Human body recognition system management methods, devices, storage media and electronic equipment

Abstract

This application proposes a human body recognition system management method, apparatus, storage medium, and electronic device. First, the human body registration identity of the corresponding user is determined based on the target facial image. If no corresponding human body registration image exists for the registered identity, any target human body image from the target human body image tracking set is selected as the corresponding human body registration image. By automatically using any target human body image from the human body image tracking set as the human body registration image when no corresponding human body registration image exists, this overcomes the technical problem of requiring the user to actively provide a human body registration image, reduces user intervention, saves manpower and time, and is more convenient.

Description

Technical Field

[0001] This application relates to the field of image recognition, and more specifically, to a human body recognition system management method, apparatus, storage medium, and electronic device. Background Technology

[0002] With the widespread adoption of smart devices and the rapid development of network technology, smart cameras are playing an increasingly important role in people's daily lives. The video captured by smart cameras forms a crucial foundation for computer vision. Computer vision includes facial recognition, human body recognition, and vehicle recognition, among others.

[0003] Human body recognition technology, a popular research area in computer vision, primarily addresses the technical challenges of recognizing and retrieving human bodies across different cameras and scenes. It can identify individuals based on their clothing, posture, hairstyle, and other information, serving as a crucial supplement to facial recognition technology. This allows for continuous tracking of individuals across different cameras when clear facial images are unavailable, enhancing the spatiotemporal continuity of data. The combination of human body recognition and facial recognition technologies opens up new application scenarios, elevating the cognitive level of artificial intelligence to a new stage.

[0004] Currently, in human body recognition technology, users usually need to actively provide one or more human body registration images. After detecting a human body in a real-world application scenario, the human body recognition algorithm matches and identifies it against the registered image. Its disadvantage is that users need to register manually. Summary of the Invention

[0005] The purpose of this application is to provide a human body recognition system management method, apparatus, storage medium, and electronic device to at least partially improve the above-mentioned problems.

[0006] To achieve the above objectives, the technical solutions adopted in the embodiments of this application are as follows:

[0007] In a first aspect, embodiments of this application provide a method for managing a human body recognition system, the method comprising:

[0008] Determine the registered identity of the corresponding user based on the target facial image;

[0009] If no corresponding human registration image exists for the registered human identity, any target human image in the target human image tracking set will be identified as the human registration image corresponding to the registered human identity.

[0010] The target human body image tracking set is a human body image tracking set associated with the target human face image, wherein the target human body image and the target human face image belong to the same target person.

[0011] Secondly, embodiments of this application provide a human body recognition system management device, the device comprising:

[0012] The identity determination unit is used to determine the human registration identity of the corresponding user based on the target facial image;

[0013] The registration unit is used to determine any target human image in the target human image tracking set as the human registration image corresponding to the human registration identity when there is no corresponding human registration image for the human registration identity.

[0014] The target human body image tracking set is a human body image tracking set associated with the target human face image, wherein the target human body image and the target human face image belong to the same target person.

[0015] Thirdly, embodiments of this application provide a storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described method.

[0016] Fourthly, embodiments of this application provide an electronic device, the electronic device comprising: a processor and a memory, the memory being used to store one or more programs; when the one or more programs are executed by the processor, the above-described method is implemented.

[0017] Compared to existing technologies, the human body recognition system management method, apparatus, storage medium, and electronic device provided in this application first determine the corresponding user's human body registration identity based on the target face image; if no corresponding human body registration image exists for the registered identity, any target human body image in the target human body image tracking set is determined as the human body registration image corresponding to the registered identity. By automatically using any target human body image in the target human body image tracking set as the human body registration image when no corresponding human body registration image exists, this overcomes the technical problem of requiring the user to actively provide a human body registration image, reduces user intervention, saves manpower and time, and is more convenient.

[0018] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart illustrating the human body recognition system management method provided in this application embodiment;

[0021] Figure 2 One of the flowcharts of the human body recognition system management method provided in the embodiments of this application;

[0022] Figure 3 A schematic diagram of the sub-steps of S1010 provided in the embodiments of this application;

[0023] Figure 4 One of the flowcharts of the human body recognition system management method provided in the embodiments of this application;

[0024] Figure 5 A schematic diagram of the sub-step S1014 provided in the embodiments of this application;

[0025] Figure 6 A schematic diagram of the coordinate system provided in the embodiments of this application;

[0026] Figure 7 One of the flowcharts of the human body recognition system management method provided in the embodiments of this application;

[0027] Figure 8 One of the flowcharts of the human body recognition system management method provided in the embodiments of this application;

[0028] Figure 9 A schematic diagram of a human body recognition system management device provided in an embodiment of this application;

[0029] Figure 10 This is a hardware structure block diagram of the server provided in an embodiment of this application.

[0030] In the diagram: 1100 - Server; 1110 - Processor; 1120 - Storage medium; 1121 - Operating system; 1122 - Data; 1123 - Application program; 1130 - Memory; 1140 - Input / output interface; 1150 - Wired or wireless network interface; 1160 - Power supply; 301 - Identity verification unit; 302 - Registration unit. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0032] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0033] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0034] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0035] In the description of this application, it should be noted that the terms "upper", "lower", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship that the product of this application is usually placed in. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0036] In the description of this application, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set" and "connection" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0037] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0038] To overcome the above technical problems, this application provides a human body recognition system management method, applicable to servers or other electronic device terminals. Please refer to... Figure 1 , Figure 1 This is a schematic flowchart of a human body recognition system management method provided in an embodiment of this application. Figure 1 As shown, the management methods for human body recognition systems include:

[0039] S1010: Determine the human registration identity of the corresponding user based on the target facial image.

[0040] Optionally, the registered identity can be an account, name, email address, or other symbol that can serve as an identifier. For example, if the target person is Zhang San and the registered identity is an account, Zhang San's facial image is the target face image. By recognizing Zhang San's facial image, the account corresponding to Zhang San in the human body recognition system can be determined.

[0041] S1012, determine if a corresponding human registration image exists for the registered human identity. If yes, proceed to S1014; otherwise, proceed to S1013.

[0042] Specifically, referring to the example above, after determining Zhang San's account in the human body recognition system, assuming his account is AAAA, determine whether AAAA has a corresponding registered human image. The registered human image is used as a template for human body recognition in the system. For example, when obtaining the human image to be recognized, compare it with the registered human image corresponding to AAAA (Zhang San). If the comparison result shows that the target person in the human image to be recognized matches the registered human image, then the human image to be recognized is Zhang San's human image.

[0043] If there is no corresponding human registration image for a registered human identity, it means that there is no template for human recognition in the human body recognition system. At this time, execute S1013 to add a corresponding human registration image for the registered human identity.

[0044] If a corresponding human registration image exists for the registered human identity, further judgment is needed on whether to update the human registration image, so S1014 is executed.

[0045] S1013, determine any target human image in the target human image tracking set as the human registration image corresponding to the human registration identity.

[0046] Among them, the target human body image tracking set is a human body image tracking set associated with the target human face image, and the target human body image and the target human face image belong to the same target person.

[0047] Please continue to refer to the example above. The target human body image tracking set is the human body image tracking set of Zhang San, which stores human body images of Zhang San.

[0048] The human body recognition system management method provided in this application automatically uses any target human body image from the target human body image tracking set as the human body registration image when no corresponding human body registration image exists for the registered human body identity. This overcomes the technical problem of requiring users to actively provide human body registration images, reduces user intervention, saves manpower and time, and is more convenient. Furthermore, it can update human body registration data in real time, improving the accuracy of subsequent image comparisons.

[0049] S1014, sequentially determine whether each target human image has a corresponding first human registration image. If yes, proceed to S1015; otherwise, proceed to S1016.

[0050] The first registered human image is a registered human image with the same viewing angle as the target human image in the human image tracking set. The viewing angle refers to the orientation of the human body relative to the image capturing lens, the direction of the target person (front, chest, or back) relative to the image acquisition device, such as front, side, and back, etc.

[0051] Continuing with the previous example, assume there are 30 registered human images corresponding to Zhang San's identity. From these 30 images, select the first registered human image whose viewpoint is the same as the target human image. It's possible that the viewpoints of the 30 registered human images are all different from the target human image. To enrich the diversity of recognition templates in the human body recognition system and improve its accuracy, update the target human image with the new registered human image, i.e., execute S1016.

[0052] If a corresponding first registered human image exists for the target human image, updating the target human image to a new registered human image may result in duplication with the first registered human image. This would fail to enrich the diversity of recognition templates in the human recognition system and instead make the system redundant. Therefore, S1015 needs to be executed.

[0053] S1015, determine whether the similarity between the target human image and the first registered human image is greater than a threshold. If yes, proceed to S1017; otherwise, proceed to S1016.

[0054] Optionally, continuing with Zhang San as the target person as an example, the differences in the human body images obtained when Zhang San wears different clothes or has different hairstyles under the same viewpoint can be significant. Using human body images that belong to Zhang San but are very different as recognition templates for the human body recognition system helps improve the recognition accuracy of the system. Therefore, if the similarity between the target human body image and the first registered human body image is less than or equal to the threshold, S1016 is executed. Conversely, if the similarity between the target human body image and the first registered human body image is greater than the threshold, to avoid redundancy in the human body recognition system, S1017 is executed.

[0055] S1016, Update the target human image to a new human registration image.

[0056] In the human body recognition system management method provided in this application embodiment, if the perspective of the human body registration image is different from that of the target human body image, in order to enrich the diversity of recognition templates in the human body recognition system and improve the recognition accuracy of the human body recognition system, the target human body image is updated to a new human body registration image; when the similarity between the target human body image and the first human body registration image is less than or equal to a threshold, human body images that belong to the same template person but have great differences are used as new recognition templates for the human body recognition system, which is beneficial to improving the recognition accuracy of the human body recognition system.

[0057] It should be noted that, in the embodiments of this application, updating the target human body image to a new human body registration image means adding the target human body image to the new human body registration image while keeping the original human body registration image unchanged. By increasing the number of human body registration images, when performing human body recognition, the image is compared and judged with a large number of human body registration images to reduce errors caused by large changes in perspective or clothing.

[0058] S1017, skip.

[0059] exist Figure 1 Based on this, regarding how to obtain the perspective of the target human body image and the content in S1013, this application embodiment also provides a possible implementation method, please refer to... Figure 2 Human body recognition system management methods also include:

[0060] S1011, obtain the perspective of each target human body image based on the human body attribute recognition algorithm.

[0061] Among them, the human attribute recognition algorithm can be either a deep single-attribute recognition algorithm (abbreviated as DeepSAR) or a deep multi-attribute recognition algorithm (abbreviated as DeepMAR).

[0062] S1013 includes:

[0063] S1013-1, determine any one of the target human images from the same viewpoint in the target human image tracking set as the human registration image corresponding to the human registration identity.

[0064] Specifically, in S1011, the viewpoint of each target human image has been acquired. Within the same target video, the target person's clothing, attire, hairstyle, and other physical characteristics remain largely consistent. Therefore, multiple human images of the target person from the same viewpoint exhibit high similarity and minimal differences. To avoid redundancy in the human recognition system, only any one image from each target human image at the same viewpoint can be identified as the human registration image corresponding to the registered identity. This reduces the storage size of the human registration images and improves the efficiency of image comparison.

[0065] Because only one image of a target human body from the same viewpoint can be used as the human body registration image, the human body registration image consists of target human body images from different viewpoints. This means registering target human body images from different viewpoints as human body registration images. Assuming that the target person's clothing, attire, hairstyle, and other physical characteristics remain largely consistent, reducing the storage size of the human body registration images avoids redundancy in the human body recognition system and further improves the efficiency of image comparison. Figure 1 Based on the above, regarding the content in S1010, this application embodiment also provides a possible implementation method, please refer to... Figure 3 S1010 includes:

[0066] S1010-1, Obtain the target face features of the target face image based on the face recognition algorithm.

[0067] Among them, the face recognition algorithm is a commonly used face recognition algorithm such as CosFace or ArcFace.

[0068] S1010-2, Determine the second face registration image from the face feature database.

[0069] The second face registration image is the pre-registered face image with the highest similarity to the target face features, and the face feature database includes at least one set of pre-registered face images.

[0070] Optionally, continuing with the previous example, the target person is Zhang San, and Zhang San's facial image is the target face image. Assume the facial feature database includes pre-registered face images of people such as Zhang San, Zhang Si, and Zhang Wu. Similarity is calculated between the target face image's facial features and the facial features of each pre-registered face image. The pre-registered face image of Zhang San with the highest similarity to the target face features is then designated as the second face registration image. The user corresponding to the second face registration image is Zhang San.

[0071] S1010-3, determine the registration identity corresponding to the second face registration image as the human body registration identity of the target user.

[0072] Because the user corresponding to the second face registration image is the same as the user corresponding to the target human body image, and the target human body image tracking set is the human body image tracking set associated with the target face image, and the target human body image and the target face image belong to the same target person, the registration identity corresponding to the second face registration image can be determined as the human body registration identity of the target user.

[0073] exist Figure 3 In order to further ensure the accuracy of the obtained human registration identity, this application embodiment also provides a possible implementation method, please refer to the following.

[0074] Before S1010-3, determine whether the similarity between the target face image and the second face registration image is higher than the face similarity threshold; if so, determine the identity corresponding to the pre-registered face feature with the highest similarity as the registered identity of the target user.

[0075] Continuing with the example above, let's assume the target person is Li Er, and the target face image is Li Er's facial image. After calculating the similarity between the target face features of the target face image and the face features of each pre-registered face image, we find that Zhang Si's pre-registered face image has the highest similarity to the target face features, but it is still below the face similarity threshold, indicating that Li Er and Zhang Si are not the same target person. Therefore, the registration identity corresponding to the second face registration image of Zhang Si cannot be determined as Li Er's registered identity.

[0076] exist Figure 1 Based on this, regarding how to obtain the target face image and target body image tracking set, embodiments of this application also provide a possible implementation method, please refer to... Figure 4 Human body recognition system management methods also include:

[0077] S102, perform face detection on the target video and store the face images of different target persons in the corresponding face image tracking set.

[0078] The face image carries attribute information and a target frame identifier. The target frame identifier carried by the face image is the identifier of the video frame in which the face image was detected. For example, if a face image is detected in the i-th frame of the video image, then the target frame identifier corresponding to that face image is i.

[0079] Optionally, the MTCNN or RetinaFace face detection algorithm is used to detect faces in the target video. The IOU tracking algorithm is used to store the face images of the same person in consecutive frames into the same face image tracking set, and to store the face images of different target persons into their respective face image tracking sets.

[0080] S103, perform human detection on the target video and store the human images of different target persons in the corresponding human image tracking set.

[0081] Among them, the human body image carries attribute information and target frame identifier. The target frame identifier carried by the human body image is the identifier of the video frame in which the human body image is detected.

[0082] Optionally, the YOLO or SSD human detection algorithm is used to detect human bodies in the target video. The IOU tracking algorithm is used to store human images of the same person in consecutive frames into the same human image tracking set, and human images of different target persons are stored in their respective human image tracking sets.

[0083] S104, based on the attribute information and target frame identifier carried by the face image and the body image respectively, determine the body image tracking set associated with each face image tracking set.

[0084] Optionally, the target video may contain multiple characters, such as four characters: A, B, C, and D. Four face image tracking sets and four body image tracking sets are obtained through steps S102 and S103. Then, based on the attribute information and target frame identifier carried by the face and body images respectively, the four face image tracking sets are associated with their corresponding body image tracking sets. For example, the face image tracking set corresponding to character A is associated with the body image tracking set corresponding to character A. This ensures that the characters corresponding to the associated face image tracking sets and body image tracking sets are the same.

[0085] S105, determine any face image in the face image tracking set as the target face image.

[0086] Optionally, the best-quality face image is selected from the face image tracking set and determined as the target face image, and the quality of the face image is obtained based on the image's sharpness, brightness, and the opening and closing state of the eyes.

[0087] S106, the human body image tracking set associated with the face image tracking set is determined as the target human body image tracking set.

[0088] Continuing with the previous example, assume the target person is person A, and person A corresponds to face image tracking set 2. Face image tracking set 2 is associated with human image tracking set 4. Any face image in face image tracking set 2 is identified as the target face image, and human image tracking set 4 is identified as the target human image tracking set.

[0089] exist Figure 4 Based on the above, regarding the content in S1014, this application embodiment also provides a possible implementation method, please refer to... Figure 5 S1014 includes:

[0090] S1014-1: Sequentially determine whether the attribute information of the face images in the face image tracking set matches the attribute information of the human images in the human image tracking set carrying the same target frame identifier. If yes, execute S1014-2; otherwise, repeat S1014-1.

[0091] Suppose that the face image tracking set includes 3 face images, and the body image tracking set includes 3 body images. The target frame identifiers carried by the 3 face images and 3 body images are 7, 8, and 9, respectively. First, it is determined whether the attribute information of the face image with target frame identifier 7 matches the attribute information of the body image with target frame identifier 7. If they match, S1014-2 is executed to determine that the face image tracking set and the body image tracking set are associated. If they do not match, it is determined whether the attribute information of the face image with target frame identifier 8 matches the attribute information of the body image with target frame identifier 8. If they match, S1014-2 is executed. If they do not match, it is determined whether the attribute information of the face image with target frame identifier 9 matches the attribute information of the body image with target frame identifier 9. If they match, S1014-2 is executed. If they do not match, it is determined that the face image tracking set and the body image tracking set are not associated. At this time, it is determined whether the face image tracking set is associated with the next body image tracking set.

[0092] In this embodiment, the face image and body image to be matched are determined by the target frame identifier to be from the same video frame, thereby ensuring the accuracy of the matching results.

[0093] S1014-2, Determine the association between the face image tracking set and the human body image tracking set.

[0094] exist Figure 5 Based on the above, regarding the content in S1014-1, this application embodiment also provides a possible implementation method, please refer to the following.

[0095] The attribute information includes the x and y coordinates of the top-left corner of the corresponding image, as well as the width and height of the corresponding image. For example... Figure 6As shown in the figure, the lower left corner of the video frame image is used as the origin O, and the two sides of the video frame image are used as the X-axis and Y-axis respectively. As described above, the face image and the human body image that have been matched through the target frame identification come from the same video frame image. The abscissa and ordinate of the upper left corner of the face image and the human body image in the coordinate system corresponding to the video frame image can be obtained respectively.

[0096] S1014-1. Determine whether the attribute information of the face image in the face image tracking set matches the attribute information of the human body image in the human body image tracking set with the same target frame identification, including:

[0097] Determine whether the first association condition, the second association condition, the third association condition, and the fourth association condition are all established simultaneously;

[0098] Among them, the first association condition is that the abscissa of the upper left corner of the face image is greater than the abscissa of the upper left corner of the human body image, the second association condition is that the ordinate of the upper left corner of the face image is greater than the ordinate of the upper left corner of the human body image, the third association condition is that the sum of the abscissa of the upper left corner of the face image and the width of the face image is less than the sum of the abscissa of the upper left corner of the human body image and the width of the human body image, and the fourth association condition is that the sum of the ordinate of the upper left corner of the face image and the height of the face image is less than the sum of the ordinate of the upper left corner of the human body image and one-third of the height of the human body image;

[0099] If so, determine that the attribute information of the face image matches the attribute information of the human body image.

[0100] Suppose the attribute information of the face image is (x1, y1, w1, h1), and the attribute information of the human body image is (x2, y2, w2, h2). x represents the abscissa value of the upper left corner point of the corresponding image, y represents the ordinate value of the upper left corner point of the corresponding image, w represents the width of the corresponding image, and h represents the height of the corresponding image. If x1>x2, y1>y2, x1+w1<x2+w2, x1+h1<x2+h2 / 3; when the four conditions are all established simultaneously, the attribute information of the face image matches the attribute information of the human body image, and the face image and the human body image belong to the same person.

[0101] On Figure 1 this basis, regarding how to obtain the target face image and the target human body image tracking set, the embodiment of the present application also provides a possible implementation manner. Please refer to Figure 7 , the human body recognition system management method further includes:

[0102] S107. Perform face detection on the target person in the target video, track the face frame, and perform human body detection on the target person in the target video, and track the human body frame.

[0103] Optionally, a face detection algorithm is used to detect faces and track face bounding boxes. Among them, the face detection algorithm is a common face detection algorithm such as MTCNN or RetinaFace; the face bounding box uses the IOU tracking algorithm to put the same face bounding boxes in consecutive frames into the same tracking chain. A human detection algorithm is used to detect humans and track human bounding boxes. Among them, the human detection algorithm is a common object detection algorithm such as YOLO or SSD; the human bounding box can use the IOU tracking algorithm to put the same human bounding boxes in consecutive frames into the same tracking chain.

[0104] S108. Establish an association relationship between the face bounding box and the human bounding box.

[0105] Among them, the method for establishing the association is specifically as follows. Assume that the attribute information of the face bounding box is (x1, y1, w1, h1), and the attribute information of the human bounding box is (x2, y2, w2, h2). x and y respectively represent the horizontal and vertical coordinate values of the upper left corner point of the box, and w and h respectively represent the width and height of the box. If the four conditions x > x2, y > y2, x1 + w1 < x2 + w2, and x1 + h1 < x2 + h2 / 3 are all satisfied, then this face bounding box and the human bounding box belong to the same person, and an association relationship is established between the face bounding box and the human bounding box. The coordinate system in S108 is the same as the coordinate system described above.

[0106] S109. Determine the face bounding box with an association relationship as the target face image, and determine the human bounding box tracking chain corresponding to the human bounding box with an association relationship as the target human image tracking set.

[0107] On the Figure 1 basis, regarding how to further improve the recognition accuracy of the human recognition system, the embodiments of the present application also provide a possible implementation method. Please refer to Figure 8 , the human recognition system management method further includes:

[0108] S101. Delete the human registration images whose storage time exceeds the preset time threshold.

[0109] Optionally, when the storage time of the human registration images in the human recognition system is too long, the target person may have grown or their appearance has changed significantly during this period. Therefore, the reference significance of the human registration images with too long storage time is relatively small, and misrecognition may occur when performing human recognition through the human registration images with too long storage time. Therefore, in order to improve the recognition accuracy, it is necessary to delete the human registration images whose storage time exceeds the preset time threshold.

[0110] In summary, the human body recognition system management method provided in this application firstly, when no corresponding human body registration image exists for a registered human body identity, any target human body image from the target human body image tracking set is automatically used as the human body registration image. This overcomes the technical problem of requiring users to actively provide human body registration images, reduces user intervention, saves manpower and time, and is more convenient. Secondly, if the perspective of the human body registration image is different from that of the target human body image, in order to enrich the diversity of recognition templates in the human body recognition system and improve the recognition accuracy, the target human body image is updated to a new human body registration image. When the similarity between the target human body image and the first human body registration image is less than or equal to a threshold, human body images belonging to the same template person but with significant differences are used as new recognition templates for the human body recognition system, which helps improve the recognition accuracy. Finally, to improve recognition accuracy, human body registration images that have been saved for more than a preset time threshold need to be deleted.

[0111] Please see Figure 9 , Figure 9 The present application provides a human body recognition system management device, which is optionally applied to the electronic device described above.

[0112] The human body recognition system management device includes an identity determination unit 301 and a registration unit 302.

[0113] The identity determination unit 301 is used to determine the human registration identity of the corresponding user based on the target facial image. Optionally, the identity determination unit 301 can perform the above-described S1010.

[0114] The registration unit 302 is used to determine any target human image in the target human image tracking set as the human registration image corresponding to the human registration identity when there is no corresponding human registration image for the human registration identity.

[0115] Among them, the target human body image tracking set is a human body image tracking set associated with the target human face image, and the target human body image and the target human face image belong to the same target person.

[0116] The registration unit is also used to determine whether there is a corresponding human registration image for the human registration identity; if so, it sequentially determines whether there is a corresponding first human registration image for each target human image.

[0117] Among them, the first human registration image is a human registration image with the same viewpoint as the target human image in the human image tracking set;

[0118] If it does not exist, the registration unit is also used to update the target human body image to a new human body registration image;

[0119] If it exists, the registration unit is also used to determine whether the similarity between the target human image and the first registered human image is greater than the threshold.

[0120] If the value is less than 1, the registration unit is also used to update the target human image to a new human registration image.

[0121] Optionally, the registration unit 302 may execute S1012-S1017 as described above.

[0122] The identity determination unit 301 is also used to obtain the viewpoint of each target human image based on a human attribute recognition algorithm. Optionally, the identity determination unit 301 can perform the above-described S1011.

[0123] The registration unit 302 is further configured to identify any one of the target human images from the same viewpoint in the target human image tracking set as the human registration image corresponding to the human registration identity. Optionally, the registration unit 302 may execute S1013-1 as described above.

[0124] The human body recognition system management device provided in this application embodiment can achieve... Figures 1 to 8 The various processes implemented in the human body recognition system management method in the method embodiments will not be described again here to avoid repetition.

[0125] This application provides an electronic device, which can be a server. The server includes a processor and a memory, the memory storing at least one instruction, at least one program, code set, or instruction set. The at least one instruction, at least one program, code set, or instruction set is loaded and executed by the processor to implement the human body recognition system management method provided in the above method embodiments.

[0126] Memory can be used to store software programs and modules. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory. Memory can primarily include a program storage area and a data storage area. The program storage area can store the operating system, application programs required for the functions, etc.; the data storage area can store data created based on the use of the device, etc. Furthermore, memory can include high-speed random access memory, and can also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, memory can also include a memory controller to provide the processor with access to the memory.

[0127] Figure 10 This is a hardware structure block diagram of a server for a human body recognition system management method provided in an embodiment of this application. For example... Figure 10As shown, the server 1100 can vary significantly due to different configurations or performance. It may include one or more Processing Units (CPUs) 1110 (processors 1110 may include, but are not limited to, microprocessors (MCUs) or programmable logic devices (FPGAs), a memory 1130 for storing data, and one or more storage media 1120 (e.g., one or more mass storage devices) for storing application programs 1123 or data 1122. The memory 1130 and storage media 1120 may be temporary or persistent storage. The program stored in the storage media 1120 may include one or more modules, each module may include a series of instruction operations on the server. Furthermore, the processor 1110 may be configured to communicate with the storage media 1120 and execute the series of instruction operations in the storage media 1120 on the server 1100. Server 1100 may also include one or more power supplies 1160, one or more wired or wireless network interfaces 1150, one or more input / output interfaces 1140, and / or one or more operating systems 1121, such as Windows Server™, MacOSX™, Unix™, Linux™, FreeBSD™, etc.

[0128] The input / output interface 1140 can be used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of server 1100. In one example, the input / output interface 1140 includes a network interface controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the input / output interface 1140 may be a radio frequency (RF) module used for wireless communication with the Internet.

[0129] Those skilled in the art will understand that Figure 10 The structure shown is for illustrative purposes only and does not limit the scope of the structure described above. For example, server 1100 may also include... Figure 10 The more or fewer components shown, or having the same Figure 10 The different configurations shown.

[0130] This application also provides a storage medium, which is a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described human body recognition system management method embodiments and achieves the same technical effects. To avoid repetition, it will not be described again here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0131] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0132] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0133] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for managing a human body recognition system, characterized in that, The method includes: Determine the registered identity of the corresponding user based on the target facial image; Determine whether a corresponding human registration image exists for the registered human identity; If so, then sequentially determine whether each target human image has a corresponding first human registration image; Wherein, the first human registration image is a human registration image with the same viewpoint as the target human image in the human image tracking set; If no corresponding first human body registration image exists, the target human body image is updated to a new human body registration image; If a corresponding first human body registration image exists, then determine whether the similarity between the target human body image and the first human body registration image is greater than a threshold. If it is smaller than the target human image, then the target human image is updated to a new human registration image; If no corresponding human registration image exists for the registered human identity, any target human image in the target human image tracking set will be identified as the human registration image corresponding to the registered human identity. The target human body image tracking set is a human body image tracking set associated with the target human face image, wherein the target human body image and the target human face image belong to the same target person.

2. The human body recognition system management method as described in claim 1, characterized in that, Before identifying any target human image in the human image tracking set as the human registration image corresponding to the human registration identity, the method further includes: The perspective of each target human body image is obtained based on the human body attribute recognition algorithm; The step of identifying any target human image in the human image tracking set as the human registration image corresponding to the human registration identity includes: Any one of the target human images from the same viewpoint in the target human image tracking set is identified as the human registration image corresponding to the human registration identity.

3. The human body recognition system management method as described in claim 1, characterized in that, The process of determining the corresponding user's registered identity based on the target facial image includes: The target facial features of the target facial image are obtained based on a facial recognition algorithm; A second face registration image is determined from a face feature database, wherein the second face registration image is a pre-registered face image with the highest similarity to the target face feature, and the face feature database includes at least one set of pre-registered face images; The registration identity corresponding to the second facial registration image is determined as the human registration identity of the target user.

4. The human body recognition system management method as described in claim 1, characterized in that, Before determining the corresponding user's registered identity based on the target facial image, the method includes: Perform face detection and track face bounding boxes for target individuals in the target video, and perform human body detection and track human body bounding boxes for target individuals in the target video. Establish an association between the face frame and the body frame; The face bounding boxes with related relationships are identified as the target face images, and the human bounding box tracking chains corresponding to the human bounding boxes with related relationships are identified as the target human image tracking set.

5. The human body recognition system management method as described in claim 1, characterized in that, Before determining the corresponding user's registered identity based on the target facial image, the method further includes: Face detection is performed on the target video, and the face images of different target persons are stored in the corresponding face image tracking set. The face image carries attribute information and target frame identifier, and the target frame identifier carried by the face image is the identifier of the video frame in which the face image is detected. Human detection is performed on the target video, and the human images of different target persons are stored in the corresponding human image tracking set. The human image carries attribute information and target frame identifier, and the target frame identifier carried by the human image is the identifier of the video frame in which the human image is detected. Based on the attribute information and target frame identifier carried by the face image and the body image respectively, determine the body image tracking set associated with each face image tracking set; Identify any face image from the face image tracking set as the target face image; The human body image tracking set associated with the face image tracking set is determined as the target human body image tracking set.

6. The human body recognition system management method as described in claim 5, characterized in that, The step of determining the human image tracking set associated with each face image tracking set based on the attribute information and target frame identifier carried by the face image and the human body image respectively includes: Sequentially determine whether the attribute information of the face images in the face image tracking set matches the attribute information of the human images in the human image tracking set that carry the same target frame identifier; If so, then the face image tracking set is determined to be associated with the human body image tracking set; If not, then repeat the process of checking whether the attribute information of the face images in the face image tracking set matches the attribute information of the human images in the human image tracking set that carry the same target frame identifier.

7. The human body recognition system management method as described in claim 6, characterized in that, The attribute information includes the x and y coordinates of the top left corner of the corresponding image, as well as the width and height of the corresponding image. The step of determining whether the attribute information of the face images in the face image tracking set matches the attribute information of the human images in the human image tracking set carrying the same target frame identifier includes: Determine whether the first association condition, the second association condition, the third association condition, and the fourth association condition are all true simultaneously; The first association condition is that the x-coordinate of the top left corner of the face image is greater than the x-coordinate of the top left corner of the human body image; the second association condition is that the y-coordinate of the top left corner of the face image is greater than the y-coordinate of the top left corner of the human body image; the third association condition is that the sum of the x-coordinate of the top left corner of the face image and the width of the face image is less than the sum of the x-coordinate of the top left corner of the human body image and the width of the human body image; and the fourth association condition is that the sum of the y-coordinate of the top left corner of the face image and the height of the face image is less than the sum of one-third of the y-coordinate of the top left corner of the human body image and the height of the human body image. If so, then it is determined that the attribute information of the face image matches the attribute information of the human body image.

8. The human body recognition system management method as described in claim 1, characterized in that, The method further includes: Human registration images that have been saved for longer than a preset time threshold will be deleted.

9. A human body recognition system management device, characterized in that, The device includes: The identity determination unit is used to determine the human registration identity of the corresponding user based on the target facial image; The registration unit is used to determine whether a corresponding human registration image exists for the registered human identity; if so, it sequentially determines whether a corresponding first human registration image exists for each target human image; wherein, the first human registration image is a human registration image with the same viewpoint as the target human image in the human image tracking set; if not, the target human image is updated to a new human registration image; if so, it determines whether the similarity between the target human image and the first human registration image is greater than a threshold; if less, the target human image is updated to a new human registration image. The registration unit is also used to determine any target human image in the target human image tracking set as the human registration image corresponding to the human registration identity when there is no corresponding human registration image for the human registration identity. The target human body image tracking set is a human body image tracking set associated with the target human face image, wherein the target human body image and the target human face image belong to the same target person.

10. The human body recognition system management device as described in claim 9, characterized in that, The identity determination unit is also used to obtain the perspective of each target human image based on the human attribute recognition algorithm; The registration unit is also used to determine any one of the target human images from the same viewpoint in the target human image tracking set as the human registration image corresponding to the human registration identity.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-8.

12. An electronic device, characterized in that, include: Processor and memory, the memory being used to store one or more programs; When the one or more programs are executed by the processor, the method as described in any one of claims 1-8 is implemented.

Images