within a preset period. The preset period is not limited and can be one hour, one day, or one week, etc.

[0095] Fifth-generation mobile communication technology (5G), as a new generation of communication technology, will provide the main foundational network support for the Internet of Things, enabling applications such as autonomous driving, smart cities, smart industry, and real-time virtual reality (VR) / augmented reality (AR). 5G network construction, as a major development direction for communication technology in the coming years, has been elevated to a strategic level by major countries worldwide, with major manufacturers and operators across the related industry chain actively deploying it. 5G network deployment is affected by spectrum resources; high-frequency communication results in shorter base station distances, making the use of streetlights and other structures to build 5G micro base stations in densely populated urban areas a viable option for 5G network deployment. Micro base stations are miniaturized versions of conventional base stations, making installation and construction much easier. Applying them to 5G network construction can significantly improve efficiency and has strong practical application value. Because 5G millimeter waves have poor penetration and significant attenuation in the air, if 5G still uses the "macro base stations" employed in the 3G and 4G eras, it will not be able to provide sufficient signal support for users at greater distances. To address this challenge, 5G has begun using entirely new base stations—micro base stations. As the name suggests, micro base stations are base stations made small enough to be considered large enough.

[0096] Facial recognition is a biometric technology that identifies individuals based on their facial features. It involves using cameras or webcams to capture images or video streams containing faces, automatically detecting and tracking faces within the images, and then performing facial recognition. This is a series of related technologies, often called image recognition or facial identification. Facial recognition technology primarily relies on visible light images, a familiar method with over 30 years of development history. However, this method has inherent limitations, particularly when ambient lighting changes, causing a sharp decline in recognition accuracy and failing to meet the needs of practical systems. Solutions to the lighting problem include 3D image facial recognition and thermal imaging facial recognition. However, these two technologies are still far from mature, and their recognition results are unsatisfactory. A rapidly developing solution is multi-source facial recognition technology based on active near-infrared imaging. It overcomes the influence of lighting changes and has achieved superior recognition performance, surpassing 3D image facial recognition in overall system performance in terms of accuracy, stability, and speed. This technology has developed rapidly in the last two to three years, gradually bringing facial recognition technology towards practical application. Like other biometric features of the human body (fingerprints, irises, etc.), the face is innate. Its uniqueness and the fact that it is not easy to be copied provide the necessary premise for identity verification. Compared with other types of biometrics, face recognition has the following characteristics: (1) Non-mandatory: Users do not need to cooperate with the face acquisition device. They can obtain face images almost unconsciously. This sampling method is not "mandatory". (2) Non-contact: Users do not need to directly contact the device to obtain face images. Concurrency: In actual application scenarios, multiple faces can be sorted, judged and identified. (3) In addition, it also conforms to the characteristics of visual characteristics: "recognizing people by their appearance", as well as the characteristics of simple operation, intuitive results and good concealment.

[0097] In terms of the research status of related technologies, automatic face detection is the foundation of all applications surrounding automatic face image analysis, including but not limited to: face recognition and verification, face tracking in surveillance situations, facial expression analysis, facial attribute recognition (gender / age recognition, beauty assessment), facial lighting adjustment and deformation, facial shape reconstruction, image and video retrieval, and organization and presentation of digital albums. Face detection is the initial step in all modern vision-based human-computer and human-robot interactive systems. Mainstream commercial digital cameras have built-in face detection to assist autofocus. Many social networks use face detection mechanisms to achieve image and / or person tagging. From the perspective of the problem domain, face detection belongs to the field of object detection. Object detection usually has two major categories: (1) General object detection: detecting multiple categories of objects in an image, such as various general object detection methods used in certain scenarios. Among them, the core of general object detection is the n(object) + 1(background) = n + 1 classification problem. This type of detection usually has a large model and is slow. Few STOA methods can achieve CPU real-time. (2) Specific category object detection: This type of detection detects only a specific type of object in an image, such as face detection, pedestrian detection, vehicle detection, etc. The core of specific category object detection is a binary classification problem of 1 (object) + 1 (background). This type of detection usually has a small model and very high speed requirements. The basic requirement here is the central processing unit real-time (CPU real-time).

[0098] From a historical perspective, the role of deep learning is very evident: In the pre-deep learning phase, classic detection algorithms were designed for specific targets, such as face detection, pedestrian detection, and various other object detection problems. However, multi-object detection required training templates for each category, essentially creating 200 specific category detection problems. In the deep learning phase, classic detection algorithms addressed general targets, such as faster algorithms like Faster Region-CNN (Faster-RCNN), the Region-based Fully Convolutional Network (R-FCN) series, the faster You Only Look Once (YOLO), and the Regression-based SingleShot MultiBoxDetector (SSD) series. Powerful deep learning allows for multi-category detection tasks with just a single Convolutional Neural Network (CNN). Although these are all multi-class methods, they can all be used to solve single-class problems. State-of-the-art (SOTA) models for specific target detection problems such as face detection and pedestrian detection are targeted improvements on these methods. Faster-RCNN series: The advantage of these methods is high performance, but the disadvantage is slow speed. They cannot operate in real-time on GPUs, failing to meet the extremely high speed requirements of face detection. Since performance is not an issue, the research focus for these methods is improving efficiency. SSD series: The advantage of these methods is speed, achieving real-time performance on GPUs. The disadvantage is poor detection of densely packed small targets, which faces happen to be. The research focus for these methods is improving the detection performance of densely packed small targets, while also needing to be as fast as possible. Real-time GPU algorithms remain limited in application.

[0099] This application aims to create a novel membership service system. Related membership service systems primarily focus on post-consumption services, such as post-consumption fee deductions, top-ups, points management, and SMS marketing. These are passive membership services. For merchants, effectively obtaining member location dynamics, such as whether a member is near the store, and conducting timely and precise marketing plays a crucial role in improving store sales and member satisfaction. In related technologies, merchants mainly promote in-store products through mass SMS messaging. However, mass SMS marketing is costly, and frequent marketing can cause member resentment and negatively impact the brand. Furthermore, existing membership management systems cannot immediately obtain member profiles upon a member entering the store, making personalized marketing services impossible.

[0100] Based on this, the following will describe an exemplary application of the embodiments of this application in a practical application scenario.

[0101] This application embodiment relies on the operator's 5G micro base station to implement a face recognition system, which can be integrated into the merchant's existing membership management platform to enrich the functionality of the membership management platform.

[0102] The membership service method based on a micro base station provided in this application includes three parts: a 5G micro base station, a face recognition system, and a membership management platform, as follows: Figure 5 As shown. The main principle is that after merchants connect to the system, when a member enters the coverage area of ​​a 5G micro base station, the merchant sends different marketing messages to the member (i.e., the target audience) through the 5G micro base station; members facing the merchant's location (i.e., the first target audience) are selected to build a temporary facial recognition database; when the member enters the store, the facial recognition system and the temporary database quickly identify the member's information, such as name and consumption preferences, and notify service personnel to greet and guide them, achieving precise service; and facial recognition payment service is provided at the time of payment, improving the member's service experience. The membership service system includes the following three modules:

[0103] Module 1: 5G Micro Base Station

[0104] 5G micro base stations include those uniformly installed by operators within shopping malls and those installed by merchants themselves. Using 5G micro base stations can fill network gaps in densely populated urban areas. While outdoor macro base stations can provide network coverage for most areas in densely populated urban areas due to the complex network of buildings and obstructions, lamppost micro base stations can fill these gaps, ensuring coverage in areas with weak signals and improving overall network capacity. Furthermore, they can offload traffic from high-traffic areas. Dense urban areas experience high traffic volumes, especially in commercial streets, tourist plazas, and areas with high student activity, where outdoor base stations struggle to effectively penetrate and cover the network, and cannot meet the network capacity demands of high-volume user groups, easily leading to network congestion. 5G micro base stations can address this by offloading traffic from these hotspots. Finally, sensitive macro base stations can be split into micro base stations. In related technologies, 5G macro base station pilot projects centrally deploy the baseband-based unit (BBU), integrating the radio frequency remote unit and antenna into an active antenna unit (AAU). Because 64T64R (64 transmit, 64 receive) AAU equipment is relatively large, it can easily cause resentment among residents. In 5G scenarios, some performance aspects can be reduced to form micro base station products, such as developing 16T16R (16 transmit, 16 receive), which has a smaller size, lower noise, and better concealment. Breaking down macro base stations, which have long been shelved due to sensitivity, into multiple micro base stations can overcome the challenge of building macro base stations in sensitive properties.

[0105] 5G micro base stations serve two main functions: acquiring phone numbers (i.e., identifiers) within their coverage area (i.e., a specific physical area) and transmitting these numbers to a membership management platform; and sending marketing SMS messages to members. The principle is as follows: the micro base station sends phone numbers within its coverage area to the merchant's membership management platform. The platform matches the member's number and returns the number and corresponding marketing SMS message to the micro base station, which then sends the SMS message to the member's number.

[0106] Module 2: Membership Management Platform

[0107] The membership management platform primarily filters members acquired from base stations for targeted message marketing, and further filters members who are interested in visiting stores; for members entering a store, service specialists are notified to greet and guide them (i.e., the first service type) to ensure precise service; and facial recognition payment service is provided during payment to improve members' service perception.

[0108] The specific implementation plan for selecting members of the store is as follows:

[0109] Suppose that n members (i.e., target objects) are obtained in the vicinity via micro base stations. n ={m1,m2,m3,...,mk ,...,m n}, n≥0, n≥k≥0, m k This refers to the phone number or PIN (identity identifier) ​​of the kth member.

[0110] Assume the k-th member is in a continuous time period T k Within, c positions were obtained, so the temporal position information of the kth member is: Tp k ={p1,p2,p3,...,p i ,...,p c}, c≥0, c≥i≥0.

[0111] The location p0 of the merchant (i.e., the service provider) is a fixed value that remains unchanged in the two-dimensional XY coordinate system.

[0112] The offset vector V of the c positions of the kth member from p0 k For V k ={p1-p0,p2-p0,p3-p0,....,p i -p0,...,p c -p0},c≥0,c≥i≥0.

[0113] The value of a is calculated using the latest squared linear fitting formula y = a·x + b. If a >= 0, it indicates that the offset vector V k It is either constantly increasing or remaining stagnant, moving away from p0; conversely, if a < 0, it indicates that the offset vector V is increasing. k It has been decreasing and is approaching p0.

[0114] Therefore, if the k-th member V k The linear fitting coefficient a < 0 indicates that the members are facing the store, and then the member data in the face database is extracted to build a temporary face database.

[0115] The membership management platform records basic membership information, with each piece of information having the following fields: {id_card, id_type, id_name, id_gender, ...}, as shown below. Figure 6 As shown.

[0116] Its functions include: adding, deleting, modifying, and querying the member database; matching member numbers sent by the micro base station (there may be multiple member numbers), and then pushing the member's facial photo during registration to the facial recognition system; and returning the member numbers successfully matched by the facial recognition system and the corresponding marketing SMS messages to the micro base station.

[0117] Module 3: Face Recognition System

[0118] The facial recognition system includes functions for dynamically adjusting the facial database, facial data acquisition, and facial detection.

[0119] 3.1 Dynamic adjustment of the face database

[0120] The dynamic adjustment of the face database is used to support the construction of a dynamic database for the face recognition system. This dynamic database can be built based on the database entry interface. The system obtains the phone numbers of users connected to the base station through the micro base station data interface and queries the membership management platform for basic member information, including member photos and names, based on the phone numbers. A dynamic personnel database is built using this basic information. The system defaults to updating the dynamic database daily, meaning it is cleared at 00:00 each day. This ensures that the personnel information in the dynamic database includes people who have entered the store, while minimizing the size of the base database and improving recognition accuracy.

[0121] 3.2 Face capture function

[0122] The face capture function uses face cameras at the merchant entrance to capture facial information, which is then transmitted to the face recognition system. This includes photos of both members and non-members. The system supports the detection of face photos captured by the front-end camera, performs quality assessments, and extracts features from photos that meet quality requirements, preparing for subsequent recognition and attribute identification. The accuracy rate is over 99% when the left, right, tilt, and angles of the face image are less than 15°.

[0123] 3.3 Face detection function

[0124] The face detection function matches the photos captured by the camera with the face photos of members pushed to the face database by the membership management platform, and displays the information of successful matches on the membership management platform. Non-members are shown face photos and non-member prompts, while members are shown member photos, membership level, member consumption records, consumption preferences, etc. Merchants can quickly identify the identity of members and provide differentiated services.

[0125] The face detection function process includes face localization, face registration, face feature extraction, and face comparison.

[0126] (1) Face localization

[0127] The input to a face localization algorithm is an image (i.e., the first image captured), and the output is a sequence of face bounding box coordinates (0, 1, or multiple face bounding boxes). For example, the `get_frontal_face_detector` locator from the dlib library can be used, outputting one or more upward-facing rectangles as face bounding boxes.

[0128] (2) Face registration

[0129] Face registration is a technique for locating the coordinates of key facial features on a face. The input to a face registration algorithm is a "face image" plus a "face bounding box," and the output is a sequence of coordinates of key facial features (i.e., key feature points). The number of key facial features is a fixed value that is preset.

[0130] (3) Facial feature extraction

[0131] This function transforms a face image into a fixed-length string of values, extracting key points for face registration into a multi-dimensional feature vector V0 = {P1, P2, P3, ..., P...}. n}, n>0 (i.e., the eigenvector of the first key feature point).

[0132] (4) Face comparison

[0133] Face comparison is an algorithm that measures the similarity between two faces. The input to the face comparison algorithm is two facial features (Note: the facial features are obtained from the previous face feature extraction algorithm, which yields the feature to be compared, V0, and the face database Sv = {V1, V2, V3, ..., V...}). m}, where m is the number of faces in the database, and the output is the similarity between two features. Here, the method of comparing vector similarity in the gradient direction is used.

[0134] The formula for calculating the gradient direction G0 of V0 (i.e., the first comparison result) is as follows:

[0135] G={g n =P n -P n-1}, P∈V, n>0 (Formula 3);

[0136] Where V0 and V1 are compared by vector D 01 The calculation formula for (i.e., the third comparison result) is as follows:

[0137] D i ={d n =P n -Q n}, P∈V0, Q∈V t n = 128, t = m (Formula 4);

[0138] The similarity Sm between G0 and G1 01 The formula for calculating (i.e., matching degree) is as follows:

[0139]

[0140] Calculate Sm sequentially using the method for calculating Sm. 01 Sm 02 ,Sm 03 ,...,Sm 0mThen select the smallest Sm 0j ,j∈[1,m], if this Sm 0j If the value of j ∈ [1, m] is less than the given threshold T (T > 0), then the value of j is the ID of a person corresponding to the face database, thus successfully matching that person. If this Sm 0j If the value of j∈[1,m] is greater than the given threshold T (T>0), the match fails, indicating that the face database does not contain this person.

[0141] If single photo information is supported, the system will search a specific database and display the results in descending order of comparison scores. The system can support a database of 100,000 images; it is recommended to have a database of fewer than 50 images for the highest recognition rate. The highest accuracy rate is achieved with the largest database, exceeding 99.5%, as shown in Table 1 below, which illustrates the relationship between the number of images in the database, detection accuracy, and processing time. The system supports configuring different thresholds for face recognition, including face quality thresholds, feature point count thresholds, face comparison thresholds, and thresholds for different attributes. Thresholds can be dynamically configured according to actual business needs, and different thresholds can be set based on different scenarios, lighting conditions, or angles. The system supports setting different comparison thresholds for different face databases to prevent situations where the similarity score is too low for a visiting member, thus improving the user experience.

[0142] Table 1: Relationship between Face Database Size, Detection Accuracy, and Time Consumption

[0143] Facebank number 50 500 5000 50000 500000 Accuracy 99.6％ 99.2％ 98.5％ 97.3％ 95.7％ Time taken (ms) 0.3 2 21 186 1955

[0144] In this application embodiment, (1) before using face recognition, the solution is to obtain the member number of the merchant who is close to the location from the micro base station to create a face database in real time, which achieves the technical effect of reducing the number of face databases to be matched, thereby achieving fast and high-precision face recognition.

[0145] (2) By using the distance range given to members upon entry, merchants send marketing information to members; when members enter the store, service specialists are quickly notified to greet and guide them, thus achieving the effect of precise service.

[0146] (3) A gradient direction comparison vector similarity scheme was adopted, and the distance and direction of the feature vectors were calculated simultaneously, which achieved a technical effect that can better distinguish whether facial features are similar.

[0147] In this embodiment, members appearing around a merchant can be targeted for rapid and precise marketing; member profiles can be quickly obtained upon entry, enabling differentiated services to be provided to members; accurate and reasonable member spending upon entry can be achieved, increasing merchant revenue; and precise member marketing can be realized, providing a comprehensive solution for service-oriented merchants.

[0148] When members enter the coverage area of ​​a 5G micro base station, merchants can send different marketing messages to them through the 5G micro base station; select members facing the merchant's location to create a temporary facial recognition database; when members enter the store, the facial recognition system and the temporary database can quickly identify member information, such as name and consumption preferences, and notify service personnel to greet and guide them, achieving precise service; and provide facial recognition payment services at the time of payment to improve members' service perception.

[0149] It should be noted that although the steps of the method in this application are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0150] Based on the foregoing embodiments, this application provides an identification device, which includes the included modules and the units included in each module, which can be implemented by a processor; of course, it can also be implemented by specific logic circuits; in the implementation process, the processor can be a central processing unit (CPU), microprocessor (MPU), digital signal processor (DSP) or field programmable gate array (FPGA), etc.

[0151] Figure 7 This is a schematic diagram of the structure of the identification device according to an embodiment of this application, as shown below. Figure 7 As shown, the device 700 includes an extraction module 701, a comparison module 702, and an identification module 703, wherein: the extraction module 701 is used to extract features from the object to be identified in the acquired first image to obtain the feature vector of the first key feature point; the comparison module 702 is used to compare the difference between the feature vector of any first key feature point and at least one other first key feature point to obtain a first comparison result corresponding to the any first key feature point; the identification module 703 is used to use the first comparison result corresponding to each first key feature point as the feature of the object to be identified, and to identify whether the object to be identified belongs to a first service type based at least on the feature vector of each first key feature point and the corresponding first comparison result.

[0152] In some embodiments, the device 700 further includes an output module, configured to output a first prompt message if the object to be identified belongs to a first service type; wherein the first prompt message is used to indicate that a first service is provided for the object to be identified; the output module is further configured to output a second prompt message if the object to be identified belongs to a second service type; wherein the second prompt message is used to indicate that a second service is provided for the object to be identified, the first service type and the second service type are different, and the second service is different from the first service.

[0153] In some embodiments, the apparatus 700 further includes an acquisition module and a determination module. The acquisition module is configured to acquire a second comparison result corresponding to a second key feature point of a first object belonging to a first service type in a database; wherein the second comparison result is obtained by comparing the difference between the feature vectors of any second key feature point of the first object and at least one other second key feature point of the first object; the comparison module 702 is configured to compare the difference between the feature vectors of any first key feature point of the object to be identified and the feature vectors of the second key feature points of the first object to obtain a third comparison result corresponding to any first key feature point; the determination module is configured to use the second comparison results corresponding to each second key feature point as features of the first object, and determine the matching degree between the object to be identified and the first object based on each first comparison result, each second comparison result and each third comparison result; the determination module is further configured to determine whether the object to be identified belongs to the first service type based on the matching degree.

[0154] In some embodiments, the determining module is further configured to determine a first degree of difference between the object to be identified and the first object based on each of the first comparison results and each of the second comparison results; determine a second degree of difference between the object to be identified and the first object based on each of the third comparison results; and determine a degree of matching between the object to be identified and the first object based on the first degree of difference and the second degree of difference.

[0155] In some embodiments, the determining module is further configured to determine, from all the obtained matching degrees, a target matching degree that represents the highest similarity between the object to be identified and the first object; determine whether the target matching degree satisfies a first condition; and if the target matching degree satisfies the first condition, determine that the object to be identified belongs to the first service type.

[0156] In some embodiments, the device 700 further includes a detection module and a selection module. The detection module is used to detect the identity and movement direction of a second object within a specific physical area. The selection module is used to select the first object from the second objects based on the identity and movement direction of the second object. An extraction module 701 is used to extract features from the first object to obtain feature vectors of the second key feature points. A comparison module 702 is used to compare the difference between the feature vectors of any second key feature point of the first object and at least one other second key feature point of the first object to obtain a second comparison result corresponding to the any second key feature point. The identity of the first object, the feature vectors of the second key feature points of the first object, and the second comparison results corresponding to the second key feature points are taken as a feature group, and the database is obtained based on the feature group of each first object.

[0157] In some embodiments, the selection module is further configured to select a target object belonging to a first service type from the second objects based on the identity identifier of the second object; the determination module is further configured to determine whether the target object is gradually approaching the service provision point based on the movement direction of the target object; if so, the target object is selected as the first object.

[0158] The descriptions of the above device embodiments are similar to those of the above method embodiments, and have similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0159] It should be noted that, in the embodiments of this application... Figure 7 The module division shown in the identification device is illustrative and represents only one logical functional division; in actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, exist as separate physical units, or be integrated into one unit with two or more units. The integrated units can be implemented in hardware, as software functional units, or a combination of both.

[0160] It should be noted that, in the embodiments of this application, if the above-described methods are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, 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 an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0161] This application provides an electronic device. Figure 8 This is a schematic diagram of the hardware entity of the electronic device according to an embodiment of this application, such as... Figure 8 As shown, the electronic device 800 includes a memory 801 and a processor 802. The memory 801 stores a computer program that can run on the processor 802. When the processor 802 executes the program, it implements the steps in the method provided in the above embodiments.

[0162] It should be noted that the memory 801 is configured to store instructions and applications executable by the processor 802, and can also cache data to be processed or already processed (e.g., image data, audio data, voice communication data and video communication data) in the processor 802 and various modules in the electronic device 800. It can be implemented by flash memory or random access memory (RAM).

[0163] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method provided in the above embodiments.

[0164] This application provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the steps in the method provided in the above-described method embodiments.

[0165] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium, storage medium, and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0166] It should be understood that the phrases "one embodiment," "an embodiment," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment," "in one embodiment," or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The descriptions of the various embodiments above tend to emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, they will not be repeated here.

[0167] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three kinds of relationships. For example, object A and / or object B can represent three situations: object A exists alone, object A and object B exist simultaneously, and object B exists alone.

[0168] It should be noted that, in this document, 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. Unless otherwise specified, 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 that element.

[0169] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple modules or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or modules can be electrical, mechanical, or other forms.

[0170] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0171] In addition, each functional module in the various embodiments of this application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the integrated modules can be implemented in hardware or in the form of hardware plus software functional units.

[0172] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0173] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, 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 an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0174] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0175] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0176] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0177] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A recognition method, characterized in that, The method includes: Feature extraction is performed on the object to be identified in the first acquired image to obtain the feature vector of the first key feature point; Compare the differences between the feature vectors of any first key feature point and at least one other first key feature point to obtain a first comparison result corresponding to any first key feature point; The first comparison result corresponding to each of the first key feature points is used as the feature of the object to be identified. Based at least on the feature vector of each of the first key feature points and the corresponding first comparison result, it is identified whether the object to be identified belongs to the first service type. The step of identifying whether the object to be identified belongs to the first service type based at least on the feature vectors of each of the key feature points and the corresponding first comparison results includes: Obtain a second comparison result corresponding to a second key feature point of a first object belonging to a first service type in the database; wherein, the second comparison result is obtained by comparing the difference between the feature vectors of any second key feature point of the first object and at least one other second key feature point of the first object; Compare the difference between the feature vectors of any first key feature point of the object to be identified and the feature vectors of the second key feature points of the first object to obtain a third comparison result corresponding to any first key feature point; the number of first key feature points is the same as the number of second key feature points. The second comparison result corresponding to each of the second key feature points is used as the feature of the first object. Based on each of the first comparison results, each of the second comparison results and each of the third comparison results, the matching degree between the object to be identified and the first object is determined. Based on the matching degree, it is determined whether the object to be identified belongs to the first service type.

2. The method according to claim 1, characterized in that, The method further includes: If the object to be identified belongs to the first service type, a first prompt message is output; wherein, the first prompt message is used to indicate that the first service is provided to the object to be identified; If the object to be identified belongs to the second service type, a second prompt message is output; wherein, the second prompt message is used to indicate that a second service is provided for the object to be identified, the first service type and the second service type are different, and the second service is different from the first service.

3. The method according to claim 1, characterized in that, The step of determining the matching degree between the object to be identified and the first object based on each of the first comparison results, each of the second comparison results, and each of the third comparison results includes: Based on each of the first comparison results and each of the second comparison results, a first degree of difference between the object to be identified and the first object is determined; Based on each of the third comparison results, a second degree of difference between the object to be identified and the first object is determined; Based on the first difference degree and the second difference degree, the matching degree between the object to be identified and the first object is determined.

4. The method according to claim 1, characterized in that, The step of determining whether the object to be identified belongs to the first service type based on the matching degree includes: From all the obtained matching scores, determine the target matching score that represents the highest similarity between the object to be identified and the first object; Determine whether the target matching degree meets specific conditions; If the target matching degree meets a specific condition, the object to be identified is determined to belong to the first service type.

5. The method according to claim 1, characterized in that, The database construction process includes: Detect the identity and movement direction of a second object within a specific physical area; Based on the identity identifier and movement direction of the second object, select the first object from the second object; Feature extraction is performed on the first object to obtain the feature vector of the second key feature point; By comparing the difference between the feature vectors of any second key feature point of the first object and at least one other second key feature point of the first object, a second comparison result corresponding to the any second key feature point is obtained. The database is obtained by taking the identity identifier of the first object, the feature vector of the second key feature point of the first object, and the second comparison result corresponding to the second key feature point as a feature group, and then taking the feature group of each first object.

6. The method according to claim 5, characterized in that, The step of selecting the first object from the second object based on the identity identifier and movement direction of the second object includes: Based on the identity identifier of the second object, select the target object belonging to the first service type from the second object; Based on the movement direction of the target object, determine whether the target object is gradually approaching the service provision point; if so, designate the target object as the first object.

7. The method according to claim 6, characterized in that, The method further includes: Obtain marketing information and send the marketing information to the target audience.

8. An identification device, characterized in that, include: The extraction module is used to extract features from the object to be identified in the first acquired image to obtain the feature vector of the first key feature point. The comparison module is used to compare the difference between the feature vectors of any first key feature point and at least one other first key feature point to obtain a first comparison result corresponding to any first key feature point. The identification module is used to take the first comparison result corresponding to each of the first key feature points as the feature of the object to be identified, and to identify whether the object to be identified belongs to the first service type based at least on the feature vector of each of the first key feature points and the corresponding first comparison result. The acquisition module is used to acquire a second comparison result corresponding to a second key feature point of a first object belonging to a first service type in the database; wherein, the second comparison result is obtained by comparing the difference between the feature vectors of any second key feature point of the first object and at least one other second key feature point of the first object; The comparison module is used to compare the difference between the feature vectors of any first key feature point of the object to be identified and the feature vectors of the second key feature points of the first object, and to obtain a third comparison result corresponding to any first key feature point; the number of first key feature points is the same as the number of second key feature points. The determination module is used to take the second comparison result corresponding to each of the second key feature points as the feature of the first object, and determine the matching degree between the object to be identified and the first object based on each of the first comparison results, each of the second comparison results and each of the third comparison results; The determining module is further configured to determine whether the object to be identified belongs to the first service type based on the matching degree.

9. An electronic device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the program, it implements the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images