CN116189028B - Image recognition method, device, electronic equipment and storage medium - Google Patents

Abstract

The disclosure discloses an image recognition method, an image recognition device, electronic equipment and a storage medium. Relates to the technical field of artificial intelligence, in particular to the technical field of deep learning and computer vision, and comprises the following specific implementation schemes: obtaining a time sequence image sequence to be identified according to the video image; responding to the received selection operation aiming at the target identification task, and obtaining a target processing strategy; and processing the time sequence image sequence to be identified based on the target processing strategy to obtain an identification result.

Description

Image recognition method, device, electronic device and storage medium

Technical Field

The present disclosure relates to the field of artificial intelligence technology, in particular to the field of deep learning and computer vision technology, and is applied to the field of image recognition technology, and specifically to an image recognition method, device, electronic device, and storage medium.

Background technique

Computer vision technology refers to the use of image acquisition equipment and computers to replace human eyes to identify, track and measure targets, and further perform image processing. The images processed in computer vision technology can include static images and dynamic images.

With the development of artificial intelligence technology, its application in computer vision is becoming more and more extensive, for example: image recognition based on artificial intelligence technology.

Summary of the invention

The present disclosure provides an image recognition method, an apparatus, an electronic device, and a storage medium.

According to one aspect of the present disclosure, there is provided an image recognition method, comprising:

According to the video image, a time-series image sequence to be identified is obtained;

In response to receiving a selection operation for a target recognition task, obtaining a target processing strategy; and

Based on the target processing strategy, the time-series image sequence to be recognized is processed to obtain the recognition result.

According to another aspect of the present disclosure, an image recognition device is provided, including: a first obtaining module, a second obtaining module, and a third obtaining module.

A first acquisition module is used to obtain a time-series image sequence to be identified based on the video image;

A second obtaining module is configured to obtain a target processing strategy in response to receiving a selection operation for a target recognition task; and

The third acquisition module is used to process the time-series image sequence to be recognized based on the target processing strategy to obtain the recognition result.

According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as described above.

According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the method as described above.

According to another aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein the computer program implements the method described above when executed by a processor.

It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

FIG1 schematically shows an exemplary system architecture to which an image recognition method and apparatus can be applied according to an embodiment of the present disclosure;

FIG2 schematically shows a flow chart of an image recognition method according to an embodiment of the present disclosure;

FIG3 schematically shows a schematic diagram of obtaining a target processing strategy in response to receiving a selection operation for a target recognition task according to an embodiment of the present disclosure;

FIG4 schematically shows a schematic diagram of processing spatial relationship features to obtain recognition results according to an embodiment of the present disclosure;

FIG5 schematically shows a schematic diagram of processing the action features of the first target object to obtain a recognition result according to an embodiment of the present disclosure;

FIG6 schematically shows a schematic diagram of obtaining a recognition result by processing the position features of the target area and the position features of the first target object according to an embodiment of the present disclosure;

FIG7 schematically shows a schematic diagram of obtaining a recognition result by processing the position feature of the first target object and the position feature of the second target object according to an embodiment of the present disclosure;

FIG8 schematically shows a schematic diagram of processing the features of the first target object and the features of the third target object to obtain a recognition result according to an embodiment of the present disclosure;

FIG9 schematically shows an overall exemplary system architecture diagram corresponding to different processing strategies according to an embodiment of the present disclosure;

FIG10 schematically shows an exemplary system architecture deployment diagram according to an embodiment of the present disclosure;

FIG. 11 schematically shows a block diagram of an image recognition device according to an embodiment of the present disclosure; and

FIG. 12 schematically shows a block diagram of an electronic device suitable for implementing an image recognition method according to an embodiment of the present disclosure.

Detailed ways

The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

With the development of artificial intelligence technology, models that meet specific tasks can be obtained based on network structure algorithms for specific tasks and model training for target recognition. For example: target detection model. However, this specific task model is limited to specific task scenarios and is difficult to meet the needs of complex application scenarios. For example: in smart security scenarios, scenario requirements may include the recognition requirements of target objects, the recognition requirements of abnormal behaviors of target objects, the recognition requirements of the movement trajectories of target objects, etc.

In view of this, an embodiment of the present disclosure provides an image recognition method, comprising: obtaining a time-series image sequence to be recognized based on a video image; obtaining a target processing strategy in response to receiving a selection operation for a target recognition task; and processing the time-series image sequence to be recognized based on the target processing strategy to obtain a recognition result.

FIG. 1 schematically shows an exemplary system architecture to which an image recognition method and apparatus can be applied according to an embodiment of the present disclosure.

It should be noted that FIG. 1 is only an example of a system architecture to which the embodiments of the present disclosure can be applied, in order to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios. For example, in another embodiment, the exemplary system architecture to which the image recognition method and apparatus can be applied may include a terminal device, but the terminal device may implement the image recognition method and apparatus provided by the embodiments of the present disclosure without interacting with the server.

As shown in Fig. 1, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103 and the server 105. The network 104 may include various connection types, such as wired and/or wireless communication links, etc.

The user can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications can be installed on the terminal devices 101, 102, and 103, such as knowledge reading applications, web browser applications, search applications, instant messaging tools, email clients, and/or social platform software, etc. (only as examples).

The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for the content browsed by the user using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

It should be noted that the image recognition method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the image recognition device provided in the embodiment of the present disclosure can generally be set in the server 105. The image recognition method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and/or the server 105. Accordingly, the image recognition device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and/or the server 105.

For example, when a user selects a target recognition task online, the first terminal device 101, the second terminal device 102, and the third terminal device 103 can obtain the identification information of the target recognition task, and then send the identification information of the target recognition task and the video image to the server 105, and the server 105 obtains the time-series image sequence to be recognized based on the video image; analyzes the identification information of the target recognition task to determine the target processing strategy; and processes the time-series image sequence to be recognized based on the target processing strategy to obtain the recognition result. Alternatively, the server or server cluster that can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and/or the server 105 determines the target processing strategy, and processes the time-series image sequence to be recognized based on the target processing strategy to obtain the recognition result.

The image recognition method provided in the embodiment of the present disclosure can also be generally performed by the first terminal device 101, the second terminal device 102, and the third terminal device 103. Accordingly, the image recognition device provided in the embodiment of the present disclosure can generally be set in the first terminal device 101, the second terminal device 102, and the third terminal device 103.

It should be understood that the number of terminal devices, networks and servers in Figure 1 is only illustrative. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision, disclosure and application of user personal information involved comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good morals.

In the technical solution of the present disclosure, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

FIG. 2 schematically shows a flow chart of an image recognition method according to an embodiment of the present disclosure.

As shown in FIG. 2 , the method includes operations S210 to S230 .

In operation S210, a time-sequential image sequence to be identified is obtained according to a video image.

In operation S220, in response to receiving a selection operation for a target recognition task,

Target processing strategy.

In operation S230 , the time-series image sequence to be recognized is processed based on the target processing strategy to obtain a recognition result.

According to an embodiment of the present disclosure, the video image may be a video image acquired offline or a video image acquired in real time.

According to the embodiments of the present disclosure, based on the video image, the time-series image sequence to be identified can be obtained by collecting each frame of the image frame. The collection cycle can also be set according to actual business needs, for example, the time-series image sequence to be identified can be obtained by collecting every 3 to 10 frames.

According to an embodiment of the present disclosure, the target recognition task may include any one of the following: an abnormal behavior recognition task, a target attribute feature recognition task, a target object behavior trajectory recognition task, etc.

According to the embodiments of the present disclosure, the user can use the visual interface of the client to interact with the server to obtain the target processing strategy. For example, the user selects the recognition task of abnormal behavior Db ₁ in the visual interface of the client, and the server can obtain the target processing strategy corresponding to the abnormal behavior Db ₁ based on the information of the abnormal behavior Db ₁ .

According to the embodiments of the present disclosure, a target processing path can be determined based on a target processing strategy, and an image feature sequence to be identified is processed based on the target processing path to obtain an abnormal behavior identification result.

For example, the target processing strategy corresponding to the recognition task of abnormal behavior Db ₁ may include an image feature extraction path and a feature processing path. The image features of the time-series image sequence to be recognized may be firstly extracted, and then the image features may be processed to obtain the recognition result.

For example, the target processing strategy corresponding to the recognition task of the target object's behavior trajectory may include an object feature extraction path, an object feature recognition path, a target object position feature extraction path, and a target object position feature recognition path. The object feature may be extracted first to recognize the target object; then the target object's position feature may be recognized to obtain the recognition result of the target object's action trajectory.

According to the embodiments of the present disclosure, since the target processing strategy is obtained in response to the received selection operation for the target recognition task, the automatic matching of the target recognition task and the target processing strategy is realized based on the data interaction between the systems to adapt to the requirements of complex application scenarios. This solves the problem in the related art that the model obtained by training the network structure algorithm based on a specific task needs to be redeveloped before it can be applied to the actual application scenario.

According to an embodiment of the present disclosure, operations S210 to S230 may be performed by an electronic device. The electronic device may be a server or a terminal device. The server may be the server 105 in FIG. 1 . The terminal device may be the first terminal device 101, the second terminal device 102, or the third terminal device 103 in FIG. 1 .

According to an embodiment of the present disclosure, operation S220 may include the following operations:

According to the selection operation, identification information of the target recognition task is determined. According to the identification information of the target recognition task, target processing strategy information is obtained from the mapping relationship between the recognition task and the processing strategy.

The method shown in FIG. 2 will be further described below with reference to FIG. 3 to FIG. 10 in combination with specific embodiments.

FIG3 schematically shows a schematic diagram of obtaining a target processing strategy in response to receiving a selection operation for a target recognition task according to an embodiment of the present disclosure.

As shown in FIG3 , in 300, the recognition task 321 may include a recognition task _T1 (321_1), a recognition task _T2 (321_2), ..., and a recognition task _Tm (321_m). The mapping relationship 323 between the recognition task and the processing strategy may include a mapping relationship 323_1 between the recognition task _T1 and the processing strategy _TA1 , a mapping relationship 323_2 between the recognition task _T2 and the processing strategy _TA2 , ..., and a mapping relationship 323_n between the recognition task _Tn and the processing strategy _TAn .

According to an embodiment of the present disclosure, according to the selection operation, the identifier of the target recognition task can be determined to be the identifier of the recognition task T ₂ , indicating that the target recognition task 322 is the recognition task T _2. According to the identifier of the recognition task T ₂ , the target processing strategy 324 can be determined from the mapping relationship 323 between the recognition task and the processing strategy to be the processing strategy TA ₂ corresponding to the recognition task T ₂ .

According to an embodiment of the present disclosure, obtaining target processing strategy information from a mapping relationship between the recognition task and the processing strategy according to the identification information of the target recognition task may include the following operations:

According to the identification information of the target recognition task, the processing strategy information to be selected is obtained from the mapping relationship between the recognition task and the processing strategy. In response to the received selection operation for the processing strategy information to be selected, the target processing strategy is obtained.

According to an embodiment of the present disclosure, the processing strategies corresponding to the identification task _T2 may include multiple ones, for example: processing strategy _TA2-1 , processing strategy _TA2-2 , ..., processing strategy _TA2-I , and the processing strategy information to be selected may include: processing strategy _TA2-1 , processing strategy _TA2-2 , ..., processing strategy TA2 _-I .

According to an embodiment of the present disclosure, the policy information to be selected can be displayed on the client through a visual interface to facilitate the user to perform a selection operation. Based on the user's selection of the policy information to be selected, the target processing policy can be obtained. For example, if the user selects the processing policy TA _2-2 , then the target processing policy corresponding to the identification task T ₂ can be the processing policy TA _2-2 .

According to an embodiment of the present disclosure, the user's selection of the policy information to be selected may also be a selection of two or more processing strategies, and the obtained target processing strategy may be a combination processing strategy of the two or more processing strategies.

For example, the user may select processing strategy TA _2-2 and processing strategy TA _2-3 , and the obtained target processing strategy may be a combined processing strategy of processing strategy TA _2-2 and processing strategy TA _2-3 .

According to an embodiment of the present disclosure, for a combination of processing strategies, the order of the processing strategies in the combination can be determined based on the selection order, or based on the association relationship of the processing logic between the processing strategies in the combination. For example: the input data in the processing strategy TA _2-2 is obtained based on the processing result of the processing strategy TA _2-3 , and the order in the combined processing strategy can be determined as processing strategy TA _2-3 →processing strategy TA _2-2 .

According to an embodiment of the present disclosure, the target processing strategy may include a target feature extraction strategy and a target feature processing strategy. Operation S230 may include the following operations:

Based on the target feature extraction strategy, the target features of the time-series image sequence to be identified are extracted. Based on the target feature processing strategy, the target features are processed to obtain the recognition result.

FIG. 4 schematically shows a schematic diagram of processing spatial relationship features to obtain recognition results according to an embodiment of the present disclosure.

As shown in FIG. 4 , in 400 , a time-series image sequence 432 to be identified is obtained according to video data 431 , and the time-series image sequence 432 to be identified may include image P ₁ ( 432_1 ) corresponding to time t ₁ , image P ₂ ( 432_2 ) corresponding to time t ₂ , ... , image P _i ( 432_i ) corresponding to time _ti .

According to an embodiment of the present disclosure, based on the target feature extraction strategy, the spatial relationship features of each image in the time-series image sequence 432 to be identified are extracted in sequence to obtain the spatial relationship features 433. The spatial relationship features 433 may include the spatial relationship features 433_1 corresponding to the time _t1 , the spatial relationship features 433_2 corresponding to the time _t2 , ..., and the spatial relationship features 433_i corresponding to the time _t1 . The spatial relationship features characterize the relative motion features between the foreground features and the background features in the image. For example, the image may include streets, houses, and multiple active objects. Multiple active objects may be the foreground features of the image, and the streets and houses may be the background features of the image. The spatial relative relationship features between the foreground features and the background features may characterize the relative motion features between the multiple active objects and the streets and houses.

According to the embodiment of the present disclosure, based on the target feature processing strategy, the feature 434 of the spatial relationship feature changing over time is obtained. According to the feature of the spatial relationship feature changing over time, the recognition result 435 is obtained.

For example, during the time interval from t ₁ to t _i , as time changes, the relative motion characteristics between multiple active objects and streets and houses may be gradually gathered in the same direction, and the recognition result may be determined as abnormal behavior of multiple active objects, which may be gang fighting.

FIG. 5 schematically shows a schematic diagram of processing the action features of the first target object to obtain a recognition result according to an embodiment of the present disclosure.

As shown in FIG. 5 , in 500 , a time-series image sequence 532 to be identified is obtained according to video data 531 , and the time-series image sequence 532 to be identified may include image P ₁ ( 532_1 ) corresponding to time t ₁ , image P ₂ ( 532_2 ) corresponding to time t ₂ , ... , image P _i ( 532_i ) corresponding to time _ti .

According to an embodiment of the present disclosure, based on the target feature extraction strategy, the action features of the first target object of each image in the time-series image sequence 532 to be identified are extracted in sequence to obtain the action features 533 of the first target object. The action features 533 of the first target object may include an action feature 533_1 of the first target object corresponding to time _t1 , an action feature 533_2 of the first target object corresponding to time _t2 , ..., an action feature 533_i of the first target object corresponding to time _t1 .

According to an embodiment of the present disclosure, the motion feature may include a motion direction feature and a motion change feature. The motion direction feature may characterize the moving direction feature of the motion of the first target object relative to the ground. The motion change feature may characterize the motion change feature of the first target object.

According to an embodiment of the present disclosure, based on the target feature processing strategy, a feature 534 of the action feature of the first target object changing over time is obtained. According to the feature of the action feature of the first target object changing over time, a recognition result 535 is obtained.

For example, at time _t1 - _t1 , as time changes, the motion direction feature of the first target object may be the same moving direction relative to the ground, and gradually shortening the distance from the ground. The motion change feature of the first target object may be that the motion of the first target object changes from a knee bending motion to a body leaning forward motion, and finally changes to a crawling motion with both hands supporting the ground. It can be determined that the first target object has currently performed an abnormal behavior, and the abnormal behavior may be a fall.

According to an embodiment of the present disclosure, the action feature of the first target object can characterize the target skeleton point feature of the first target object. Based on the target feature processing strategy, the action feature of the first target object is processed to obtain the feature of the action feature of the first target object changing over time, which may include the following operations:

Based on the target feature processing strategy, the target skeleton point features are processed to obtain the change trend features of the target skeleton points over time; based on the change trend features of the target skeleton points over time, the change features of the action features of the first target object over time are obtained.

For example, the target skeleton points may include skeleton points of the human body trunk, limbs, head, etc. The change of the skeleton point features can more accurately characterize the change of the action features of the first target object. Based on the target feature processing strategy, the target skeleton point features are processed to improve the accuracy of image recognition.

FIG. 6 schematically shows a schematic diagram of obtaining a recognition result by processing the position features of the target area and the position features of the first target object according to an embodiment of the present disclosure.

As shown in FIG6 , in 600 , a time-series image sequence 632 to be identified is obtained according to video data 631 , and the time-series image sequence 632 to be identified may include image P ₁ ( 632_1 ) corresponding to time t ₁ , image P ₂ ( 632_2 ) corresponding to time t ₂ , ... , image P _i ( 632_i ) corresponding to time _ti .

According to an embodiment of the present disclosure, based on the target feature extraction strategy, the position features of the first target object and the position features of the target area of each image in the time-series image sequence 632 to be identified are extracted in sequence to obtain the position features 633 of the first target object and the position features 634 of the target area. The position features 633 of the first target object may include the position features 633_1 of the first target object corresponding to the time _t1 , the position features 633_2 of the first target object corresponding to the time _t2 , ..., and the position features 633_i of the first target object corresponding to the time _t1 . The position features 634 of the target area may include the position features 633_1 of the second target object corresponding to the time _t1 , the position features 634_2 of the target area corresponding to the time _t2 , ..., and the position features 634_i of the target area corresponding to the time _t1 .

According to an embodiment of the present disclosure, based on the target feature processing strategy, the position feature of the target area and the position feature of the first target object are processed to obtain a first change feature 635, wherein the first change feature characterizes the feature of the first relative position relationship between the first target object and the target area changing over time. According to the first change feature 635, a recognition result 636 is obtained.

For example, the position feature of the first target object may be a pixel coordinate feature, and the feature of the first relative position relationship between the first target object and the target area changing over time may include that the distance between the pixel coordinates of the first target object and the edge coordinates of the target area gradually decreases until the pixel coordinates of the first target object are within the target area. This indicates that the first target object has entered the target area, and the recognition result obtained may be that the first object has entered the target area.

FIG7 schematically shows a schematic diagram of obtaining a recognition result by processing the position feature of the first target object and the position feature of the second target object according to an embodiment of the present disclosure;

As shown in FIG. 7 , in 700 , a time-series image sequence 732 to be identified is obtained according to video data 731 , and the time-series image sequence 732 to be identified may include image P ₁ ( 732_1 ) corresponding to time t ₁ , image P ₂ ( 732_2 ) corresponding to time t ₂ , ... , image P _i ( 732_i ) corresponding to time _ti .

According to an embodiment of the present disclosure, based on the target feature extraction strategy, the position features of the first target object and the second target object of each image in the time-series image sequence 732 to be identified are sequentially extracted to obtain the position features 733 of the first target object and the position features 734 of the second target object. The position features 733 of the first target object may include the position features 733_1 of the first target object corresponding to the time _t1 , the position features 733_2 of the first target object corresponding to the time _t2 , ..., and the position features 733_i of the first target object corresponding to the time _t1 . The position features 734 of the second target object may include the position features 733_1 of the second target object corresponding to the time _t1 , the position features 734_2 of the second target object corresponding to the time _t2 , ..., and the position features 734_i of the second target object corresponding to the time _t1 .

According to an embodiment of the present disclosure, based on the target feature processing strategy, the position feature of the first target object and the position feature of the second target object are processed to obtain a second change feature 735. According to the second change feature 735, a recognition result 736 is obtained.

According to an embodiment of the present disclosure, the second change feature characterizes the feature that the second relative position relationship between the first target object and the second target object changes over time. The first target object may represent a person and parts such as the ear, head or mouth of the person, and the second target object may represent any object, such as a mobile phone, a cigarette, etc. The feature that the relative position relationship between the first target object and the second target object changes over time may be the feature that the relative position between the person and the mobile phone changes over time.

For example, the relative positions of the first target object and the second target object are gradually getting closer, indicating that the distance between the mobile phone and the ear or head of a person is gradually shortening. It can be determined that the behavior of the first target object is that the person is using the mobile phone to make a call, and the recognition result is that the first target object has made a call.

FIG8 schematically shows a schematic diagram of processing features of a first target object and features of a third target object to obtain a recognition result according to an embodiment of the present disclosure.

As shown in FIG8 , in 800 , a time-series image sequence 832 to be identified is obtained according to video data 831 , and the time-series image sequence 832 to be identified may include image P ₁ ( 832_1 ) corresponding to time t ₁ , image P ₂ ( 832_2 ) corresponding to time t ₂ , ... , image P _i ( 832_i ) corresponding to time _ti .

According to an embodiment of the present disclosure, based on the target feature extraction strategy, the position features of the first target object and the third target object of each image in the time-series image sequence 832 to be identified are sequentially extracted to obtain the position features 833 of the first target object and the position features 834 of the second target object. The position features 833 of the first target object may include the position features 833_1 of the first target object corresponding to the time _t1 , the position features 833_2 of the first target object corresponding to the time _t2 , ..., and the position features 833_i of the first target object corresponding to the time _t1 . The position features 834 of the second target object may include the position features 833_1 of the second target object corresponding to the time _t1 , the position features 834_2 of the second target object corresponding to the time _t2 , ..., and the position features 834_i of the second target object corresponding to the time _t1 .

According to an embodiment of the present disclosure, the video image 831 may be acquired by acquisition devices in different acquisition directions, or may be acquired by acquisition devices in the same acquisition direction.

According to an embodiment of the present disclosure, based on the target processing strategy, the following operations may be performed:

The features of the first target object and the features of the third target object are matched to obtain a feature matching result 835. When the feature matching result meets a predetermined threshold, a recognition result 836 is obtained based on the features of the first target object and the features of the third target object.

According to an embodiment of the present disclosure, when the feature matching result meets a predetermined threshold, obtaining a recognition result according to the feature of the first target object and the feature of the third target object may include the following operations:

When the feature matching result meets a predetermined threshold, the first target object and the third target object are determined to be the same target object. According to the features of the first target object and the features of the third target object, the target object attribute features are obtained. According to the target object attribute features, the recognition result is obtained.

According to an embodiment of the present disclosure, the features of the first target object and the features of the third target object may include features of different parts of the human body and action features, such as facial features, clothing color features, walking posture features, and the like.

According to an embodiment of the present disclosure, a feature matching result is obtained by using the similarity between the feature of the first target object and the feature of the third target object. When the similarity meets a predetermined threshold, it means that the feature of the first target object matches the feature of the third target object, and it can be determined that the first target object and the third target object are the same object.

According to an embodiment of the present disclosure, when it is determined that the first target object and the third target object are the same object, it can be used to count the crowd flow in a certain time period.

According to an embodiment of the present disclosure, the feature of the first target object may further include a feature of a relative motion direction between the first target object and a background object in the image.

For example, if the relative motion direction of the first target object in the image relative to the background object (e.g., a house) is away from the house, the flow of people leaving the house in a certain time period can be counted based on the relative motion direction characteristics of the first target object and the background object in the image. Similarly, if the relative motion direction of the first target object relative to the background object (e.g., a house) is entering the house, the flow of people entering the house in a certain time period can be counted based on the relative motion direction characteristics of the first target object and the background object in the image.

FIG. 9 schematically shows an overall exemplary system architecture diagram corresponding to different processing strategies according to an embodiment of the present disclosure.

As shown in FIG9 , in 900 , processing modules corresponding to different processing strategies are configured. When the target recognition task is the first abnormal behavior recognition, the processing modules corresponding to the target processing strategy may include: a detection and tracking module 903, a key point detection module 904, and a first abnormal behavior recognition module 905. Based on the target processing strategy corresponding to the first abnormal behavior recognition task, the video image 901 of the first acquisition device is processed to obtain a time-series image sequence 902 to be recognized. The detection and tracking module 903 processes the time-series image sequence 902 to be recognized to obtain the skeleton point features of the first target object. The key point monitoring module 904 processes the skeleton point features of the first target object to obtain the target skeleton point features. The target skeleton point features are processed using the first abnormal behavior recognition module 905 to obtain the recognition result of the first abnormal behavior.

According to the embodiment of the present disclosure, when the target recognition task is changed, the target processing strategy is changed accordingly. Then, other behavior recognition modules corresponding to the target recognition task are used to process the time-series image sequence 902 to be recognized, which will not be described in detail here.

According to an embodiment of the present disclosure, when the target recognition task is an input/output calculation and statistics task or an attribute recognition task, the video image 910 of the second acquisition device can be processed together with the video image 901 of the second acquisition device. For example: the detection and tracking module 912 is used to process the time-series image sequence 911 to be recognized to obtain the characteristics of the third target object. For example, the detection and tracking module 903 processes the time-series image sequence 902 to be recognized to obtain the characteristics of the first target object. The feature matching module 913 is used to process the characteristics of the first target object and the characteristics of the third target object to obtain a feature matching result. Then, the attribute recognition module 914 is used to process the characteristics of the first target object and the characteristics of the third target object to obtain an attribute recognition result. Or the input/output counting module 915 is used to perform calculation and statistics on the same target object to obtain an input/output counting and statistics result.

FIG. 10 schematically shows an exemplary system architecture deployment diagram according to an embodiment of the present disclosure.

As shown in FIG10 , in the exemplary system architecture deployment diagram 1000, an algorithm architecture layer 1001, an application layer 1002, and a deployment layer 1003 are included. In the algorithm architecture layer 1001, input data 10011 may include image files, single-lens video images, and multi-lens video images. The algorithm principle 10012 is to obtain target features by performing target detection on the input data. The target features are processed using multi-target tracking technology, feature association technology, facial recognition technology, trajectory fusion technology, and other technologies to obtain output data 10013. The specific processing strategy is determined according to the target recognition task, and the output data 10013 may include attribute recognition results, abnormal behavior recognition results, trajectory/traffic counts, and the like.

The application layer 1002 deploys various application function modules, including an abnormal behavior warning function module 10021, a traffic density monitoring function module 10022, an inbound and outbound traffic control function module 10023, a video structuring function module 10024, an attribute analysis function module 10025, etc.

The deployment layer 1003 includes the native inference library (Paddle Inference) 10031, the service-oriented deployment framework (PaddleServing) 10032, and the deep learning optimized inference optimizer (TensorRT) 10033.

Figure 11 schematically shows a block diagram of an image recognition device according to an embodiment of the present disclosure.

As shown in FIG. 11 , the image recognition device 1100 includes a first obtaining module 1101 , a second obtaining module 1102 and a third obtaining module 1103 .

The first acquisition module is used to obtain a time-series image sequence to be identified according to the video image.

The second obtaining module is used to obtain a target processing strategy in response to receiving a selection operation for a target recognition task.

The third acquisition module is used to process the time-series image sequence to be recognized based on the target processing strategy to obtain the recognition result.

According to an embodiment of the present disclosure, the second obtaining module 1102 may include a determining unit and a first obtaining unit.

The determination unit is used to determine the identification information of the target recognition task according to the selection operation.

The first obtaining unit is used to obtain target processing strategy information from the mapping relationship between the recognition task and the processing strategy according to the identification information of the target recognition task.

According to an embodiment of the present disclosure, the first obtaining unit may include a first obtaining sub-unit and a second obtaining sub-unit.

The first obtaining subunit is used to obtain the processing strategy information to be selected from the mapping relationship between the recognition task and the processing strategy according to the identification information of the target recognition task.

The second obtaining subunit is used to obtain a target processing strategy in response to a received selection operation on the processing strategy information to be selected.

According to an embodiment of the present disclosure, the target processing strategy includes a target feature extraction strategy and a target feature processing strategy, and the third obtaining module 1103 may include: a second obtaining unit and a third obtaining unit.

The second obtaining unit is used to extract the target features of the time-series image sequence to be identified based on the target feature extraction strategy.

The third obtaining unit is used to process the target feature based on the target feature processing strategy to obtain the recognition result.

According to an embodiment of the present disclosure, the landmark feature includes a spatial relationship feature, and the third obtaining unit includes: a third obtaining sub-unit and a fourth obtaining sub-unit.

The third obtaining subunit is used to process the spatial relationship features based on the target feature processing strategy to obtain the characteristics of the spatial relationship features that change over time.

The fourth obtaining subunit is used to obtain the recognition result according to the characteristics of the spatial relationship characteristics that change over time.

According to an embodiment of the present disclosure, the target feature includes an action feature of the first target object, and the third obtaining unit includes: a fifth obtaining sub-unit and a sixth obtaining sub-unit.

The fifth obtaining subunit is used to process the action feature of the first target object based on the target feature processing strategy to obtain the feature of the action feature of the first target object that changes with time.

The sixth obtaining subunit is used to obtain a recognition result according to the characteristics of the action characteristics of the first target object that change over time.

According to an embodiment of the present disclosure, the action feature of the first target object represents the target skeleton point feature of the first target object, and the sixth acquisition subunit is used to process the target skeleton point feature based on the target feature processing strategy to obtain the change trend feature of the target skeleton point over time. According to the change trend feature of the target skeleton point over time, the feature of the action feature of the first target object changing over time is obtained.

According to an embodiment of the present disclosure, the target feature includes a position feature of the target area and a position feature of the first target object, and the third obtaining unit includes: a seventh obtaining subunit and an eighth obtaining subunit.

The seventh obtaining subunit is used to process the position feature of the target area and the position feature of the first target object based on the target feature processing strategy to obtain a first change feature, wherein the first change feature represents a feature of a first relative position relationship between the first target object and the target area that changes over time. The eighth obtaining subunit is used to obtain a recognition result based on the first change feature.

According to an embodiment of the present disclosure, the target feature includes a position feature of a first target object and a position feature of a second target object, and the third obtaining unit includes: a ninth obtaining subunit and a tenth obtaining subunit.

The ninth obtaining subunit is used to process the position feature of the first target object and the position feature of the second target object based on the target feature processing strategy to obtain a second change feature, wherein the second change feature represents a feature of a second relative position relationship between the first target object and the second target object that changes over time. The tenth obtaining subunit is used to obtain a recognition result based on the second change feature.

According to an embodiment of the present disclosure, the target features include: features of a first target object and features of a third target object, and the features of the first target object and the features of the third target object are extracted from at least two images with different acquisition directions; the third acquisition unit includes: an eleventh acquisition subunit and a twelfth acquisition subunit.

The eleventh obtaining subunit is used to match the features of the first target object and the features of the third target object based on the target feature processing strategy to obtain a feature matching result.

The twelfth obtaining subunit is used to obtain a recognition result according to the features of the first target object and the features of the third target object when the feature matching result meets a predetermined threshold.

According to an embodiment of the present disclosure, the twelfth obtaining subunit is used to: determine that the first target object and the third target object are the same target object when the feature matching result meets a predetermined threshold value. Obtain target object attribute features based on the features of the first target object and the features of the third target object. Obtain recognition results based on the target object attribute features.

According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

According to an embodiment of the present disclosure, an electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method as described above.

According to an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause a computer to execute the method described above.

According to an embodiment of the present disclosure, a computer program product includes a computer program, and when the computer program is executed by a processor, the computer program implements the method as described above.

FIG. 12 shows a schematic block diagram of an example electronic device 1200 that can be used to implement an embodiment of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and/or required herein.

As shown in FIG. 12 , the device 1200 includes a computing unit 1201, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1202 or a computer program loaded from a storage unit 1208 into a random access memory (RAM) 1203. In the RAM 1203, various programs and data required for the operation of the device 800 can also be stored. The computing unit 1201, the ROM 1202, and the RAM 1203 are connected to each other via a bus 804. An input/output (I/O) interface 1205 is also connected to the bus 1204.

A number of components in the device 1200 are connected to the I/O interface 1205, including: an input unit 1206, such as a keyboard, a mouse, etc.; an output unit 1207, such as various types of displays, speakers, etc.; a storage unit 1208, such as a disk, an optical disk, etc.; and a communication unit 1209, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1209 allows the device 1200 to exchange information/data with other devices through a computer network such as the Internet and/or various telecommunication networks.

The computing unit 1201 may be a variety of general and/or special processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 1201 performs the various methods and processes described above, such as image recognition methods. For example, in some embodiments, the image recognition method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 1208. In some embodiments, part or all of the computer program may be loaded and/or installed on the device 1200 via ROM 1202 and/or communication unit 1209. When the computer program is loaded into RAM 1203 and executed by the computing unit 1201, one or more steps of the image recognition method described above may be performed. Alternatively, in other embodiments, the computing unit 1201 may be configured to perform the image recognition method in any other appropriate manner (e.g., by means of firmware).

Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and/or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, implements the functions/operations specified in the flow chart and/or block diagram. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communications network). Examples of communications networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server arises through computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims (25)

1. An image recognition method, comprising: According to the video image, a time-series image sequence to be identified is obtained; In response to receiving a selection operation for a target recognition task, a target processing strategy corresponding to the target recognition task is obtained from a mapping relationship between recognition tasks and processing strategies, wherein the target processing strategy includes a combination processing strategy corresponding to the target recognition task, and the execution order of each processing strategy in the combination processing strategy is determined based on the association relationship of the processing logic of each processing strategy; the target recognition task includes any one of the following: an abnormal behavior recognition task, a target attribute feature recognition task, and a target object behavior trajectory recognition task; and Based on the target processing path corresponding to the target processing strategy, the time-series image sequence to be identified is processed to obtain an identification result; the target processing strategy includes a target feature extraction strategy and a target feature processing strategy. 2. The method according to claim 1, wherein, in response to receiving a selection operation for a target recognition task, obtaining a target processing strategy corresponding to the target recognition task from a mapping relationship between recognition tasks and processing strategies comprises: According to the selection operation, determining identification information of the target recognition task; and According to the identification information of the target recognition task, the target processing strategy information is obtained from the mapping relationship between the recognition task and the processing strategy. 3. The method according to claim 2, wherein the step of obtaining the target processing strategy information from a mapping relationship between the recognition task and the processing strategy according to the identification information of the target recognition task comprises: According to the identification information of the target recognition task, obtaining the processing strategy information to be selected from the mapping relationship between the recognition task and the processing strategy; and In response to the received selection operation on the to-be-selected processing strategy information, the target processing strategy is obtained. 4. The method according to claim 1, wherein the step of processing the time-series image sequence to be identified based on the target processing path corresponding to the target processing strategy to obtain the identification result comprises: Extracting target features of the time-series image sequence to be identified based on the target feature extraction strategy; and Based on the target feature processing strategy, the target feature is processed to obtain the recognition result. 5. The method according to claim 4, wherein the target feature includes a spatial relationship feature, and the processing of the target feature based on the target feature processing strategy to obtain the recognition result comprises: Based on the target feature processing strategy, the spatial relationship feature is processed to obtain a feature of the spatial relationship feature changing over time; and The recognition result is obtained according to the characteristics of the spatial relationship characteristics changing with time. 6. The method according to claim 4, wherein the target feature comprises an action feature of the first target object, and the step of processing the target feature based on the target feature processing strategy to obtain the recognition result comprises: Based on the target feature processing strategy, the action feature of the first target object is processed to obtain a feature of the action feature of the first target object changing over time; and The recognition result is obtained according to the characteristics of the action characteristics of the first target object that change over time. 7. The method according to claim 6, wherein the motion feature of the first target object represents the target bone point feature of the first target object, and the processing of the motion feature of the first target object based on the target feature processing strategy to obtain the feature of the motion feature of the first target object changing over time comprises: Based on the target feature processing strategy, the target skeleton point features are processed to obtain the change trend features of the target skeleton point over time; and According to the change trend characteristics of the target skeleton points over time, the characteristics of the action characteristics of the first target object changing over time are obtained. 8. The method according to claim 4, wherein the target feature comprises a position feature of a target area and a position feature of a first target object, and the step of processing the target feature based on the target feature processing strategy to obtain the recognition result comprises: Based on the target feature processing strategy, the position feature of the target area and the position feature of the first target object are processed to obtain a first change feature, wherein the first change feature represents a feature of a first relative position relationship between the first target object and the target area that changes over time; and The recognition result is obtained according to the first change feature. 9. The method according to claim 4, wherein the target feature comprises a position feature of a first target object and a position feature of a second target object, and the step of processing the target feature based on the target feature processing strategy to obtain the recognition result comprises: Based on the target feature processing strategy, the position feature of the first target object and the position feature of the second target object are processed to obtain a second change feature, wherein the second change feature represents a feature of a second relative position relationship between the first target object and the second target object that changes over time; and The recognition result is obtained according to the second change feature. 10. The method according to claim 4, wherein the target features include: features of a first target object and features of a third target object, and the features of the first target object and the features of the third target object are extracted from at least two images with different acquisition directions; and the processing of the target features based on the target feature processing strategy to obtain the recognition result comprises: Based on the target feature processing strategy, matching the features of the first target object and the features of the third target object to obtain a feature matching result; and When the feature matching result meets a predetermined threshold, the recognition result is obtained according to the feature of the first target object and the feature of the third target object. 11. The method according to claim 10, wherein, when the feature matching result satisfies a predetermined threshold, obtaining the recognition result according to the feature of the first target object and the feature of the third target object comprises: When the feature matching result satisfies a predetermined threshold, determining that the first target object and the third target object are the same target object; Obtaining target object attribute features according to the features of the first target object and the features of the third target object; and The recognition result is obtained according to the attribute characteristics of the target object. 12. An image recognition device, comprising: A first acquisition module is used to obtain a time-series image sequence to be identified based on the video image; A second obtaining module is used for, in response to receiving a selection operation for a target recognition task, obtaining a target processing strategy corresponding to the target recognition task from a mapping relationship between recognition tasks and processing strategies, wherein the target processing strategy includes a combination processing strategy corresponding to the target recognition task, and the execution order of each processing strategy in the combination processing strategy is determined based on the association relationship of the processing logic of each processing strategy; the target recognition task includes any one of the following: an abnormal behavior recognition task, a target attribute feature recognition task, and a target object behavior trajectory recognition task; and The third acquisition module is used to process the time-series image sequence to be identified based on the target processing path corresponding to the target processing strategy to obtain the identification result; the target processing strategy includes a target feature extraction strategy and a target feature processing strategy. 13. The apparatus according to claim 12, wherein the second obtaining module comprises: a determining unit, configured to determine identification information of a target recognition task according to the selection operation; and The first obtaining unit is used to obtain the target processing strategy information from the mapping relationship between the recognition task and the processing strategy according to the identification information of the target recognition task. 14. The apparatus according to claim 13, wherein the first obtaining unit comprises: A first obtaining subunit is used to obtain the processing strategy information to be selected from the mapping relationship between the recognition task and the processing strategy according to the identification information of the target recognition task; and The second obtaining subunit is used to obtain the target processing strategy in response to the received selection operation for the to-be-selected processing strategy information. 15. The device according to claim 12, wherein the third obtaining module comprises: a second obtaining unit, configured to extract target features of the time-series image sequence to be identified based on the target feature extraction strategy; and The third obtaining unit is used to process the target feature based on the target feature processing strategy to obtain the recognition result. 16. The apparatus according to claim 15, wherein the target feature comprises a spatial relationship feature, and the third obtaining unit comprises: A third obtaining subunit is used to process the spatial relationship feature based on the target feature processing strategy to obtain a feature of the spatial relationship feature changing over time; and The fourth obtaining subunit is used to obtain the recognition result according to the characteristics of the spatial relationship characteristics that change over time. 17. The device according to claim 15, wherein the target feature comprises an action feature of the first target object, and the third obtaining unit comprises: a fifth obtaining subunit, configured to process the motion feature of the first target object based on the target feature processing strategy to obtain a feature of the motion feature of the first target object that changes over time; and The sixth obtaining subunit is used to obtain the recognition result according to the characteristics of the action characteristics of the first target object that change over time. 18. The device according to claim 17, wherein the action feature of the first target object represents a target skeleton point feature of the first target object, and the sixth obtaining subunit is used for: Based on the target feature processing strategy, the target skeleton point feature is processed to obtain a change trend feature of the target skeleton point over time; and According to the change trend characteristics of the target skeleton points over time, the characteristics of the action characteristics of the first target object changing over time are obtained. 19. The device according to claim 15, wherein the target feature comprises a position feature of the target area and a position feature of the first target object, and the third obtaining unit comprises: a seventh obtaining subunit, configured to process the position feature of the target area and the position feature of the first target object based on the target feature processing strategy to obtain a first change feature, wherein the first change feature represents a feature of a first relative position relationship between the first target object and the target area that changes over time; and An eighth obtaining subunit is used to obtain the recognition result according to the first change feature. 20. The apparatus according to claim 15, wherein the target feature comprises a position feature of a first target object and a position feature of a second target object, and the third obtaining unit comprises: a ninth obtaining subunit, configured to process the position feature of the first target object and the position feature of the second target object based on the target feature processing strategy to obtain a second change feature, wherein the second change feature represents a feature of a second relative position relationship between the first target object and the second target object changing over time; and The tenth obtaining subunit is used to obtain the recognition result according to the second change feature. 21. The device according to claim 15, wherein the target features include: features of a first target object and features of a third target object, the features of the first target object and the features of the third target object are extracted from at least two images with different acquisition directions; and the third obtaining unit includes: an eleventh obtaining subunit, configured to match the features of the first target object and the features of the third target object based on the target feature processing strategy to obtain a feature matching result; and The twelfth obtaining subunit is used to obtain the recognition result according to the feature of the first target object and the feature of the third target object when the feature matching result meets a predetermined threshold. 22. The apparatus according to claim 21, wherein the twelfth obtaining subunit is used for: When the feature matching result satisfies a predetermined threshold, determining that the first target object and the third target object are the same target object; Obtaining target object attribute features according to the features of the first target object and the features of the third target object; and The recognition result is obtained according to the attribute characteristics of the target object. 23. An electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 11. 24. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the method according to any one of claims 1-11. 25. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 11.