Target recognition method, device, system and computer readable storage medium

Abstract

The application discloses a target identification method, device and system and a computer readable storage medium. The method comprises the following steps: receiving target data to be identified sent by a terminal device; wherein the target data to be identified comprises target characteristics to be identified and target attributes; determining a target storage space corresponding to the target attributes from a plurality of storage spaces; wherein at least one first reference characteristic corresponding to the attributes is pre-stored in each storage space; comparing the target characteristics to be identified with at least one first reference characteristic in the target storage space to obtain at least one target comparison result; determining a target identification result according to the at least one target comparison result, and sending the target identification result to the terminal device. In the above manner, the comparison efficiency can be improved.

Description

Technical Field

[0001] This application relates to the field of target recognition technology, and in particular to target recognition methods, apparatus, systems and computer-readable storage media. Background Technology

[0002] Currently, target recognition technologies, such as facial recognition, generally employ neural network models to extract facial features, and then compare these features to achieve facial recognition. Using the aforementioned facial feature extraction algorithms, further facial comparison can be performed.

[0003] The drawback is that, during the comparison process, the feature to be compared is usually compared with all benchmark features, and the output is the similarity between the two features. Summary of the Invention

[0004] This application provides a target identification method, apparatus, system, and computer-readable storage medium, which can improve comparison efficiency.

[0005] One technical solution adopted in this application is to provide a target recognition method, which includes: receiving target data to be identified sent by a terminal device; wherein the target data to be identified includes target features and target attributes; determining a target storage space corresponding to the target attribute from multiple storage spaces; wherein each storage space pre-stores at least one first reference feature corresponding to the attribute; comparing the target features to be identified with at least one first reference feature in the target storage space to obtain at least one target comparison result; determining a target recognition result based on the at least one target comparison result, and sending the target recognition result to the terminal device.

[0006] The method of comparing the target feature to be identified with at least one first reference feature in the target storage space to obtain at least one target comparison result includes: determining whether a second reference feature exists in the target storage space; wherein each second reference feature is associated with a first reference feature; if so, comparing the target feature to be identified with at least one first reference feature in the target storage space to obtain at least one first comparison result; and comparing the target feature to be identified with the second reference feature to obtain at least one second comparison result; and obtaining at least one target comparison result based on the correlation between at least one first comparison result and at least one second comparison result.

[0007] Specifically, obtaining at least one target alignment result based on the correlation between at least one first alignment result and at least one second alignment result includes: taking a weighted average of the correlated first alignment results and the first alignment result to obtain a target alignment result.

[0008] The target recognition result includes successful or unsuccessful recognition. After determining the target recognition result based on at least one target comparison result, the process includes: if the target recognition result is successful, determining the best target comparison result among the at least one target comparison results; determining whether there is an associated second benchmark feature for the first benchmark feature corresponding to the best target comparison result; if so, updating the first benchmark feature or the second benchmark feature using the target feature to be recognized; if not, using the target feature to be recognized as the second benchmark feature.

[0009] The process of updating the first or second benchmark feature using the target feature to be identified includes: comparing the first comparison result and the second comparison result; if the first comparison result is higher than the second comparison result, replacing the second benchmark feature with the target feature to be identified; if the first comparison result is lower than the second comparison result, replacing the first benchmark feature with the target feature to be identified.

[0010] After determining the target recognition result based on at least one target comparison result, the method further includes: if the target recognition result is a recognition failure, then the feature of the target to be recognized is stored as the first reference feature in the target storage space.

[0011] The method further includes: if there is no target storage space corresponding to the target attribute among multiple storage spaces, comparing the target feature to be identified with at least one first reference feature in each storage space to obtain at least one target comparison result.

[0012] Before receiving the target data to be identified sent by the terminal device, the process includes: receiving the image to be processed sent by the terminal device; extracting features from the image to be processed to obtain the features to be processed; wherein the features to be processed are 128-dimensional features; and storing the features to be processed that meet the image quality requirements as the first reference features in the corresponding storage space.

[0013] Another technical solution adopted in this application is to provide a target recognition device, which includes: a processor; a memory connected to the processor, the memory being used to store a computer program, and the memory including a plurality of storage spaces, each storage space pre-storing at least one first reference feature corresponding to an attribute; and a communication module connected to the processor for communicating with a terminal device; wherein the processor is used to execute the computer program to control the memory and the communication module to implement the method provided by the above technical solution.

[0014] Another technical solution adopted in this application is to provide a target recognition system, which includes: a terminal device; and a target recognition device, which is communicatively connected to the terminal device, and the target recognition device is the same as the target recognition device provided in the above technical solution.

[0015] Another technical solution adopted in this application is to provide a computer-readable storage medium for storing a computer program, which, when executed by a processor, is used to implement the method provided by the above technical solution.

[0016] The beneficial effects of this application are as follows: Unlike the prior art, the target identification method of this application pre-stores at least one first reference feature corresponding to the attribute in the storage space, so that when receiving the target data to be identified sent by the terminal device, the target storage space corresponding to the target attribute is determined from multiple storage spaces, and the target feature to be identified in the target data to be identified is compared with at least one first reference feature in the target storage space to obtain at least one target comparison result. This reduces the number of comparisons between the target feature to be identified and the first reference feature, improves the comparison efficiency, and thus enables the target identification result to be obtained quickly. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0018] Figure 1 This is a flowchart illustrating the first embodiment of the target recognition method provided in this application;

[0019] Figure 2 This is a schematic diagram illustrating an application scenario of the target recognition method provided in this application;

[0020] Figure 3 This is a flowchart illustrating the second embodiment of the target recognition method provided in this application;

[0021] Figure 4 This is a flowchart illustrating the third embodiment of the target recognition method provided in this application;

[0022] Figure 5 This is a flowchart illustrating an embodiment of step 43 provided in this application;

[0023] Figure 6 This is a flowchart illustrating the fourth embodiment of the target recognition method provided in this application;

[0024] Figure 7 This is a schematic diagram of the structure of an embodiment of the target recognition device provided in this application;

[0025] Figure 8 This is a schematic diagram of the structure of an embodiment of the target recognition system provided in this application;

[0026] Figure 9 This is a schematic diagram of an embodiment of the computer-readable storage medium provided in this application. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It is understood that the specific embodiments described herein are only for explaining this application and not for limiting it. Furthermore, it should be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all structures. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0028] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0029] See Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the target recognition method provided in this application. The method includes:

[0030] Step 11: Receive the target data to be identified sent by the terminal device; wherein, the target data to be identified includes the target features and target attributes.

[0031] In some embodiments, the target data to be identified may be generated based on a face image or a human eye image.

[0032] Step 12: Determine the target storage space corresponding to the target attribute from multiple storage spaces; wherein each storage space pre-stores at least one first baseline feature corresponding to the attribute.

[0033] In some embodiments, since the identification process essentially involves comparing the target feature to be identified with a large number of first reference features, the time required for comparison can be long if the number of first reference features stored in the storage space is large. Therefore, these first reference features are classified according to their corresponding attributes so that different categories of first reference features are stored in a storage space.

[0034] For example, this attribute could be age, which means dividing the first baseline feature into different age groups, with each age group corresponding to a storage space. For instance, if 0-15 years old is divided into the first age group, 15-35 years old into the second age group, 35-60 years old into the third age group, and 60 years old and above into the fourth age group, then there will be four storage spaces.

[0035] For example, this attribute could be age or whether or not glasses are worn. The first baseline feature is divided into different age groups, with each age group corresponding to a separate storage space. For instance, if 0-15 years old is the first age group, 15-35 years old the second, 35-60 years old the third, and over 60 years old the fourth, then there will be four storage spaces. Within each storage space, the first baseline feature is further divided based on whether or not glasses are worn, resulting in a first baseline feature corresponding to wearing glasses and a first baseline feature corresponding to not wearing glasses.

[0036] In some embodiments, each storage space may correspond to a database.

[0037] Step 13: Compare the features of the target to be identified with at least one first reference feature in the target storage space to obtain at least one target comparison result.

[0038] In some embodiments, the comparison result can be the similarity between the target feature to be identified and the first reference feature. Specifically, a traversal approach can be used to compare the target feature to be identified with each of the first reference features in the target storage space to obtain the corresponding similarity.

[0039] In some embodiments, if there is no target storage space corresponding to the target attribute among the multiple storage spaces, the target feature to be identified is compared with at least one first reference feature in each storage space to obtain at least one target comparison result.

[0040] Step 14: Determine the target recognition result based on at least one target comparison result, and send the target recognition result to the terminal device.

[0041] In some embodiments, the target comparison result with the highest similarity among at least one target comparison result is determined. If the target comparison result with the highest similarity meets the preset conditions, the recognition is determined to be successful. The successful recognition is then taken as the target recognition result and the target recognition result is sent to the terminal device.

[0042] If the target comparison result with the highest similarity does not meet the preset conditions, the recognition is determined to be a failure. The recognition failure is then taken as the target recognition result and sent to the terminal device.

[0043] In some embodiments, sending the successfully identified target recognition result to the terminal device will display a success message on the terminal device. Sending the unsuccessfully identified target recognition result to the terminal device will display a failure message on the terminal device, requesting that data be collected again.

[0044] In one application scenario, combined Figure 2 To explain:

[0045] The terminal device acquires an image of the target to be identified and extracts features from the image to obtain target data. The terminal device then sends the target data to the target recognition device.

[0046] The target identification device receives target data to be identified sent by the terminal device; the target data to be identified includes target features and target attributes.

[0047] The target recognition device determines the target storage space corresponding to the target attribute from multiple storage spaces; wherein, each storage space pre-stores at least one first reference feature corresponding to the attribute; compares the target feature to be recognized with at least one first reference feature in the target storage space to obtain at least one target comparison result; and determines the target recognition result based on at least one target comparison result.

[0048] The target recognition device sends the target recognition results to the terminal device.

[0049] In this embodiment, by pre-storing at least one first reference feature corresponding to an attribute in the storage space, when receiving target data to be identified sent by the terminal device, the target storage space corresponding to the target attribute is determined from multiple storage spaces. The target feature to be identified in the target data is compared with at least one first reference feature in the target storage space to obtain at least one target comparison result. This reduces the number of comparisons between the target feature to be identified and the first reference feature, improves comparison efficiency, and enables the target identification result to be obtained quickly, so that the terminal device can quickly receive the target identification result and improve the user experience.

[0050] See Figure 3 , Figure 3 This is a flowchart illustrating a second embodiment of the target recognition method provided in this application. The method includes:

[0051] Step 31: Receive the target data to be identified sent by the terminal device; wherein, the target data to be identified includes the target features and target attributes.

[0052] Step 32: Determine the target storage space corresponding to the target attribute from multiple storage spaces; wherein each storage space pre-stores at least one first baseline feature corresponding to the attribute.

[0053] Steps 31 and 32 have the same or similar technical solutions as those in the above embodiments, and will not be described in detail here.

[0054] Step 33: Determine whether a second baseline feature exists in the target storage space; wherein each second baseline feature is associated with a first baseline feature.

[0055] The second reference feature can be obtained when the target recognition result is successful. Specifically, the first reference feature corresponding to successful recognition is determined, and then the target feature to be recognized is associated with the first reference feature corresponding to successful recognition as the second reference feature and stored in the target storage space.

[0056] If yes, proceed to step 34; otherwise, it means that the target feature to be identified does not have a corresponding first reference feature in the target storage space, or the original image of the target feature to be identified does not meet the image quality requirements, so the target feature to be identified does not have much feature information and therefore cannot be compared.

[0057] Step 34: Compare the features of the target to be identified with at least one first reference feature in the target storage space to obtain at least one first comparison result.

[0058] In some embodiments, the first comparison result may be the similarity between the target feature to be identified and the first reference feature. Specifically, a traversal approach can be used to compare the target feature to be identified with each of the first reference features in the target storage space to obtain the corresponding similarity.

[0059] Step 35: Compare the features of the target to be identified with the second benchmark features to obtain at least one second comparison result.

[0060] In some embodiments, the second comparison result may be the similarity between the target feature to be identified and the second benchmark feature. Specifically, a traversal approach can be used to compare the target feature to be identified with each second benchmark feature in the target storage space to obtain the corresponding similarity.

[0061] Step 36: Based on the correlation between at least one first alignment result and at least one second alignment result, obtain at least one target alignment result.

[0062] Since each second baseline feature is associated with a first baseline feature, the first and second alignment results are correlated. Therefore, the correlated first alignment results need to be processed to obtain the corresponding target alignment result. For example, a weighted average can be taken from the correlated first alignment results to obtain a target alignment result. The weighting coefficients can all be 1, or they can be divided according to the image quality of the second and first baseline features, with higher coefficients for higher image quality and lower coefficients for lower image quality.

[0063] If the first benchmark feature has no associated second benchmark feature, then the first comparison result corresponding to the first benchmark feature is directly used as the target comparison result.

[0064] Step 37: Determine the target recognition result based on at least one target comparison result, and send the target recognition result to the terminal device.

[0065] In some embodiments, the target comparison result with the highest similarity among at least one target comparison result is determined. If the target comparison result with the highest similarity meets the preset conditions, the recognition is determined to be successful. The successful recognition is then taken as the target recognition result and the target recognition result is sent to the terminal device.

[0066] If the target comparison result with the highest similarity does not meet the preset conditions, the recognition is determined to be a failure. The recognition failure is then taken as the target recognition result and sent to the terminal device.

[0067] In some embodiments, sending the successfully identified target recognition result to the terminal device will display a success message on the terminal device. Sending the unsuccessfully identified target recognition result to the terminal device will display a failure message on the terminal device, requesting that data be collected again.

[0068] In some embodiments, after determining the target recognition result based on at least one target comparison result, refer to Figure 4 This can include the following processes:

[0069] Step 41: If the target recognition result is successful, then determine the best target comparison result among at least one target comparison result.

[0070] The best target alignment result is the target alignment result with the highest similarity among at least one target alignment results.

[0071] Step 42: Determine whether there is a related second benchmark feature for the first benchmark feature corresponding to the best target alignment result.

[0072] The optimal target alignment result may be obtained directly from the first benchmark feature, or it may be obtained from both the first and second benchmark features. Therefore, it is necessary to determine whether the first benchmark feature corresponding to the optimal target alignment result has a related second benchmark feature.

[0073] If the first benchmark feature corresponding to the best target alignment result has an associated second benchmark feature, proceed to step 43; if the first benchmark feature corresponding to the best target alignment result does not have an associated second benchmark feature, proceed to step 44.

[0074] Step 43: Update the first or second baseline features using the features of the target to be identified.

[0075] When it is determined that the first reference feature corresponding to the best target alignment result has a related second reference feature, it indicates that the best target alignment result is obtained from the first reference feature and the second reference feature. Therefore, the first or second reference feature corresponding to the best target alignment result can be updated using the features of the target to be identified.

[0076] In some embodiments, see Figure 5 Step 43 can be the following process:

[0077] Step 51: Compare the first alignment result with the second alignment result.

[0078] The first comparison result is obtained by comparing the feature of the target to be identified with the first benchmark feature, and the second comparison result is obtained by comparing the feature of the target to be identified with the second benchmark feature corresponding to the first benchmark feature.

[0079] Step 52: If the first comparison result is higher than the second comparison result, replace the second baseline feature with the feature of the target to be identified.

[0080] If the first comparison result is higher than the second comparison result, it means that the target feature to be identified has more feature information than the second reference feature, and the target feature to be identified is used to replace the second reference feature.

[0081] Step 53: If the first comparison result is lower than the second comparison result, replace the first baseline feature with the feature of the target to be identified.

[0082] If the first comparison result is lower than the second comparison result, it means that the target feature to be identified has more feature information than the first reference feature, and the target feature to be identified is used to replace the first reference feature.

[0083] Step 44: Use the features of the target to be identified as the second baseline features.

[0084] If the first benchmark feature corresponding to the best target alignment result does not have an associated second benchmark feature, then the target feature to be identified can be used as the second benchmark feature to associate it with the first benchmark feature.

[0085] In other embodiments, if the target recognition result is a recognition failure, the feature of the target to be recognized is stored as a first reference feature in the target storage space.

[0086] In this embodiment, by pre-storing at least one first reference feature corresponding to an attribute in the storage space, when receiving target data to be identified sent by the terminal device, the target storage space corresponding to the target attribute is determined from multiple storage spaces. The target feature to be identified in the target data is compared with at least one first reference feature in the target storage space to obtain at least one target comparison result. This reduces the number of comparisons between the target feature to be identified and the first reference feature, improves comparison efficiency, and enables the target identification result to be obtained quickly, so that the terminal device can quickly receive the target identification result and improve the user experience.

[0087] Furthermore, by combining the first and second comparison results, the first or second benchmark feature can be dynamically updated, which is beneficial for obtaining the latest benchmark features in the storage space. This allows for real-time updates, which can improve the accuracy of subsequent comparisons.

[0088] In some embodiments, before receiving the target data to be identified sent by the terminal device, refer to Figure 6 , Figure 6 This is a flowchart illustrating the fourth embodiment of the target recognition method provided in this application. The method includes:

[0089] Step 61: Receive the image to be processed sent by the terminal device.

[0090] In some embodiments, the target recognition device needs to pre-store the first reference features, thus requiring the pre-collection of the image to be processed and the extraction of the corresponding first reference features from the image to be processed.

[0091] Step 62: Extract features from the image to be processed to obtain the features to be processed; the features to be processed are 128-dimensional features.

[0092] In some embodiments, a corresponding feature extraction network can be used to extract features to obtain the features to be processed. This feature extraction network can be built based on a convolutional neural network, or it can be built based on a residual network or a generative adversarial network.

[0093] Specifically, a corresponding feature extraction network is used to extract features, resulting in 512-dimensional floating-point features. Then, quantization techniques are used to reduce the dimensionality of the 512-dimensional features to 128-dimensional integer features.

[0094] Step 63: Store the features to be processed that meet the image quality requirements as the first reference features in the corresponding storage space.

[0095] When the feature to be processed is obtained, the corresponding attribute is also generated. When the feature to be processed meets the image quality requirements, the feature to be processed is stored as the first reference feature in the storage space corresponding to the attribute.

[0096] In this embodiment, the target recognition device uses quantization technology to extract 128-dimensional features from the image. Only the feature information data is saved, without involving the original image, which improves the privacy of the original data. Furthermore, the 128-dimensional features occupy less storage space, reducing storage requirements, and can accelerate model inference processing time. Additionally, by determining whether the image quality meets the requirements, a condition for storage is set, thereby improving the quality of the stored features and enhancing subsequent recognition accuracy.

[0097] See Figure 7 , Figure 7 This is a schematic diagram of an embodiment of the target identification device provided in this application. The target identification device 70 includes: a processor 71, a memory 72, and a communication module 73.

[0098] The memory 72 is connected to the processor 71. The memory 72 is used to store computer programs and includes multiple storage spaces, each of which stores at least one first reference feature corresponding to the attribute in advance.

[0099] The communication module 73 is connected to the processor 71 and is used to communicate with the terminal device.

[0100] The processor 71 is used to execute computer programs to control the memory 72 and the communication module 73, and to implement the following methods:

[0101] The system receives target data to be identified sent by a terminal device; wherein the target data to be identified includes target features and target attributes; determines a target storage space corresponding to the target attribute from multiple storage spaces; wherein each storage space pre-stores at least one first reference feature corresponding to the attribute; compares the target features to be identified with at least one first reference feature in the target storage space to obtain at least one target comparison result; determines the target identification result based on the at least one target comparison result, and sends the target identification result to the terminal device.

[0102] It is understood that when the processor 71 is used to execute computer programs, it is also used to implement the technical solutions of any embodiment of this application, which will not be elaborated here.

[0103] See Figure 8 , Figure 8 This is a schematic diagram of an embodiment of a computer-readable storage medium provided in this application. The target recognition system 80 includes: a terminal device 81 and a target recognition device 70.

[0104] The target identification device 70 is communicatively connected to the terminal device 81, and the target identification device 70 is the same as the target identification device 70 in the above embodiment.

[0105] In one application scenario, the terminal device 81 can be a face comparison terminal, and the target recognition device 70 can be a server. Therefore, the face comparison terminal remotely connects to the server. The server is used to store multiple comparison features.

[0106] First, images such as ID photos or live photos are sent to the server as the initial samples for database entry. The extracted 512-dimensional floating-point features are then reduced to 128-dimensional integer features using quantization techniques to reduce the storage length of the features and effectively reduce the storage overhead of the database on the server. Simultaneously, the image quality of the features to be stored is compared with a preset image quality to determine if it meets the storage requirements. If it does not meet the requirements, the user is prompted to re-upload the photo until the server determines it is acceptable before saving. At the same time, age information or specific identity information of the images is saved. Different age groups, such as 0-15, 15-35, 35-60, and over 60, are categorized into different attributes to identify the characteristics of different groups. The 128-dimensional features corresponding to the first sample are then stored in the databases of the servers corresponding to the different attributes.

[0107] Then, the face comparison terminal collects external facial features, along with attribute identifiers for the facial features, with age used as the attribute identifier here; the face comparison terminal sends a request to the server; based on the attribute identifier, it determines whether to traverse and search the database of all stored features.

[0108] Specifically, the server searches the database corresponding to the attribute identifier. If no match is found, it compares the baseline features with the facial features from all databases. If a match is found, it compares the baseline features with the facial features from the corresponding database.

[0109] The server receives facial features from the face comparison terminal and compares them with baseline features in the corresponding database, generating a comparison result. It then selects the comparison result with the highest similarity from the database. The server checks if the highest similarity score exceeds a set threshold. If it does, the recognition is successful, and the facial features are stored as secondary sample data in the corresponding database. If the comparison fails, the facial features are not saved.

[0110] At the same time, the server contains the first baseline features uploaded by the user and the second baseline features collected and saved by the face comparison terminal.

[0111] The feature information collected from the face comparison terminal is sent to the server. Based on the current face features, the server retrieves the first and second baseline features from the corresponding database and calculates their similarity. The weighted average of the two similarities is taken to obtain the final similarity score. The server then checks if the final similarity score is greater than a set threshold; if it is, the comparison is considered successful. If the similarity score of the first baseline feature is higher than that of the second baseline feature, the current face feature value replaces the second baseline feature for storage, and the original second baseline feature is deleted to update the database.

[0112] See Figure 9 , Figure 9 This is a schematic diagram of an embodiment of the computer-readable storage medium provided in this application. The computer-readable storage medium 90 is used to store a computer program 91. When the computer program 91 is executed by a processor, it is used to implement the following method steps:

[0113] The system receives target data to be identified sent by a terminal device; wherein the target data to be identified includes target features and target attributes; determines a target storage space corresponding to the target attribute from multiple storage spaces; wherein each storage space pre-stores at least one first reference feature corresponding to the attribute; compares the target features to be identified with at least one first reference feature in the target storage space to obtain at least one target comparison result; determines the target identification result based on the at least one target comparison result, and sends the target identification result to the terminal device.

[0114] It is understood that when computer program 91 is executed by the processor, it is also used to implement the technical solutions of any embodiment of this application.

[0115] The technical solution of this application can be applied to scenarios such as security deployment, intelligent building access control systems, and unified authentication.

[0116] In summary, the technical solution of this application does not involve image information of the original data during the relevant identity verification process, but only involves feature data, which can improve the privacy of the original data. Feature values ​​are extracted by uploading the first sample data through the terminal device, and quantization technology is used to reduce storage space. Attribute identifiers are used to distinguish databases storing different types of features. When comparing the current facial feature values ​​sampled by the terminal device, the database range for comparison is selected according to the attribute identifiers for accurate comparison and improved comparison efficiency. Furthermore, when both the first and second benchmark features exist simultaneously, the comparison results can be used to update the first or second benchmark features in real time, which helps to improve the accuracy of feature comparison.

[0117] In the several embodiments provided in this application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0118] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0119] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0120] If the integrated units in the other embodiments described above 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 this application, in essence, or the part that contributes to the prior art, or all or 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.) or processor to execute all or part of the steps 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), random access memory (RAM), magnetic disks, or optical disks.

[0121] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A target recognition method, characterized in that, The method includes: The device receives target data to be identified sent by a terminal device; wherein the target data to be identified includes target features and target attributes. A target storage space corresponding to the target attribute is determined from multiple storage spaces; wherein each storage space pre-stores at least one first benchmark feature corresponding to the attribute, and each first benchmark feature is classified according to its corresponding attribute so that first benchmark features of the same category are stored in one storage space. The method of comparing the target feature to be identified with at least one first reference feature in the target storage space to obtain at least one target comparison result includes: determining whether a second reference feature exists in the target storage space; wherein each second reference feature is associated with one first reference feature; when a first reference feature is successfully identified, the successfully identified target feature is used as a second reference feature associated with the successfully identified first reference feature; if so, the target feature to be identified is compared with at least one first reference feature in the target storage space to obtain at least one first comparison result; and the target feature to be identified is compared with the second reference feature to obtain at least one second comparison result; and at least one target comparison result is obtained based on the correlation between the at least one first comparison result and the at least one second comparison result. The target identification result is determined based on the at least one target comparison result, and the target identification result is sent to the terminal device.

2. The method according to claim 1, characterized in that, The step of obtaining at least one target alignment result based on the correlation between the at least one first alignment result and the at least one second alignment result includes: A weighted average is taken from the first alignment result and the first alignment result that are related to each other to obtain a target alignment result.

3. The method according to claim 2, characterized in that, The target recognition result includes successful or unsuccessful recognition; after determining the target recognition result based on the at least one target comparison result, the process includes: If the target identification result is successful, then the best target comparison result among at least one target comparison result is determined; Determine whether the first benchmark feature corresponding to the best target alignment result has a related second benchmark feature; If so, update the first reference feature or the second reference feature using the target feature to be identified; If not, the feature of the target to be identified shall be used as the second reference feature.

4. The method according to claim 3, characterized in that, The step of updating the first reference feature or the second reference feature using the features of the target to be identified includes: Compare the first alignment result with the second alignment result; If the first comparison result is higher than the second comparison result, the second reference feature is replaced with the target feature to be identified; If the first comparison result is lower than the second comparison result, the first benchmark feature is replaced with the target feature to be identified.

5. The method according to claim 2, characterized in that, After determining the target recognition result based on the at least one target comparison result, the method further includes: If the target recognition result is a recognition failure, the feature of the target to be recognized is stored as the first reference feature in the target storage space.

6. The method according to claim 1, characterized in that, The method further includes: If there is no target storage space corresponding to the target attribute among multiple storage spaces, the target feature to be identified is compared with at least one first reference feature in each storage space to obtain at least one target comparison result.

7. The method according to claim 1, characterized in that, Before receiving the target data to be identified sent by the receiving terminal device, the following is included: Receive the image to be processed sent by the terminal device; Feature extraction is performed on the image to be processed to obtain the features to be processed; wherein, the features to be processed are 128-dimensional features; The features to be processed that meet the image quality requirements are stored as the first reference features in the corresponding storage space.

8. A target recognition device, characterized in that, The target identification device includes: processor; A memory connected to the processor, the memory being used to store computer programs, and the memory including a plurality of storage spaces, each storage space pre-stores at least one first reference feature corresponding to an attribute, wherein each first reference feature is classified according to its corresponding attribute so that first reference features of the same category are stored in one storage space. A communication module, connected to the processor, is used to communicate with the terminal device; The processor is used to execute the computer program to control the memory and the communication module, and to implement the method as described in any one of claims 1-7.

9. A target recognition system, characterized in that, The target recognition system includes: Terminal equipment; A target identification device, which is communicatively connected to the terminal device, is the target identification device as described in claim 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program, which, when executed by a processor, is used to implement the method as described in any one of claims 1-7.

Images