WO2022043741A1 - Network training method and apparatus, person re-identification method and apparatus, storage medium, and computer program - Google Patents

Network training method and apparatus, person re-identification method and apparatus, storage medium, and computer program Download PDF

Info

Publication number
WO2022043741A1
WO2022043741A1 PCT/IB2020/060047 IB2020060047W WO2022043741A1 WO 2022043741 A1 WO2022043741 A1 WO 2022043741A1 IB 2020060047 W IB2020060047 W IB 2020060047W WO 2022043741 A1 WO2022043741 A1 WO 2022043741A1
Authority
WO
WIPO (PCT)
Prior art keywords
network
pedestrian
identification
identification network
person
Prior art date
Application number
PCT/IB2020/060047
Other languages
French (fr)
Chinese (zh)
Inventor
庄伟铭
张学森
张帅
伊帅
Original Assignee
商汤国际私人有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 商汤国际私人有限公司 filed Critical 商汤国际私人有限公司
Publication of WO2022043741A1 publication Critical patent/WO2022043741A1/en

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/60—Protecting data
    • G06F21/62—Protecting access to data via a platform, e.g. using keys or access control rules
    • G06F21/6218—Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database
    • G06F21/6245—Protecting personal data, e.g. for financial or medical purposes

Definitions

  • Pedestrian re-identification also known as pedestrian re-identification
  • pedestrian re-identification is a technology that uses computer vision technology to determine whether there is a specific pedestrian in an image or a video sequence.
  • pedestrian re-identification technology has been widely used in many fields and industries, such as intelligent video detection, intelligent security and so on.
  • SUMMARY Embodiments of the present disclosure propose a network training and pedestrian re-identification method and device, a storage medium, and a technical solution for a computer program. According to an aspect of the embodiments of the present disclosure, a network training method is provided.
  • the method is applied to a cloud server, where the cloud server includes a first person re-identification network, and the method includes: sending first network parameters corresponding to the first pedestrian re-identification network; receiving second network parameters returned by the multiple edge servers, wherein, for any of the edge servers, the edge servers include A second person re-identification network, an identity classification network and a local image dataset, the second person re-identification network and the first person re-identification network have the same network structure, and the second network parameter is the edge obtained after the end server trains the second pedestrian re-identification network according to the local image data set, the identity classification network and the first network parameters; Two network parameters, updating the first pedestrian re-identification network to obtain the updated first pedestrian re-identification network.
  • the first pedestrian re-identification network is updated according to the second network parameters returned by the multiple edge servers to obtain the updated first row
  • the person re-identification network includes: receiving weights corresponding to the second network parameters returned by the plurality of edge servers, wherein, for any of the edge servers, the weights corresponding to the second network parameters are the Determined by the edge server according to the second pedestrian re-identification network before training and the second pedestrian re-identification network after training; according to the weights corresponding to the second network parameters returned by the multiple edge servers , performing a weighted average on the second network parameters returned by the plurality of edge servers to obtain the updated first network parameters; according to the updated first network parameters, the first pedestrian The re-identification network is updated to obtain the updated first pedestrian re-identification network.
  • the method further includes: sending a shared image data set to the multiple edge servers; receiving pseudo tags returned by the multiple edge servers, wherein for any of the edge servers an end server, where the pseudo label is generated by the side server according to the shared image dataset and the trained second pedestrian re-identification network; according to the shared image dataset and the multiple side servers For the returned pseudo-label, the updated first pedestrian re-identification network is trained to obtain the trained first pedestrian re-identification network.
  • the returning according to the shared image data set and the multiple side servers training the updated first pedestrian re-identification network, and obtaining the trained first pedestrian re-identification network comprising: determining, according to the pseudo-tags returned by the multiple edge servers, Average pseudo-label; train the updated first person re-identification network according to the shared image data set and the average pseudo-label, to obtain the trained first person re-identification network.
  • a network training method is provided, the method is applied to an edge server, and the edge server includes a second person re-identification network, an identity classification network and a local image dataset, so The method includes: receiving a first network parameter corresponding to a first pedestrian re-identification network sent by a cloud server, wherein the first pedestrian-re-identification network and the second pedestrian-re-identification network have the same network structure; The local image data set, the identity classification network and the first network parameters are used to train the second pedestrian re-identification network, and the trained second pedestrian re-identification network is obtained, wherein the first pedestrian re-identification network is obtained.
  • the two-person re-identification network corresponds to the second network parameter; and the second network parameter is sent to the cloud server.
  • the second pedestrian re-identification network is trained according to the local image data set, the identity classification network and the first network parameters to obtain the trained
  • the second person re-identification network includes: training the second person re-identification network and the identity classification network according to the local image data set and the first network parameters, to obtain the trained second person re-identification network.
  • a person re-identification network and the trained identity classification network is further includes: storing the trained identity classification network in the edge server.
  • the local image data set includes image data corresponding to multiple identities; and the dimension of the identity classification network is related to the number of the multiple identities.
  • the method further includes: receiving a shared image data set sent by the cloud server; generating a pseudo-label according to the shared image data set and the trained second person re-identification network ; Send the pseudo tag to the cloud server.
  • the method further includes: determining a first feature vector according to the second pedestrian re-identification network before training and the local image data set, and determining a first feature vector according to the second pedestrian re-identification network after training and the local image data set;
  • the two-person re-identification network and the local image data set determine a second feature vector; determine the cosine distance between the first feature vector and the second feature vector; determine the second feature vector according to the cosine distance the weight corresponding to the network parameter; sending the weight corresponding to the second network parameter to the cloud server.
  • the edge server is an image acquisition device; and the local image data set is acquired according to the image acquisition device.
  • the edge server is connected to at least one image acquisition device, and the edge server and the at least one image acquisition device are located in the same geographical area; the local image data set is the The edge server is obtained from the at least one image acquisition device.
  • a pedestrian re-identification method including: performing pedestrian re-identification processing on at least one frame of images to be identified obtained within a target geographic area through a target pedestrian re-identification network, and determining the pedestrian The re-identification result; wherein, the target pedestrian re-identification network is obtained by training the above-mentioned network training method.
  • the target person re-identification network is an updated first person re-identification network or a trained first person re-identification network.
  • the target pedestrian The re-identification network is the second person re-identification network after training.
  • a network training device is provided, the network training device is applied to a cloud server, the cloud server includes a first person re-identification network, and the device includes: a sending part, which is is configured to send the first network parameters corresponding to the first pedestrian re-identification network to the plurality of edge servers; the receiving part is configured to receive the second network parameters returned by the multiple edge servers, wherein for any a described edge server,
  • the edge server includes a second person re-identification network, an identity classification network and a local image dataset, the second person re-identification network and the first person re-identification network have the same network structure, and the first person re-identification network has the same network structure.
  • the second network parameter is obtained by the side server after training the second person re-identification network according to the local image data set, the identity classification network and the first network parameter; the update part is configured as According to the second network parameters returned by the multiple edge servers, the first pedestrian re-identification network is updated to obtain the updated first pedestrian re-identification network.
  • a network training apparatus is provided, the apparatus is applied to an edge server, and the edge server includes a second pedestrian re-identification network, an identity classification network and a local image dataset, so
  • the apparatus includes: a receiving part configured to receive a first network parameter corresponding to a first pedestrian re-identification network sent by a cloud server, wherein the first pedestrian-re-identification network and the second pedestrian-re-identification network have the same network structure; the network training part is configured to train the second pedestrian re-identification network according to the local image data set, the identity classification network and the first network parameters, and obtain the trained the second pedestrian re-identification network, wherein the second pedestrian re-identification network corresponds to a second network parameter; the sending part is configured to send the second network parameter to the cloud server according to an embodiment of the present disclosure.
  • a pedestrian re-identification device comprising: a pedestrian re-identification part configured to perform pedestrian re-identification processing on at least one frame of images to be identified obtained within a target geographical area through a target pedestrian re-identification network, and determine Pedestrian re-identification result; wherein, the target pedestrian re-identification network is obtained by training the above-mentioned network training method.
  • an electronic device comprising: a processor; a memory configured to store instructions executable by the processor; wherein the processor is configured to invoke the instructions stored in the memory, to perform the above method.
  • a computer-readable storage medium having computer program instructions stored thereon, the computer program instructions implementing the foregoing method when executed by a processor.
  • a computer program is provided, including computer-readable code, when the computer-readable code is executed in an electronic device, the processor in the electronic device implements the foregoing method when executed.
  • the cloud server including the first pedestrian re-identification network the first network parameters corresponding to the first pedestrian re-identification network are sent to multiple edge servers, and the multiple edge servers are received.
  • the returned second network parameter wherein, for any edge server, the edge server includes a second person re-identification network, an identity classification network and a local image dataset that have the same network structure as the first person re-identification network,
  • the second network parameter is obtained by the edge server after training the second pedestrian re-identification network according to the local image data set, the identity classification network and the first network parameters, and then according to the second network parameters returned by the multiple edge servers, to
  • the first person re-identification network is updated, and the updated first person re-identification network is obtained.
  • the cloud server combines multiple side servers to train the pedestrian re-identification network.
  • FIG. 1 shows a flow chart of a network training method according to an embodiment of the present disclosure
  • Fig. 2 shows a structural diagram of an exemplary network training provided according to an embodiment of the present disclosure
  • Fig. 3 shows an embodiment according to the present disclosure
  • a flowchart of a network training method according to an embodiment of the present disclosure
  • FIG. 4 shows an exemplary schematic diagram of determining the weight of a second network parameter provided according to an embodiment of the present disclosure
  • FIG. 5 shows an exemplary schematic diagram provided according to an embodiment of the present disclosure.
  • Fig. 6 shows an exemplary network structure diagram of a cloud server-side server-terminal device provided according to an embodiment of the present disclosure
  • FIG. 7 shows an exemplary network training provided according to an embodiment of the present disclosure 8 shows a block diagram of a network training apparatus according to an embodiment of the present disclosure;
  • FIG. 9 shows a block diagram of a network training apparatus according to an embodiment of the present disclosure;
  • FIG. 11 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
  • exemplary is used exclusively herein to mean “serving as an example, embodiment, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.
  • the term “and/or” in this article is only an association relationship to describe associated objects, which means that there can be three kinds of relationships, for example, A and/or B, which can mean: A exists alone, A and B exist simultaneously, and A and B exist independently B these three cases.
  • at least one herein refers to any one of multiple or any combination of at least two of multiple, for example, including at least one of A, B, and C, may mean including from A, B, and C. Any one or more elements selected from the set of B and C.
  • FIG. 1 shows a flowchart of a network training method according to an embodiment of the present disclosure.
  • the network training method can be executed by a cloud server, and the cloud server includes a first pedestrian re-identification network.
  • the network training method may be implemented by a cloud server invoking computer-readable instructions stored in a memory. As shown in FIG.
  • the method may include: In step S11, sending first network parameters corresponding to the first pedestrian re-identification network to a plurality of edge servers.
  • the second network parameters returned by multiple edge servers are received, wherein, for any edge server, the edge server includes a second pedestrian re-identification network, an identity classification network and a local image data set, the second pedestrian The re-identification network and the first person re-identification network have the same network structure, and the second network parameters are obtained after the edge server trains the second person re-identification network according to the local image data set, the identity classification network and the first network parameters. of.
  • the first pedestrian re-identification network is updated according to the second network parameters returned by the multiple edge servers to obtain an updated first pedestrian re-identification network.
  • the cloud server combines multiple side servers to train the pedestrian re-identification network.
  • the image data set is still saved in the side server and does not need to be uploaded to the cloud server, so that the pedestrian re-identification network can be effectively trained while protecting the data. privacy.
  • communication bandwidth can be effectively saved.
  • the cloud server can combine multiple side servers to perform network training based on the federated learning algorithm.
  • a person re-identification network is jointly trained by multiple communities, and each community is set up with a side server.
  • the image data set (the image data set collected by the image acquisition equipment set in the community or nearby) still remains It is stored in the community (local side server) and does not need to be uploaded to other communities (other side servers), thus protecting data privacy.
  • different edge servers have different data sizes.
  • the data between the end servers is heterogeneous.
  • the traditional federated learning algorithm uses multiple side servers for network training, the weights of the second network parameters obtained by the network training in the side servers are set according to the amount of data in different side servers.
  • the cloud server updates the first pedestrian re-identification network by using the weights of the second network parameters obtained in the side server based on this weight determination method. , which will lead to lower accuracy of the updated first person re-identification network.
  • the first pedestrian re-identification network is updated according to the second network parameters returned by multiple edge servers, and the updated first pedestrian re-identification network is obtained, including: receiving multiple The weight corresponding to the second network parameter returned by the edge server, wherein, for any edge server, the weight corresponding to the second network parameter is the weight of the edge server according to the second pedestrian re-identification network before training and the second pedestrian weight after training.
  • weighted average is performed on the second network parameters returned by the multiple edge servers to obtain the updated first network parameters; according to the updated first network parameters;
  • the first network parameters are updated, and the first pedestrian re-identification network is updated to obtain the updated first pedestrian re-identification network.
  • the weight of the second network parameter sent by the side server is determined by the side server according to the second pedestrian re-identification network before training and the second pedestrian re-identification network after training, that is, the weight of the second network parameter It is determined according to the training effect of the edge server, so that the cloud server performs a weighted average of the second network parameters returned by the multiple edge servers according to the weights corresponding to the second network parameters returned by each edge server to obtain a higher-precision network parameter.
  • the updated first network parameters, and then after the first pedestrian re-identification network is updated according to the updated first network parameters, the accuracy of the updated first pedestrian re-identification network is effectively improved.
  • the data between different edge servers is heterogeneous, which leads to The performance of the trained second person re-identification network obtained by training the local image data set, the identity classification network and the first network parameters is better than the updated first person re-identification obtained by training the cloud server in conjunction with multiple side servers.
  • the internet since the local image datasets in different edge servers are collected in different scenarios (lighting, angle), the data between different edge servers is heterogeneous, which leads to The performance of the trained second person re-identification network obtained by training the local image data set, the identity classification network and the first network parameters is better than the updated first person re-identification obtained by training the cloud server in conjunction with multiple side servers.
  • the second person re-identification network trained in each side server can be used as the teacher network
  • the updated first person re-identification network in the cloud server can be used as the student network
  • the teacher network can be used to The student network is trained (using the updated second person re-identification network to train the updated first-person re-identification network) to improve the stability and convergence of the first-person re-identification network training process.
  • the method further includes: sending a shared image data set to multiple edge servers; receiving pseudo tags returned by the multiple edge servers, wherein, for any edge server, the pseudo tag is the edge server
  • the server is generated according to the shared image data set and the trained second person re-identification network; according to the shared image data set and the pseudo-labels returned by multiple side servers, the updated first person re-identification network is trained to obtain The first person re-identification network after training.
  • the cloud server receives the pseudo-label returned by each side server. Since the pseudo-label is generated by the side-end server according to the shared image data set and the trained second person re-identification network, the pseudo-label can be used to represent the trained second person re-identification network.
  • the network characteristics of the person re-identification network therefore, according to the shared image data set and the pseudo-labels returned by multiple side servers, the updated first person re-identification network is trained, which is equivalent to synthesizing the network characteristics of each side server.
  • the updated first person re-identification network is trained, so that the stability and convergence of the training process of the first person re-identification network can be effectively improved.
  • the shared image dataset refers to an image dataset that both the cloud server and each side server can use for network training.
  • FIG. 2 shows a structural diagram of an exemplary network training provided according to an embodiment of the present disclosure. As shown in FIG.
  • the second person re-identification network trained in the multiple edge servers constitutes a teacher network 1, a teacher network 2, and a teacher network N, where N is the number of multiple edge servers, and N>1 .
  • the updated first person re-identification network in the cloud server constitutes the student network.
  • the teacher network 1 uses the shared image data set to generate pseudo-labels and sends the pseudo-labels to the cloud server; the teacher network 2 uses the shared image data set to generate pseudo-labels, and sends the pseudo-labels to the cloud server; Image datasets generate pseudo-labels, and send pseudo-labels to The pedestrian re-identification network is trained to obtain a trained second pedestrian re-identification network corresponding to the second network parameters, and then the second network parameters are sent to the cloud server.
  • the edge server is connected to at least one image acquisition device, and the edge server and the at least one image acquisition device are located in the same geographical area; the local image data set is obtained by the edge server from the at least one image acquisition device. obtained.
  • an edge server may be provided within the geographical area, and in this case, the at least one image acquisition device does not need to have storage capability and computing power.
  • the edge server is connected to each image acquisition device, and further acquires images from each image acquisition device to construct a local image data set.
  • the edge server receives the first network parameters corresponding to the first pedestrian re-identification network sent by the cloud server, and trains the second pedestrian re-identification network according to the local image data set and the first network parameters, and obtains a parameter corresponding to the second network parameters.
  • the trained second person re-identification network and then sends the second network parameters to the cloud server.
  • the second person re-identification network is trained according to the local image data set, the identity classification network and the first network parameters, and the trained second person re-identification network is obtained, including: according to the local image The data set and the first network parameters are used to train the second person re-identification network and the identity classification network, and the trained second person re-identification network and the trained identity classification network are obtained.
  • the local image dataset includes image data corresponding to multiple identities; the dimension of the identity classification network is related to the number of multiple identities.
  • the trained person re-identification network is a network for identifying images, therefore, in the process of training the person re-identification network, a local image dataset including image data corresponding to multiple identities needs to be used, and the identity Classification network, the dimension of the identity classification network is related to the number of multiple identities included in the local image dataset. For example, if the local image dataset includes image data corresponding to 100 identities, the dimension of the identity classification network is 100. That is, the identity classification network includes 100 different identity classes.
  • the side-end server constructs the local second person re-identification network and the identity classification network as a combined network, and uses the first network parameters received from the cloud server and the local image data set to train the combined network, and then obtains the combined network after training network, wherein the trained combined network includes a trained second person re-identification network and a trained identity classification network, and the trained second person re-identification network corresponds to a second network parameter. Further, the edge server sends the second network parameter to the cloud server. Since the first person re-identification network and the second person re-identification network have the same network structure, the first person re-identification network can be updated by using the second network parameters.
  • the method further includes: storing the trained identity classification network in the edge server. Since the first person re-identification network trained in the cloud server does not need to use the classifier network in the actual pedestrian re-identification process, in order to save the communication bandwidth and ensure that the cloud is used in the joint training process based on the federated learning algorithm The network structure between the server and the side server is consistent. The side server only sends the second network parameters corresponding to the trained second person re-identification network to the cloud server, and stores the trained identity classification network on the side server. local.
  • the method further includes: receiving a shared image data set sent by the cloud server; generating a pseudo-label according to the shared image data set and the trained second person re-identification network; sending the pseudo-label to the cloud server .
  • FIG. 2 Still taking the above FIG. 2 as an example, as shown in FIG.
  • the side server receives the shared image data set sent by the cloud server, and uses the shared image data set and the locally trained second person re-identification network to generate pseudo labels, and then the edge
  • the terminal server sends a pseudo-label to the cloud server, because the pseudo-label can be used to represent the network characteristics of the trained second pedestrian re-identification network, so that the cloud server can re-identify the updated first pedestrian in the cloud server according to the pseudo-label.
  • the first person re-identification network after training is obtained after the recognition network performs network training, so that the network performance of the first person re-identification network after training is closer to that of the second person re-identification network after training in the side server.
  • the edge server A obtains images from the image acquisition device 1 and the image acquisition device 2 respectively to construct a local image dataset.
  • the side server B is connected to the terminal device 3, the terminal device 4 and the terminal device 5, and the terminal device 3, the terminal device 4 and the terminal device 5 are image acquisition devices (the image acquisition device 3, the image acquisition device 4 and the image acquisition device 5, for example , the image acquisition device is a camera), the edge server B, the image acquisition device 3, the image acquisition device 4, and the image acquisition device 5 are set in the same geographical area (for example, the same community, or the same company), and the edge server B is Image capture device 3, image capture device 4, and image capture device 5 acquire images to construct a local image dataset.
  • the cloud server combines two side servers (side server A and side server B) to train the pedestrian re-identification network. During the training process, the image data set is still stored locally on each side server without uploading to the cloud server.
  • Cloud-edge architecture The cloud server communicates directly with the smart cameras, and the cloud coordinates multiple smart cameras to train at the same time.
  • the smart camera caches pictures on the edge, and deletes and cleans them regularly to reduce the storage pressure on the edge server.
  • this architecture requires smart cameras to have certain computing power, storage and communication capabilities.
  • the edge gateway (that is, the above-mentioned edge-end server) is connected to multiple smart cameras, the cloud server is connected to multiple edge gateways, and the pedestrian re-identification training image is transmitted from the smart camera to the edge gateway, and cached in the edge gateway, the edge gateway Federated learning training with cloud servers.
  • the data remains at the edge gateway, and data privacy can still be protected.
  • typical application scenarios such as multiple communities jointly training a pedestrian re-identification model, each community has an edge gateway connected to multiple smart cameras, through federated learning, the data is still retained in the community, not transmitted to other Community or cloud server to protect data privacy.
  • the multiple side servers may also be partly image acquisition devices (for example, smart cameras) that communicate directly with the cloud server. ), part of which is an edge server connected to at least one image acquisition device, which is not specifically limited in this embodiment of the present disclosure.
  • FIG. 7 shows a structural diagram of an exemplary network training provided according to an embodiment of the present disclosure. As shown in FIG. 7 , the cloud server can communicate with multiple side servers (side server 1, side server 2, >... , side server N), and the first line included in the cloud server The person re-identification network and the second person re-identification network included in the edge server have the same network structure.
  • Each edge server also includes a local image dataset and an identity classification network.
  • the cloud server sends the first network parameters corresponding to the first pedestrian re-identification network to the multiple edge servers, and each edge server uses the local image data set and the identity classification network to re-identify the second pedestrian after receiving the first network parameters.
  • the network is trained to obtain a trained second person re-identification network corresponding to the second network parameters, and a trained identity classification network.
  • each side server only sends the second network parameters corresponding to the trained second person re-identification network to the cloud server.
  • the cloud server updates the first pedestrian re-identification network according to the received second network parameters returned by the multiple edge servers, and obtains the updated first pedestrian re-identification network. Then, the first network parameters corresponding to the updated first pedestrian re-identification network are sent to multiple side servers for cyclic training, until the updated recognition accuracy of the first pedestrian re-identification network in the cloud server reaches the threshold, or the cycle is repeated. When the number of training reaches the preset number, the training ends.
  • the general federated learning algorithm (Federated Averaging, FedAvg) requires that the multi-party models (person re-identification deep learning model, ie the above-mentioned person re-identification network) to be synchronized must be exactly the same.
  • the classifier layer of the deep learning model for person re-identification depends on how many different pedestrians each party's data contains, so the classifier layers of the multi-party models participating in training may be different, resulting in participation in federated learning.
  • Multi-party models may be different, so the federated learning algorithm FedAvg is not applicable in the above application scenarios.
  • the federated learning algorithm is improved in the present disclosure, that is, the models of multiple parties participating in the federated learning are allowed to be partially different, so the federated learning can be better applied to the training of person re-identification.
  • each edge server can use a The weight determination method of the training effect to determine the weight of the second network parameter corresponding to the trained second pedestrian re-identification network, so that the cloud server combines the second network parameters returned by each side server to carry out the first pedestrian re-identification network. After the update, the updated first person re-identification network with higher accuracy is obtained.
  • the specific steps of the weight determination method based on the training effect are as described in the relevant parts of the foregoing embodiments, and are not repeated here. Since the local image datasets in different side servers are collected under different scenes (lighting, angle), the data between different side servers is heterogeneous, which leads to each side server according to the local image dataset. The performance of the trained second person re-identification network obtained by training with the first network parameters is better than that of the updated first person re-identification network trained by the cloud server in conjunction with multiple side servers.
  • the knowledge evaporation algorithm can be used, based on the updated second person re-identification network and shared image data set in each side server, and the cloud server
  • the updated first person re-identification network is trained, thereby effectively improving the stability and convergence of the training process of the first person re-identification network.
  • the specific training process based on the knowledge steaming library algorithm is described in the relevant part of the above-mentioned embodiment, and details are not repeated here.
  • the present disclosure proposes a method of using a knowledge library, taking the local model of multiple parties participating in federated learning as the teacher model, and the model of the cloud server as the student model, and using the knowledge library method to better integrate the knowledge of the teacher model It is passed to the student model, thereby improving the stability and convergence of the model training.
  • the weight determination method based on the training effect and the network training based on the knowledge evaporation algorithm can be used separately, or can be used in combination.
  • Embodiments of the present disclosure There is no specific limitation on this.
  • the pedestrian re-identification network can be jointly trained based on the network structure shown in FIG.
  • the cloud server communicates directly, and the data is still stored locally during the training process, and there is no need to upload it to the cloud server, so that the pedestrian re-identification network can be obtained through effective training in the cloud server, while protecting the data privacy of multiple companies or institutions.
  • company A provides company B with a pedestrian re-identification network training service.
  • each image acquisition device for example, a smart camera
  • company A can jointly train the pedestrian re-identification network based on the network structure shown in Figure 7, company A can be used as a cloud server, and each image acquisition device in company B can be used as multiple side servers.
  • the data is still stored locally in company B and does not need to be uploaded to company A, so that the person re-identification network can be obtained through effective training in company B while protecting the data privacy of company A.
  • Embodiments of the present disclosure also provide a pedestrian re-identification method.
  • the pedestrian re-identification method can be executed by a terminal device or other processing device, wherein the terminal device can be an image acquisition device (eg, a smart camera), user equipment (User Equipment, UE), mobile device, user terminal, terminal, cellular phone , cordless phones, Personal Digital Assistant (PDA), handheld devices, computing devices, in-vehicle devices, wearable devices, etc.
  • the terminal device can be an image acquisition device (eg, a smart camera), user equipment (User Equipment, UE), mobile device, user terminal, terminal, cellular phone , cordless phones, Personal Digital Assistant (PDA), handheld devices, computing devices, in-vehicle devices, wearable devices, etc.
  • Other processing devices may be servers or cloud servers, or the like.
  • the pedestrian re-identification method may be implemented by the processor calling computer-readable instructions stored in the memory.
  • the method may include: performing pedestrian re-identification processing on at least one frame of images to be identified obtained within the target geographical area through a target pedestrian re-identification network, and determining a pedestrian re-identification result; wherein, the target pedestrian re-identification network adopts the foregoing embodiment.
  • the network training method is trained.
  • the target pedestrian re-identification network may perform pedestrian re-identification processing on at least one frame of the to-be-identified image within the target geographic area, and determine whether there is a characteristic pedestrian in the at least one frame of the to-be-identified image.
  • the target person re-identification network is the updated first person re-identification network or the training The first pedestrian re-identification network after training.
  • the updated first person re-identification network or the trained first person re-identification network in the cloud server is universal, that is, it can be applied to any application scenario, therefore, the updated first person re-identification network in the cloud server can be used.
  • the pedestrian re-identification network or the trained first pedestrian re-identification network realizes the pedestrian re-identification processing of at least one frame of the image to be recognized obtained within the target geographic area, so as to obtain the pedestrian re-identification result.
  • the target pedestrian re-identification network is the trained second pedestrian re-identification network.
  • Pedestrian Re-Identification Network in the case that an edge server is included within the target geographic area, and the edge server includes a trained second pedestrian re-identification network, the target pedestrian re-identification network is the trained second pedestrian re-identification network.
  • the network training method for the cloud server and the edge server it can be seen that since the local image data sets in different edge servers are collected in different scenarios (lighting, angle), the data between different edge servers is It has heterogeneity.
  • the trained second person re-identification network obtained by different side servers according to the local image data set is personalized and more suitable for local scenes, which leads to the trained second person re-identification in each side server.
  • the performance of the network is better than the updated first person re-identification network trained by the cloud server combined with multiple side servers.
  • the trained second pedestrian that is more suitable for the local scene of the target geographical area can be used.
  • the re-identification network performs pedestrian re-identification processing on at least one frame of the image to be identified, so as to improve the accuracy of the processing result.
  • further training iterations can be performed according to the data generated by the edge server, and a low-cost model can be continuously updated and updated.
  • FIG. 8 shows a block diagram of a network training apparatus according to an embodiment of the present disclosure.
  • the network training device is applied to a cloud server, and the cloud server includes a first pedestrian re-identification network. As shown in FIG.
  • the apparatus 80 includes: a sending part 81, configured to send a first network parameter corresponding to the first pedestrian re-identification network to a plurality of edge servers; a receiving part 82, configured to receive a plurality of edge servers The second network parameter returned by the server, wherein, for any edge server, the edge server includes a second pedestrian re-identification network, an identity classification network and a local image dataset, the second pedestrian re-identification network and the first pedestrian re-identification network
  • the networks have the same network structure, and the second network parameters are obtained by the edge server after training the second pedestrian re-identification network according to the local image data set, the identity classification network and the first network parameters; the updating part 83 is configured to The second network parameters returned by the multiple edge servers update the first pedestrian re-identification network to obtain the updated first pedestrian re-identification network.
  • the updating part 83 includes: a receiving sub-part, configured to receive weights corresponding to the second network parameters returned by multiple edge servers, wherein, for any edge server, the second network parameter The corresponding weight is determined by the edge server according to the second pedestrian re-identification network before training and the second pedestrian re-identification network after training; The weights corresponding to the two network parameters are weighted and averaged on the second network parameters returned by the multiple edge servers to obtain the updated first network parameters; the second update sub-section is configured to, according to the updated first network parameters, The first pedestrian re-identification network is updated to obtain the updated first pedestrian re-identification network.
  • the sending part 81 is further configured to send the shared image data set to multiple edge servers; the receiving part 82 is further configured to receive pseudo tags returned by multiple edge servers, wherein, For any side server, the pseudo-label is generated by the side server according to the shared image data set and the trained second person re-identification network; the apparatus 80 further includes: a network training part configured to The pseudo-label returned by the edge server is used to train the updated first pedestrian re-identification network, and the trained first pedestrian re-identification network is obtained.
  • the network training part is further configured to: determine an average pseudo-label according to pseudo-labels returned by multiple edge servers;
  • the person re-identification network is trained, and the first person re-identification network after training is obtained.
  • FIG. 9 shows a block diagram of a network training apparatus according to an embodiment of the present disclosure.
  • the network training device is applied to a side server, and the side server includes a second person re-identification network, an identity classification network and a local image data set. As shown in FIG.
  • the apparatus 90 includes: a receiving part 91, configured to receive the first network parameter corresponding to the first pedestrian re-identification network sent by the cloud server, wherein the first pedestrian-re-identification network and the second pedestrian-re-identification network
  • the recognition network has the same network structure
  • the network training part 92 is configured to train the second person re-identification network according to the local image data set, the identity classification network and the first network parameters to obtain the trained second person re-identification network network, wherein the second pedestrian re-identification network corresponds to the second network parameter
  • the sending part 93 is configured to send the second network parameter to the cloud server.
  • the network training part 92 is further configured to: train the second person re-identification network and the identity classification network according to the local image data set and the first network parameters, to obtain a trained second person re-identification network and an identity classification network.
  • Person re-identification network and trained identity classification network the apparatus 90 further includes: a storage part configured to store the trained identity classification network in the edge server.
  • the local image dataset includes image data corresponding to multiple identities; the dimension of the identity classification network is related to the number of multiple identities.
  • the receiving part 91 is further configured to receive the shared image data set sent by the cloud server; the apparatus 90 further includes: a pseudo-label generating part, configured to The two-person re-identification network generates a pseudo-label; the sending part 93 is further configured to send the pseudo-label to the cloud server.
  • the apparatus 90 further includes: a first determining part configured to determine a first feature vector according to the second person re-identification network before training and the local image data set, and The pedestrian re-identification network and the local image data set determine a second feature vector; a second determination part is configured to determine a cosine distance between the first feature vector and the second feature vector; a third determination part is configured to determine the cosine distance according to the cosine The distance is used to determine the weight corresponding to the second network parameter; the sending part 93 is further configured to send the weight corresponding to the second network parameter to the cloud server.
  • the side server is an image acquisition device; the local image data set is acquired according to the image acquisition device.
  • the edge server is connected to at least one image acquisition device, and the edge server and the at least one image acquisition device are located in the same geographical area; the local image data set is obtained by the edge server from the at least one image acquisition device owned.
  • a pedestrian re-identification device including: a pedestrian re-identification part configured to perform pedestrian re-identification processing on at least one frame of an image to be identified obtained within a target geographic area through a target pedestrian re-identification network , and determine the pedestrian re-identification result; wherein, the target pedestrian re-identification network is trained by the above-mentioned network training method.
  • the target person re-identification network is an updated first person re-identification network or a trained first person re-identification network.
  • the target pedestrian re-identification network is the trained second pedestrian re-identification network.
  • Pedestrian Re-Identification Network in some embodiments, the functions or included parts of the apparatus for network training/person re-identification provided by the embodiments of the present disclosure may be configured to execute the methods described in the above method embodiments, and the specific implementation may refer to the above method embodiments The description of , for brevity, is not repeated here.
  • a "part" may be a part of a circuit, a part of a processor, a part of a program or software, etc., of course, a unit, a module, or a non-modular form.
  • Embodiments of the present disclosure further provide a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the foregoing method is implemented.
  • the computer-readable storage medium may be a non-volatile computer-readable storage medium.
  • An embodiment of the present disclosure further provides an electronic device, including: a processor; a memory configured to store instructions executable by the processor; wherein the processor is configured to invoke the instructions stored in the memory to execute the above method.
  • Embodiments of the present disclosure also provide a computer program product, including computer-readable codes, and when the computer-readable codes are run on a device, a processor in the device executes the network training for implementing the network training provided in any of the above embodiments /Directive for pedestrian re-identification methods.
  • Embodiments of the present disclosure further provide another computer program product configured to store computer-readable instructions, which, when executed, cause the computer to perform the operations of the network training/person re-identification method provided by any of the foregoing embodiments.
  • the electronic device may be provided as a terminal, server or other form of device.
  • FIG. 10 shows a block diagram of an electronic device according to an embodiment of the present disclosure. As shown in FIG.
  • the electronic device 800 may be an image capture device (eg, a smart camera), a mobile phone, a computer, a digital broadcasting terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant Wait for the terminal. 10 , the electronic device 800 may include one or more of the following components: a processing component 802, a memory 804, a power supply component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, a sensor component 814 , and the communication component 816.
  • the processing component 802 generally controls the overall operation of the electronic device 800, such as operations associated with display, phone calls, data communications, camera operations, and recording operations.
  • the processing component 802 may include one or more processors 820 to execute instructions to perform all or part of the steps of the methods described above. Additionally, processing component 802 may include one or more modules to facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802. Memory 804 is configured to store various types of data to support operation at electronic device 800 . Examples of such data include instructions for any application or method operating on electronic device 800, contact data, phonebook data, messages, pictures, videos, and the like.
  • Memory 804 may be implemented by any type of volatile or non-volatile storage device or combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Magnetic Disk or Optical Disk.
  • Power supply assembly 806 provides power to various components of electronic device 800 .
  • Power supply components 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 800 .
  • Multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user.
  • the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user.
  • the touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may not only sense the boundaries of a touch or swipe action, but also detect the duration and pressure associated with the touch or swipe action.
  • the multimedia component 808 includes a front-facing camera and/or a rear-facing camera. When the electronic device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data.
  • Audio component 810 is configured to output and/or input audio signals.
  • the audio component 810 includes a microphone (MIC) that is configured to receive external audio signals when the electronic device 800 is in an operating mode, such as a calling mode, a recording mode, and a voice recognition mode.
  • the received audio signal may be further stored in memory 804 or transmitted via communication component 816 .
  • audio component 810 also includes a speaker for outputting audio signals.
  • the I/O interface 812 provides an interface between the processing component 802 and a peripheral interface module, and the above-mentioned peripheral interface module may be a keyboard, a click wheel, a button, and the like. These buttons may include, but are not limited to: home button, volume buttons, start button, and lock button.
  • Sensor assembly 814 includes one or more sensors for providing status assessments of various aspects of electronic device 800 .
  • the sensor component 814 can detect the open/closed state of the electronic device 800, the relative positioning of the components, for example, the components are the display and the keypad of the electronic device 800, and the sensor component 814 can also detect the electronic device 800 or the electronic device 800— Changes in the positions of components, presence or absence of user contact with the electronic device 800 , orientation or acceleration/deceleration of the electronic device 800 , and changes in the temperature of the electronic device 800 .
  • Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact.
  • Sensor assembly 814 may also include a light sensor, such as a complementary metal oxide semiconductor (CMOS) or charge coupled device (CCD) image sensor, for use in imaging applications.
  • CMOS complementary metal oxide semiconductor
  • CCD charge coupled device
  • the sensor component 814 may further include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor or a temperature sensor.
  • Communication component 816 is configured to facilitate wired or wireless communication between electronic device 800 and other devices.
  • the electronic device 800 may access a wireless network based on a communication standard, such as a wireless network (WiFi), a second generation mobile communication technology (2G) or a third generation mobile communication technology (3G), or a combination thereof.
  • the communication component 816 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel.
  • the communication component 816 also includes a near field communication (NFC) module to facilitate short-range communication.
  • NFC near field communication
  • the NFC module may be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
  • RFID radio frequency identification
  • IrDA infrared data association
  • UWB ultra-wideband
  • BT Bluetooth
  • electronic device 800 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable A programmed gate array (FPGA), controller, microcontroller, microprocessor or other electronic component implementation is used to perform the above method.
  • ASICs application specific integrated circuits
  • DSPs digital signal processors
  • DSPDs digital signal processing devices
  • PLDs programmable logic devices
  • FPGA field programmable A programmed gate array
  • controller microcontroller, microprocessor or other electronic component implementation is used to perform the above method.
  • FIG. 11 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
  • the electronic device 1900 may be provided as a server.
  • the electronic device 1900 includes a processing component 1922, which further includes one or more processors, and memory resources represented by memory 1932 for storing instructions executable by the processing component 1922, such as application programs.
  • An application program stored in memory 1932 may include one or more modules, each corresponding to a set of instructions.
  • the processing component 1922 is configured to execute instructions to perform the above-described methods.
  • the electronic device 1900 may also include a power supply assembly 1926 configured to perform power management of the electronic device 1900, a wired or wireless network interface 1950 configured to connect the electronic device 1900 to a network, and an input output (I/O) interface 1958 .
  • Electronic device 1900 may operate based on an operating system stored in memory 1932, such as Microsoft's server operating system (Windows ServerTM), Apple's graphical user interface-based operating system (Mac OS XTM), A multi-user, multi-process computer operating system (UnixTM), a free and open source Unix-like operating system (LinuxTM), an open source Unix-like operating system (FreeBSDTM) or similar.
  • a non-volatile computer-readable storage medium such as a memory 1932 comprising computer program instructions executable by the processing component 1922 of the electronic device 1900 to accomplish the above method.
  • Embodiments of the present disclosure may be systems, methods and/or computer program products.
  • the computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of the embodiments of the present disclosure.
  • a computer-readable storage medium may be a tangible device that can hold and store instructions for use by the instruction execution device.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the above.
  • Non-exhaustive list of computer readable storage media include: portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM) or flash memory), static random access memory (SRAM), portable compact disk read only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanically coded device, such as a printer on which instructions are stored Hole cards or raised structures in grooves, and any suitable combination of the above.
  • RAM random access memory
  • ROM read only memory
  • EPROM erasable programmable read only memory
  • flash memory static random access memory
  • SRAM static random access memory
  • CD-ROM compact disk read only memory
  • DVD digital versatile disk
  • memory stick floppy disk
  • mechanically coded device such as a printer on which instructions are stored Hole cards or raised structures in grooves, and any suitable combination of the above.
  • Computer-readable storage media are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (eg, light pulses through fiber optic cables), or through electrical wires transmitted electrical signals.
  • the computer readable program instructions described herein can be downloaded to various computing/processing devices from a computer readable storage medium, or to an external computer or external storage device over a network such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers.
  • the network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network, and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing/processing device .
  • Computer program instructions for performing the operations of embodiments of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting data, or programmed in one or more Source or object code written in any combination of languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as the "C" language or similar programming languages.
  • the computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server implement.
  • the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider to connect).
  • LAN local area network
  • WAN wide area network
  • the electronic circuit may be utilizing state information of computer readable program instructions to personalize an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit may The computer readable program instructions are executed to implement various aspects of the embodiments of the present disclosure. Aspects of embodiments of the present disclosure are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions.
  • These computer readable program instructions can be provided to the processor of a general purpose computer, special purpose computer or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device , resulting in means for implementing the functions/acts specified in one or more blocks of the flowchart and/or block diagrams.
  • These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions make a computer, a programmable data processing apparatus and/or other equipment work in a specific manner, so that the computer-readable medium storing the instructions includes An article of manufacture that includes implementing one or more of the blocks specified in the flowchart and/or block diagram Instructions for various aspects of function/action.
  • Computer-readable program instructions can also be loaded into a computer, other programmable data processing apparatus, or other equipment, so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other equipment to produce a computer-implemented process. , thereby causing instructions executed on a computer, other programmable data processing apparatus, or other device to implement the functions/acts specified in one or more blocks of the flowcharts and/or block diagrams.
  • the flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure.
  • each block in the flowchart or block diagram may represent a module, program segment or part of an instruction, the module, program segment or part of the instruction including one or more functions for implementing the specified logic function. executable instructions.
  • the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations can be implemented by dedicated hardware-based systems that perform the specified functions or actions , or can be implemented using a combination of dedicated hardware and computer instructions.
  • the computer program product can be specifically implemented by hardware, software or a combination thereof.
  • the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as Software Development Kit (SDK), etc. Wait.
  • SDK Software Development Kit
  • Various embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and not limited to the disclosed embodiments. Numerous modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles of the various embodiments, practical application or improvement of technology in the market, or to enable others of ordinary skill in the art to understand the various embodiments disclosed herein.
  • the embodiments of the present disclosure relate to a network training and pedestrian re-identification method and device, a storage medium, and a computer program.
  • the cloud server includes a first pedestrian-re-identification network, and the method includes: sending a message to a plurality of edge servers. First network parameters corresponding to the first pedestrian re-identification network; receiving second network parameters returned by multiple edge servers, wherein, for any edge server, the edge server includes a second pedestrian re-identification network, an identity classification network and the local image data set, the second pedestrian re-identification network and the first pedestrian re-identification network have the same network structure, and the second network parameter is the edge server according to the local image data set, the identity classification network and the first network parameters.
  • the first person-re-identification network is updated to obtain the updated first-person-re-identification network. Since the cloud server combines multiple side servers to train the pedestrian re-identification network, the image data set is still stored in the side server during the training process, and does not need to be uploaded to the cloud server, so that the pedestrian re-identification network can be effectively trained while protecting the data. privacy.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Bioethics (AREA)
  • Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Databases & Information Systems (AREA)
  • Computer Hardware Design (AREA)
  • Computer Security & Cryptography (AREA)
  • Software Systems (AREA)
  • Human Computer Interaction (AREA)
  • Multimedia (AREA)
  • Image Analysis (AREA)

Abstract

The embodiments of the present disclosure relate to a network training method and apparatus, a person re-identification method and apparatus, a storage medium, and a computer program. A cloud server comprises a first person re-identification network. The method comprises: sending, to a plurality of edge servers, a first network parameter corresponding to a first person re-identification network; receiving second network parameters returned by the plurality of edge servers, wherein each second network parameter is obtained by each edge server by means of training a second person re-identification network included in each edge server according to a local image data set, an identity classification network and the first network parameter; and updating the first person re-identification network according to the second network parameters returned by the plurality of edge servers, so as to obtain an updated first person re-identification network.

Description

网 络训练、 行人重识 别方法及 装置、 存储介质 、 计算机程序 相关申请的交叉引用 本公开基于 申请号为 202010864291.4、 申请日为 2020年 08月 25 日的中国专利申请 提出， 并要求该中国专利申请的优先权， 该中国专利申请的全部内容在此引入本申请作为 参考。 技术领域 本公开涉及计算机技术领域， 尤其涉及一种网络训练、 行人重识别方法及装置、 存储 介质、 计算机程序。 背景技术 行人重识别 (Person Re-identification),也称为行人再识别，是利用计算机视觉技术判断 图像或者视频序列中是否存在特定行人的技术。 目前， 行人重识别技术已广泛应用于多个 领域和行业， 如应用于智能视频检测、 智能安保等。 由于行人重识别技术在处理图像或视 频帧序列的过程中， 涉及了人脸、 人体、 个人身份等隐私数据， 因此， 亟需一种可以避免 隐私数据泄露的行人重识别方法。 发明内容 本公开 实施例提出了一种网络训练、 行人重识别方法及装置、 存储介质、 计算机程序 的技术方案。 根据本公开 实施例的一方面，提供了一种网络训练方法，所述方法应用于云端服务器， 所述云端服务器中包括第一行人重识别网络， 所述方法包括： 向多个边端服务器发送所述 第一行人重识别网络对应的第一网络参数； 接收所述多个边端服务器返回的第二网络参数, 其中， 针对任一所述边端服务器， 所述边端服务器中包括第二行人重识别网络、 身份分类 网络和本地图像数据集， 所述第二行人重识别网络和所述第一行人重识别网络具有相同的 网络结构， 所述第二网络参数是所述边端服务器根据所述本地图像数据集、 所述身份分类 网络和所述第一网络参数对所述第二行人重识别网络进行训练之后得到的； 根据所述多个 边端服务器返回的所述第二网络参数， 对所述第一行人重识别网络进行更新， 得到更新后 的所述第一行人重识别网络。 在一种可 能的实现方式中， 所述根据所述多个边端服务器返回的所述第二网络参数， 对所述第一行人重识别网络进行更新， 得到更新后的所述第一行人重识别网络， 包括： 接 收所述多个边端服务器返回的所述第二网络参数对应的权重， 其中， 针对任一所述边端服 务器， 所述第二网络参数对应的权重是所述边端服务器根据训练前的所述第二行人重识别 网络和训练后的所述第二行人重识别网络确定得到的； 根据所述多个边端服务器返回的所 述第二网络参数对应的权重， 对所述多个边端服务器返回的所述第二网络参数进行加权平 均， 得到更新后的所述第一网络参数； 根据更新后的所述第一网络参数， 对所述第一行人 重识别网络进行更新， 得到更新后的所述第一行人重识别网络。 在一种可能的 实现方式中， 所述方法还包括： 向所述多个边端服务器发送共享图像数 据集； 接收所述多个边端服务器返回的伪标签， 其中， 针对任一所述边端服务器， 所述伪 标签是所述边端服务器根据所述共 享图像数据集以及训练后的所述第二行人重识别网络 生成的； 根据所述共享图像数据集和所述多个边端服务器返回的伪标签， 对更新后的所述 第一行人重识别网络进行训练， 得到训练后的所述第一行人重识别网络。 在一种可能的 实现方式中， 所述根据所述共享图像数据集和所述多个边端服务器返回 的伪标签， 对更新后的所述第一行人重识别网络进行训练， 得到训练后的所述第一行人重 识别网络， 包括： 根据所述多个边端服务器返回的伪标签， 确定平均伪标签； 根据所述共 享图像数据集和所述平均伪标签， 对更新后的所述第一行人重识别网络进行训练， 得到训 练后的所述第一行人重识别网络。 根据本公开 实施例的一方面，提供了一种网络训练方法，所述方法应用于边端服务器， 所述边端服务器中包括第二行人重识别网络、 身份分类网络和本地图像数据集， 所述方法 包括： 接收云端服务器发送的第一行人重识别网络对应的第一网络参数， 其中， 所述第一 行人重识别网络和所述第二行人重识别网络具有相同的网络结构； 根据所述本地图像数据 集、 所述身份分类网络和所述第一网络参数， 对所述第二行人重识别网络进行训练， 得到 训练后的所述第二行人重识别网络， 其中， 所述第二行人重识别网络对应第二网络参数； 向所述云端服务器发送所述第二网络参数。 在一种可 能的实现方式中， 所述根据所述本地图像数据集、 所述身份分类网络和所述 第一网络参数， 对所述第二行人重识别网络进行训练， 得到训练后的所述第二行人重识别 网络， 包括： 根据所述本地图像数据集和所述第一网络参数， 对所述第二行人重识别网络 和所述身份分类网络进行训练， 得到训练后的所述第二行人重识别网络和训练后的所述身 份分类网络。 在一种可 能的实现方式中， 所述方法还包括： 将训练后的所述身份分类网络存储在所 述边端服务器中。 在一种可 能的实现方式中， 所述本地图像数据集中包括多个身份对应的图像数据； 所 述身份分类网络的维度与所述多个身份的个数相关。 在一种可 能的实现方式中， 所述方法还包括： 接收所述云端服务器发送的共享图像数 据集； 根据所述共享图像数据集和训练后的所述第二行人重识别网络， 生成伪标签； 向所 述云端服务器发送所述伪标签。 在一种可 能的实现方式中， 所述方法还包括： 才艮据训练前的所述第二行人重识别网络 和所述本地图像数据集确定第一特征向量， 以及根据训练后的所述第二行人重识别网络和 所述本地图像数据集， 确定第二特征向量； 确定所述第一特征向量和所述第二特征向量之 间的余弦距离； 根据所述余弦距离， 确定所述第二网络参数对应的权重； 向所述云端服务 器发送所述第二网络参数对应的权重。 在一种可 能的实现方式中， 所述边端服务器为图像采集设备； 所述本地图像数据集是 根据所述图像采集设备采集得到的。 在一种可 能的实现方式中， 所述边端服务器与至少一个图像采集设备连接， 所述边端 服务器和所述至少一个图像采集设备位于相同地理区域范围； 所述本地图像数据集是所述 边端服务器从所述至少一个图像采集设备中获取得到的。 根据本公开 实施例的一方面， 提供了一种行人重识别方法， 包括： 通过目标行人重识 别网络对在目标地理区域范围内获取到的至少一帧待识别图像进行行人重识别处理， 确定 行人重识别结果；其中，所述目标行人重识别网络采用如上所述的网络训练方法训练得到。 在一种可 能的实现方式中， 所述目标行人重识别网络为更新后的第一行人重识别网络 或训练后的第一行人重识别网络。 在一种可 能的实现方式中， 在所述目标地理区域范围内包括边端服务器， 且所述边端 月艮务器中包括训练后的第二行人重识别网络的情况下， 所述目标行人重识别网络为训练后 的第二行人重识别网络。 根据本公开 实施例的一方面， 提供了一种网络训练装置， 所述网络训练装置应用于云 端服务器， 所述云端服务器中包括第一行人重识别网络， 所述装置包括： 发送部分， 被配 置为向多个边端服务器发送所述第一行人重识别网络对应的第一网络参数； 接收部分， 被 配置为接收所述多个边端服务器返回的第二 网络参数， 其中， 针对任一所述边端服务器， 所述边端服务器中包括第二行人重识别网络、 身份分类网络和本地图像数据集， 所述第二 行人重识别网络和所述第一行人重识别网络具有相同的网络结构， 所述第二网络参数是所 述边端服务器根据所述本地图像数据集、 所述身份分类网络和所述第一网络参数对所述第 二行人重识别网络进行训练之后得到的； 更新部分， 被配置为根据所述多个边端服务器返 回的所述第二网络参数 , 对所述第一行人重识别网络进行更新， 得到更新后的所述第一行 人重识别网络。 根据本公开 实施例的一方面，提供了一种网络训练装置，所述装置应用于边端服务器， 所述边端服务器中包括第二行人重识别网络、 身份分类网络和本地图像数据集， 所述装置 包括： 接收部分， 被配置为接收云端服务器发送的第一行人重识别网络对应的第一网络参 数, 其中， 所述第一行人重识别网络和所述第二行人重识别网络具有相同的网络结构； 网 络训练部分，被配置为根据所述本地图像数据集、所述身份分类网络和所述第一网络参数, 对所述第二行人重识别网络进行训练， 得到训练后的所述第二行人重识别网络， 其中， 所 述第二行人重识别网络对应第二网络参数； 发送部分， 被配置为向所述云端服务器发送所 述第二网络参数 O 根据本公 开实施例的一方面， 提供了一种行人重识别装置， 包括： 行人重识别部分， 被配置为通过 目标行人重识别网络对在目标地理区域范围内获取到的至少一帧待识别图 像进行行人重识别处理， 确定行人重识别结果； 其中， 所述目标行人重识别网络采用如上 所述的网络训练方法训练得到。 根据本公开 实施例的一方面， 提供了一种电子设备， 包括： 处理器； 被配置为存储处 理器可执行指令的存储器； 其中， 所述处理器被配置为调用所述存储器存储的指令， 以执 行上述方法。 根据本公开 实施例的一方面， 提供了一种计算机可读存储介质， 其上存储有计算机程 序指令， 所述计算机程序指令被处理器执行时实现上述方法。 根据本公开 实施例的一方面， 提供了一种计算机程序， 包括计算机可读代码， 当所述 计算机可读代码在电子设备中运行时， 所述电子设备中的处理器执行时实现上述方法。 在本公开 实施例中， 在包括第一行人重识别网络的云端服务器中， 通过向多个边端服 务器发送第一行人重识别网络对应的第一网络参数， 以及接收多个边端服务器返回的第二 网络参数， 其中， 针对任一边端服务器， 边端服务器中包括和第一行人重识别网络具有相 同的网络结构的第二行人重识别网络、 身份分类网络和本地图像数据集， 第二网络参数是 边端服务器根据本地图像数据集、 身份分类网络和第一网络参数对第二行人重识别网络进 行训练之后得到的， 进而根据多个边端服务器返回的第二网络参数， 对第一行人重识别网 络进行更新， 得到更新后的第一行人重识别网络。 云端服务器联合多个边端服务器对行人 重识别网络进行训练， 训练过程中图像数据集仍然保存在边端服务器中， 无需上传至云端 服务器， 从而可以在有效训练行人重识别网络的同时保护了数据隐私性。 应 当理解的是， 以上的一般描述和后文的细节描述仅是示例性和解释性的， 而非限制 本公开实施例。 根据下面参考附图对示例性实施例的详细说明， 本公开实施例的其它特征 及方面将变得清楚。 附图说明 此处的附 图被并入说明书中并构成本说明书的一部分， 这些附图示出了符合本公开的 实施例， 并与说明书一起用于说明本公开实施例的技术方案。 图 1示出根据本公开实施例的一种网络训练方法的流程图； 图 2示出根据本公开实施例提供的示例性的一种网络训练的结构图； 图 3示出根据本公开实施例的一种网络训练方法的流程图； 图 4示出根据本公开实施例提供的示例性的一种确定第二网络参数的权重的示意图； 图 5 示出根据本公开实施例提供的示例性的一种云端服务器-边端服务器的网络结构 图; 图 6示出根据本公开实施例提供的示例性的一种云端服务器 -边端服务器 -终端设备的 网络结构图； 图 7示出根据本公开实施例提供的示例性的一种网络训练的结构图； 图 8示出根据本公开实施例的一种网络训练装置的框图； 图 9示出根据本公开实施例的一种网络训练装置的框图； 图 10示出根据本公开实施例的一种电子设备的框图； 图 11示出根据本公开实施例的一种电子设备的框图。 具体实施方式 以下将参考附图详细说明本公开的各种示例性实施例、 特征和方面。 附图中相同的附 图标记表示功能相同或相似的元件。 尽管在附图中示出了实施例的各种方面， 但是除非特 别指出， 不必按比例绘制附图。 在这里 专用的词 “示例性 ”意为 “用作例子、 实施例或说明性”。 这里作为“示例性 ”所说 明的任何实施例不必解释为优于或好于其它实施例。 本文 中术语 "和 /或”，仅仅是一种描述关联对象的关联关系，表示可以存在三种关系， 例如， A和 /或 B, 可以表示： 单独存在 A, 同时存在 A和 B , 单独存在 B这三种情况。 另外， 本文中术语 “至少一种” 表示多种中的任意一种或多种中的至少两种的任意组合， 例如， 包括 A、 B、 C中的至少一种， 可以表示包括从 A、 B和 C构成的集合中选择的任意一个或多 个元素。 另外， 为了更好地说明本公开实施例， 在下文的具体实施方式中给出了众多的具体 细节。 本领域技术人员应当理解， 没有某些具体细节， 本公开实施例同样可以实施。 在 一些实例中， 对于本领域技术人员熟知的方法、 手段、 元件和电路未作详细描述， 以便于 凸显本公开实施 例的主旨。 图 1示出根据本公开实施例的一种网络训练方法的流程图。 该网络训练方法可以由云 端服务器执行， 云端服务器中包括第一行人重识别网络。 在一些可能的实现方式中， 该网 络训练方法可以通过云端服务器调用存储器中存储的计算机可读指令的方式来实现。 如图 1所示， 该方法可以包括： 在步骤 S 11中， 向多个边端服务器发送第一行人重识别网络对应的第一网络参数。 在 步骤 S12中， 接收多个边端服务器返回的第二网络参数， 其中， 针对任一边端服务 器， 边端服务器中包括第二行人重识别网络、 身份分类网络和本地图像数据集， 第二行人 重识别网络和第一行人重识别网络具有相同的网络结构， 第二网络参数是边端服务器根据 本地图像数据集、 身份分类网络和第一网络参数对第二行人重识别网络进行训练之后得到 的。 在 步骤 S13中， 根据多个边端服务器返回的第二网络参数， 对第一行人重识别网络进 行更新， 得到更新后的第一行人重识别网络。 云端服务器联合多个边端服务器对行人重识别网络进行训练， 训练过程中图像数据集 仍然保存在边端服务器中， 无需上传至云端服务器， 从而可以在有效训练行人重识别网络 的同时保护了数据隐私性。 此外， 由于无需将图像数据集上传至云端服务器， 可以有效节 约通信带宽。 云端服务器联合多个边端服务器对行人重识别网络进行训练时， 云端服务器可以基于 联邦学习算法联合多个边端服务器进行网络训练。 例如， 多个社区联合训练一个行人重识 别网络， 每个社区都设置一个边端服务器， 通过联邦学习算法， 图像数据集（设置在社区 内部或附近的图像采集设备采集得到的图像数据集）仍然存储在社区内（本地边端服务器）, 无需上传至其它社区 （其它边端服务器）， 从而保护了数据隐私性。 实际应用中， 由于不同边端服务器中的本地图像数据集的数据量不相同， 使得不同边 端服务器之间的数据具有异构性。 传统的联邦学习算法在利用多个边端服务器进行网络训 练时， 根据不同边端服务器中的数据量来设置边端服务器中网络训练得到的第二网络参数 的权重。 但是， 由于数据量的多少并不能直接反映网络训练的训练效果， 因此， 云端服务 器利用边端服务器 中基于这种权重确定方法得到的第二网络参数的权重对第一行人重识 别网络进行更新 , 会导致更新后的第一行人重识别网络的精度较低。 在一种可 能的实现方式中， 根据多个边端服务器返回的第二网络参数， 对第一行人重 识别网络进行更新， 得到更新后的第一行人重识别网络， 包括： 接收多个边端服务器返回 的第二网络参数对应的权重， 其中， 针对任一边端服务器， 第二网络参数对应的权重是边 端服务器根据训练前的第二行人重识别网络和训练后的第二行人重识别网络确定得到的； 根据多个边端服务器返回的第二网络参数对应的权重， 对多个边端服务器返回的第二网络 参数进行加权平均， 得到更新后的第一网络参数； 根据更新后的第一网络参数， 对第一行 人重识别网络进行更新， 得到更新后的第一行人重识别网络。 由于边端服务器发送的第二网络参数的权重是边端服务器根据训练前的 第二行人重 识别网络和训练后的第二行人重识别网络确定得到的， 也就是说， 第二网络参数的权重是 根据边端服务器的训练效果确定的， 使得云端服务器根据各边端服务器返回的第二网络参 数对应的权重， 对多个边端服务器返回的第二网络参数进行加权平均后得到精度较高的更 新后的第一网络参数， 进而根据更新后的第一网络参数， 对第一行人重识别网络进行更新 之后， 有效提高了更新后的第一行人重识别网络的精度。 实际应用中， 由于不同边端服务器中的本地图像数据集是在不同场景 （光照、 角度） 下采集得到的， 使得不同边端服务器之间的数据具有异构性， 进而导致各边端服务器根据 本地图像数据集、 身份分类网络和第一网络参数训练得到的训练后的第二行人重识别网络 的性能， 优于云端服务器联合多个边端服务器训练得到的更新后的第一行人重识别网络。 因此， 可以基于知识蒸馆算法， 将各边端服务器中训练后的第二行人重识别网络作为教师 网络， 将云端服务器中更新后的第一行人重识别网络作为学生网络， 利用教师网络对学生 网络进行训练 （利用更新后的第二行人重识别网络对更新后的第一行人重识别网络进行训 练）， 以提高第一行人重识别网络训练过程的稳定性和收敛性。 在 一种可能的实现方式中， 该方法还包括： 向多个边端服务器发送共享图像数据集； 接收多个边端服务器返回的伪标签， 其中， 针对任一边端服务器， 伪标签是边端服务器根 据共享图像数据集以及训练后的第二行人重识别网络生成的； 根据共享图像数据集和多个 边端服务器返回的伪标签， 对更新后的第一行人重识别网络进行训练， 得到训练后的第一 行人重识别网络。 云端服务 器接收各边端服务器返回的伪标签， 由于该伪标签是边端服务器根据共享图 像数据集以及训练后的第二行人重识别网络生成的， 该伪标签可以用于表示训练后的第二 行人重识别网络的网络特性， 因此，根据共享图像数据集和多个边端服务器返回的伪标签， 对更新后的第一行人重识别网络进行训练， 相当于综合各边端服务器的网络特性对更新后 的第一行人重识别网络进行训练， 从而可以有效提高第一行人重识别网络训练过程的稳定 性和收敛性。 其中， 共享图像数据集指的是云端服务器和各边端服务器均可用于进行网络 训练的图像数据集。 图 2示出根据本公开实施例提供的示例性的一种网络训练的结构图。 如图 2所示， 多 个边端服务器中训练后的第二行人重识别网络构成教师网络 1、 教师网络 2. 教师网 络 N， 其中， N为多个边端服务器的个数， N〉1。 云端服务器中更新后的第一行人重识 别网络构成学生网络。 教师网络 1利用共享图像数据集生成伪标签 以及将伪标签 «发送 至云端服务器； 将教师网络 2利用共享图像数据集生成伪标签匕， 以及将伪标签匕发送至云 端服务器； ； 教师网络 利用共享图像数据集生成伪标签 , 以及将伪标签垢发送至
Figure imgf000008_0001
行人重识别网络进行训练， 得到对应第二网络参数的训练后的第二行人重识别网络， 进而 向云端服务器发送第二网络参数。 在一种可 能的实现方式中， 边端服务器与至少一个图像采集设备连接， 边端服务器和 至少一个图像采集设备位于相同地理区域范围； 本地图像数据集是边端服务器从至少一个 图像采集设备中获取得到的。 在相 同地理区域范围内设置有至少一个图像采集设备的情况下， 可以在该地理区域范 围内设置一个边端服务器， 此时， 无需该至少一个图像采集设备具备存储能力和算力。 边 端服务器与各图像采集设备连接， 进而从各图像采集设备获取图像以构建本地图像数据集。 边端服务器接收云端服务器发送的第一行人重识别网络对应的第一网络参数， 根据本地图 像数据集和第一网络参数， 对第二行人重识别网络进行训练， 得到对应第二网络参数的训 练后的第二行人重识别网络， 进而向云端服务器发送第二网络参数。 在一种可 能的实现方式中， 根据本地图像数据集、 身份分类网络和第一网络参数， 对 第二行人重识别网络进行训练， 得到训练后的第二行人重识别网络， 包括： 根据本地图像 数据集和第一网络参数， 对第二行人重识别网络和身份分类网络进行训练， 得到训练后的 第二行人重识别网络和训练后的身份分类网络。 在一种可 能的实现方式中， 本地图像数据集中包括多个身份对应的图像数据； 身份分 类网络的维度与多个身份的个数相关。 由于训练后的行人重识别网络是对图像进行身份识别的网络， 因此， 在对行人重识别 网络进行训练的过程中， 需要用到包括多个身份对应的图像数据的本地图像数据集， 以及 身份分类网络， 身份分类网络的维度与本地图像数据集中包括的多个身份的个数相关。 例 如， 本地图像数据集中包括 100个身份对应的图像数据， 则身份分类网络的维度为 100。 也 就是说， 身份分类网络中包括 100个不同身份类别。 边端服务器将本地的第二行人重识别网络和身份分类网络构建为组合网络， 并利用从 云端服务器接收到的第一网络参数和本地图像数据集对组合网络进行训练， 进而得到训练 后的组合网络， 其中， 训练后的组合网络中包括训练后的第二行人重识别网络和训练后的 身份分类网络， 训练后的第二行人重识别网络对应第二网络参数。 进而边端服务器将第二 网络参数发送至云端服务器。 由于第一行人重识别网络和第二行人重识别网络具有相同的 网络结构， 因此， 可以利用第二网络参数对第一行人重识别网络进行更新。 在一种可 能的实现方式中， 该方法还包括： 将训练后的身份分类网络存储在边端服务 器中。 由于云端服务器中训练得到的第一行人重识别网络在实际进行行人重识别处理过程 中， 无需用到分类器网络， 因此， 为了节约通信带宽， 以及确保基于联邦学习算法进行联 合训练过程中云端服务器和边端服务器中网络结构的一致性， 边端服务器仅将训练后的第 二行人重识别网络对应的第二网络参数发送至云端服务器， 而将训练后的身份分类网络存 储在边端服务器本地。 在 一种可能的实现方式中， 该方法还包括： 接收云端服务器发送的共享图像数据集； 根据共享图像数据集和训练后的第二行人重识别网络， 生成伪标签； 向云端服务器发送伪 标签。 仍 以上述图 2为例， 如图 2所示， 边端服务器接收云端服务器发送的共享图像数据集， 以及利用共享图像数据集和本地训练后的第二行人重识别网络生成伪标签， 进而边端服务 器向云端服务器发送伪标签， 由于该伪标签可以用于表示训练后的第二行人重识别网络的 网络特性， 以使得云端服务器根据该伪标签对云端服务器中更新后的第一行人重识别网络 进行网络训练后得到的训练后的第一行人重识别网络， 使得训练后的第一行人重识别网络 的网络性能与边端服务器中训练后的第二行人重识别网络更接近， 从而有效提高第一行人 重识别网络训练过程的稳定性和收敛性。
Figure imgf000010_0001
地理区域范围 （例如， 同一社区， 或同一公司）， 边端服务器 A分别从图像采集设备 1和图 像采集设备 2获取图像以构建本地图像数据集。 边端服务器 B与终端设备 3、 终端设备 4和终 端设备 5连接， 终端设备 3、 终端设备 4和终端设备 5为图像采集设备（图像采集设备 3、 图 像采集设备 4和图像采集设备 5 , 例如， 图像采集设备为摄像头）， 边端服务器 B、 图像采集 设备 3、 图像采集设备 4和图像采集设备 5设置在相同地理区域范围 （例如， 同一社区， 或 同一公司 ）， 边端服务器 B分别从图像采集设备 3、 图像采集设备 4和图像采集设备 5获取图 像以构建本地图像数据集。云端服务器联合 2个边端服务器（边端服务器 A和边端服务器 B） 对行人重识别网络进行训练， 训练过程中图像数据集仍然保存在各边端服务器本地， 无需 上传至云端服务器， 从而可以在有效训练行人重识别网络的同时保护了数据隐私性。 根据上述论述可知 ， 本公开提出了两种联邦学习和行人重识别结合的训练架构： 云边 架构和端边云架构。 云边架构 ： 云端服务器直接和智能摄像头进行通信， 云端协调多个智能摄像头同时进 行训练。 智能摄像头将图片缓存在边端， 并定时删除清理以减少边端服务器的存储压力。 且这种架构要求智能摄像头有一定的算力、 存储和通信能力。 云边端架构 ： 边缘网关 （即上述的边端服务器）连接多个智能摄像头， 云端服务器连 接多个边缘网关， 行人重识别训练图片从智能摄像头传入边缘网关， 并缓存在边缘网关， 边缘网关与云端服务器进行联邦学习的训练。 在这个过程中， 数据仍然保留在边缘网关， 数据隐私仍然能得到保护。 其中， 典型的应用场景， 如多个社区联合训练一个行人重识别 模型， 每个社区都有一台边缘网关连接多个智能摄像头， 通过联邦学习的方式， 数据仍然 保留在社区内， 不传输到其他社区或者云端服务器， 以此保护了数据隐私。 在一种可 能的实现方式中， 云端服务器联合多个边端服务器对行人重识别网络进行训 练时， 多个边端服务器还可以部分是直接与云端服务器进行通信的图像采集设备（例如， 智能摄像头）， 部分是与至少一个图像采集设备连接的边端服务器， 本公开实施例对此不 做具体限定。 图 7示出根据本公开实施例提供的示例性的一种网络训练的结构图。 如图 7所示， 云端 服务器可以和多个边端服务器 （边端服务器 1、 边端服务器 2, >.. ... , 边端服务器 N）进行 通信， 且云端服务器中包括的第一行人重识别网络和边端服务器中包括的第二行人重识别 网络具有相同的网络结构。 各边端服务器中还包括本地图像数据集以及身份分类网络。 云 端服务器向多个边端服务器发送第一行人重识别网络对应的第一网络参数， 各边端服务器 接收到第一网络参数后， 利用本地图像数据集和身份分类网络对第二行人重识别网络进行 训练， 得到对应第二网络参数的训练后的第二行人重识别网络， 以及训练后的身份分类网 络。 为了确保云端服务器和各边端服务器联合训练的网络结构一致， 各边端服务器仅将训 练后的第二行人重识别网络对应的第二网络参数发送至云端服务器。 云端服务器根据接收 到的多个边端服务器返回的第二网络参数， 对第一行人重识别网络进行更新， 得到更新后 的第一行人重识别网络。 进而将更新后的第一行人重识别网络对应的第一网络参数发送至 多个边端服务器进行循环训练， 直至云端服务器中更新后的第一行人重识别网络的识别精 度达到阈值， 或者循环训练的次数达到预设次数， 结束训练。 通用 的联邦学习算法（ Federated Averaging, FedAvg）要求进行同步的多方的模型（行 人重识别深度学习模型， 即上述的行人重识别网络） 必须是完全一样的。 行人重识别深度 学习模型的分类器层 （即上述的身份分类网络）取决于每一方的数据包含多少个不同的行 人， 所以参与训练的多方模型的分类器层可能会不同， 导致参与联邦学习的多方的模型可 能会有区别， 因此联邦学习算法 FedAvg在上述的应用场景场景下不适用。 根据上述内容可 知， 由于本公开中改进了联邦学习算法， 即允许参与联邦学习的多方的模型有部分不同， 所以使联邦学习能更好的应用在行人重识别的训练上。 由于不同边端服务器中的本地图像数据集的数据量不相同， 使得不同边端服务器之间 的数据具有异构性。 在联合多个边端服务器对第一行人重识别网络进行训练时， 为了降低 数据异构性对更新后的第一行人重识别网络的精度的影响， 在各边端服务器中可以采用基 于训练效果的权重确定方法 来确定训练后的第二行人重识别网络对应的第二网络参数的 权重， 从而使得云端服务器联合各边端服务器返回的第二网络参数对第一行人重识别网络 进行更新后， 得到精度较高的更新后的第一行人重识别网络。 基于训练效果的权重确定方 法的具体步骤如上述实施例相关部分所述， 在此不再赘述。 由于不同边端服务器中的本地图像数据集是在不同场景（光照、角度）下采集得到的， 使得不同边端服务器之间的数据具有异构性， 进而导致各边端服务器根据本地图像数据集 和第一网络参数训练得到的训练后的第二行人重识别网络的性能， 优于云端服务器联合多 个边端服务器训练得到的更新后的第一行人重识别网络。 为了提高第一行人重识别网络训 练过程的稳定性和收敛性， 可以采用知识蒸馆算法， 基于各边端服务器中更新后的第二行 人重识别网络、共享图像数据集，对云端服务器中更新后的第一行人重识别网络进行训练， 从而有效提高了第一行人重识别网络训练过程的稳定性和收敛性。 基于知识蒸馆算法的具 体训练过程如上述实施例相关部分所述， 在此不再赘述。 基 于上述内容可知， 本公开提出使用知识蒸馆的方法， 将参与联邦学习的多方的本地 模型当成教师模型， 云端服务器的模型作为学生模型， 用知识蒸馆的方法更好的将教师模 型的知识传递到学生模型， 以此提高了模型训练的稳定性和收敛性。 在基 于图 7所示的网络结构对行人重识别网络进行训练的过程中， 基于训练效果的权 重确定方法和基于知识蒸馆算法的网络训练可以分别单独使用， 也可以综合使用， 本公开 实施例对此不做具体限定。 在 一种应用场景中， 例如， 在多个公司或者机构要联合进行行人重识别网络的训练， 以提高训练后的行人重识别网络的鲁棒性的情况下， 为了避免将多方数据汇总到同一个服 务器上产生的数据隐私泄露的问题， 可以基于图 7所示的网络结构来对行人重识别网络进 行联合训练， 其中， 多个公司或者机构作为边端服务器， 多个公司或者机构与同一个云端 服务器进行直接通信， 训练过程中数据仍然保存在本地， 无需上传至云端服务器， 从而可 以在云端服务器中通过有效训练得到行人重识别网络的同时， 保护了多个公司或者机构的 数据隐私性。 在一种应用场景中， 例如， 公司 A为公司 B提供行人重识别网络的训练服务， 如果将公 司 B的各图像采集设备（例如， 智能摄像头）的图像数据都上传至公司 A, 将会产生数据隐 私泄漏问题。此时，公司 A可以基于图 7所示的网络结构来对行人重识别网络进行联合训练， 公司 A可以作为云端服务器，公司 B中的各图像采集设备可以作为多个边端服务器，训练过 程中数据仍然保存在公司 B本地， 无需上传至公司 A, 从而可以在公司 B中通过有效训练得 到行人重识别网络的同时保护了公司 A的数据隐私性。 本公开 实施例还提供一种行人重识别方法。 该行人重识别方法可以由终端设备或其它 处理设备执行，其中，终端设备可以为图像采集设备（例如，智能摄像头）、用户设备（User Equipment, UE）、移动设备、用户终端、终端、蜂窝电话、无绳电话、个人数字处理（ Personal Digital Assistant, PDA）、 手持设备、 计算设备、 车载设备、 可穿戴设备等。 其它处理设备 可为服务器或云端服务器等。 在一些可能的实现方式中， 该行人重识别方法可以通过处理 器调用存储器中存储的计算机可读指令的方式来实现。 该方法可以包括： 通 过目标行人重识别网络对在目标地理区域范围内获取到的至少一帧待识别 图像进 行行人重识别处理， 确定行人重识别结果； 其中， 目标行人重识别网络采用前述实施例的 网络训练方法训练得到。 目标行人重识别网络可以对目标地理区域范围内的至少一帧待识别图像进行行人重 识别处理， 确定该至少一帧待识别图像中是否存在特性行人。 在一种可 能的实现方式中， 目标行人重识别网络为更新后的第一行人重识别网络或训 练后的第一行人重识别网络。 由于云端服务器中更新后的第一行人重识别网络或训练后的第一行人 重识别网络具 有普适性， 即可以应用于任意应用场景， 因此， 可以利用云端服务器中更新后的第一行人 重识别网络或训练后的第一行人重识别网络， 实现对在目标地理区域范围内获取到的至少 一帧待识别图像的行人重识别处理， 以得到行人重识别结果。 在一种可能的 实现方式中， 在目标地理区域范围内包括边端服务器， 且边端服务器中 包括训练后的第二行人重识别网络的情况下， 目标行人重识别网络为训练后的第二行人重 识别网络。 结合上述云端服务器和边端服务器的网络训练方法实施例可知， 由于不同边端服务器 中的本地图像数据集是在不同场景 （光照、 角度） 下采集得到的， 使得不同边端服务器之 间的数据具有异构性， 不同边端服务器根据本地图像数据集训练得到的训练后的第二行人 重识别网络具有个性化， 更适应本地场景， 进而导致各边端服务器中训练后的第二行人重 识别网络的性能， 优于云端服务器联合多个边端服务器训练得到的更新后的第一行人重识 别网络。 因此， 在目标地理区域范围内包括边端服务器， 且边端服务器中包括训练后的第 二行人重识别网络的情况下， 可以利用更适应目标地理区域范围的本地场景的训练后的第 二行人重识别网络， 对至少一帧待识别图像进行行人重识别处理， 以提高处理结果的准确 性。 本公开 中，在进行一次部署之后，能根据边端服务器产生的数据进行进一步训练迭代, 可以达到低成本的模型持续更新迭代。 可 以理解， 本公开实施例提及的上述各个方法实施例， 在不违背原理逻辑的情况下， 均可以彼此相互结合形成结合后的实施例， 限于篇幅， 本公开实施例不再赘述。 本领域技 术人员可以理解， 在具体实施方式的上述方法中， 各步骤的具体执行顺序应当以其功能和 可能的内在逻辑确定。 此外 ， 本公开实施例还提供了网络训练 /行人重识别装置、 电子设备、 计算机可读存储 介质、 程序， 上述均可用来实现本公开实施例提供的任一种网络训练 /行人重识别方法， 相 应技术方案和描述和参见方法部分的相应记载， 不再赘述。 图 8示出根据本公开实施例的一种网络训练装置的框图。 该网络训练装置应用于云端 服务器， 云端服务器中包括第一行人重识别网络。 如图 8所示， 装置 80包括： 发送部分 81 , 被配置为向多个边端服务器发送第一行人重识别网络对应的第一网络参 数； 接收部分 82, 被配置为接收多个边端服务器返回的第二网络参数， 其中， 针对任一边 端服务器， 边端服务器中包括第二行人重识别网络、 身份分类网络和本地图像数据集， 第 二行人重识别网络和第一行人重识别网络具有相同的网络结构， 第二网络参数是边端服务 器根据本地图像数据集、 身份分类网络和第一网络参数对第二行人重识别网络进行训练之 后得到的； 更新部分 83 , 被配置为根据多个边端服务器返回的第二网络参数， 对第一行人重识别 网络进行更新， 得到更新后的第一行人重识别网络。 在一种可能的实现方式中， 更新部分 83 , 包括： 接 收子部分， 被配置为接收多个边端服务器返回的第二网络参数对应的权重， 其中， 针对任一边端服务器， 第二网络参数对应的权重是边端服务器根据训练前的第二行人重识 别网络和训练后的第二彳亍 重识别网络确定得到的； 第一更新子部分 ， 被配置为根据多个边端服务器返回的第二网络参数对应的权重， 对 多个边端服务器返回的第二网络参数进行加权平均， 得到更新后的第一网络参数； 第二更新子部分， 被配置为根据更新后的第一网络参数， 对第一行人重识别网络进行 更新， 得到更新后的第一行人重识别网络。 在一种可能的实现方式中， 发送部分 81 , 还被配置为向多个边端服务器发送共享图像 数据 集； 接 收部分 82, 还被配置为接收多个边端服务器返回的伪标签， 其中， 针对任一边端服 务器， 伪标签是边端服务器根据共享图像数据集以及训练后的第二行人重识别网络生成的; 装置 80, 还包括： 网络训练部分， 被配置为根据共享图像数据集和多个边端服务器返回的伪标签， 对更 新后 的第一行人重识别网络进行训练， 得到训练后的第一行人重识别网络。 在一种可能的实现方式中， 网络训练部分， 还被配置为： 根据 多个边端服务器返回的伪标签， 确定平均伪标签； 根据共享图像数据集和平均伪标签， 对更新后的第一行人重识别网络进行训练， 得到 训练后的第一行人重识别网络。 图 9示出根据本公开实施例的一种网络训练装置的框图。 该网络训练装置应用于边端 服 务器， 边端服务器中包括第二行人重识别网络、 身份分类网络和本地图像数据集。 如图 9所示 ， 装置 90包括： 接 收部分 91 , 被配置为接收云端服务器发送的第一行人重识别网络对应的第一网络参 数， 其中， 第一行人重识别网络和第二行人重识别网络具有相同的网络结构； 网络训练部分 92, 被配置为根据本地图像数据集、 身份分类网络和第一网络参数， 对 第二行人重识别网络进行训练， 得到训练后的第二行人重识别网络， 其中， 第二行人重识 别网络对应第二网络参数； 发送部分 93 , 被配置为向云端服务器发送第二网络参数。 在一种可 能的实现方式中， 网络训练部分 92, 还被配置为： 根据本地 图像数据集和第一网络参数， 对第二行人重识别网络和身份分类网络进行训 练， 得到训练后的第二行人重识别网络和训练后的身份分类网络。 在一种可 能的方式中， 装置 90, 还包括： 存储部分 ， 被配置为将训练后的身份分类网络存储在边端服务器中。 在一种可 能的方式中， 本地图像数据集中包括多个身份对应的图像数据； 身份分类网 络的维度与多个身份的个数相关。 在一种可 能的方式中， 接 收部分 91 , 还被配置为接收云端服务器发送的共享图像数据集； 装置 90, 还包括： 伪标签生成部分， 被配置为根据共享图像数据集和训练后的第二行人重识别网络， 生 成伪标签； 发送部分 93 , 还被配置为向云端服务器发送所述伪标签。 在一种可 能的方式中， 装置 90, 还包括： 第一确定部分， 被配置为根据训练前的第二行人重识别网络和本地图像数据集确定第 一特征向量， 以及根据训练后的第二行人重识别网络和本地图像数据集， 确定第二特征向 量； 第二确定部分， 被配置为确定第一特征向量和第二特征向量之间的余弦距离； 第三确定部分， 被配置为根据余弦距离， 确定第二网络参数对应的权重； 发送部分 93 , 还被配置为向云端服务器发送第二网络参数对应的权重。 在一种可 能的方式中， 边端服务器为图像采集设备； 本地图像数据集是根据图像采集 设备采集得到的。 在一种可能的方式中， 边端服务器与至少一个图像采集设备连接， 边端服务器和至少 一个图像采集设备位于相同地理区域范围；本地图像数据集是边端服务器从至少一个图像 采集设备中获取得到的。 本公开 实施例还提供一种行人重识别装置， 包括： 行人重识别部分， 被配置为通过目 标行人重识别网络对在 目标地理区域范围内获取到的至少一帧待识别图像进行行人重识 别处理， 确定行人重识别结果； 其中， 目标行人重识别网络采用上述网络训练方法训练得 到。 在一种可 能的实现方式中， 目标行人重识别网络为更新后的第一行人重识别网络或训 练后的第一行人重识别网络。 在一种可 能的实现方式中， 在目标地理区域范围内包括边端服务器， 且边端服务器中 包括训练后的第二行人重识别网络的情况下， 目标行人重识别网络为训练后的第二行人重 识别网络。 在一些 实施例中，本公开实施例提供的网络训练 /行人重识别装置具有的功能或包含的 部分可以被配置为执行上文方法实施例描述的方法， 其具体实现可以参照上文方法实施例 的描述， 为了简洁， 这里不再赘述。 在本公 开实施例以及其他的实施例中， “部分” 可以是部分电路、 部分处理器、 部分 程序或软件等等， 当然也可以是单元， 还可以是模块也可以是非模块化的。 本公开实施例还提出一种计算机可读存储介质， 其上存储有计算机程序指令， 所述计 算机程序指令被处理器执行时实现上述方法。计算机可读存储介质可以是非易失性计算机 可读存储介质。 本公开 实施例还提出一种电子设备， 包括： 处理器； 被配置为存储处理器可执行指令 的存储器； 其中， 所述处理器被配置为调用所述存储器存储的指令， 以执行上述方法。 本公开 实施例还提供了一种计算机程序产品， 包括计算机可读代码， 在计算机可读代 码在设备上运行的情况下， 设备中的处理器执行用于实现如上任一实施例提供的网络训练 /行人重识别方法的指令。 本公开 实施例还提供了另一种计算机程序产品， 被配置为存储计算机可读指令， 指令 被执行时使得计算机执行上述任一实施例提供的网络训练 /行人重识别方法的操作。 电子设备可以被提供为终端、 服务器或其它形态的设备。 图 10示出根据本公开实施例的一种电子设备的框图。 如图 10所示， 电子设备 800可以 是图像采集设备 (例如， 智能摄像头)、 移动电话， 计算机， 数字广播终端， 消息收发设 备， 游戏控制台， 平板设备， 医疗设备， 健身设备， 个人数字助理等终端。 参 照图 10, 电子设备 800可以包括以下一个或多个组件： 处理组件 802, 存储器 804, 电源组件 806, 多媒体组件 808, 音频组件 810, 输入 /输出 ( I/ O ) 的接口 812, 传感器组件 814, 以及通信组件 816。 处理组件 802通常控制电子设备 800的整体操作， 诸如与显示， 电话呼叫， 数据通信， 相机操作和记录操作相关联的操作。 处理组件 802可以包括一个或多个处理器 820来执行指 令， 以完成上述的方法的全部或部分步骤。 此外， 处理组件 802可以包括一个或多个模块， 便于处理组件 802和其他组件之间的交互。 例如， 处理组件 802可以包括多媒体模块， 以方 便多媒体组件 808和处理组件 802之间的交互。 存储 器 804被配置为存储各种类型的数据以支持在电子设备 800的操作。 这些数据的示 例包括用于在电子设备 800上操作的任何应用程序或方法的指令， 联系人数据， 电话簿数 据， 消息， 图片， 视频等。 存储器 804可以由任何类型的易失性或非易失性存储设备或者 它们的组合实现，如静态随机存取存储器 ( SRAM ),电可擦除可编程只读存储器 ( EEPROM ), 可擦除可编程只读存储器 ( EPROM ),可编程只读存储器 ( PROM ), 只读存储器 ( ROM ), 磁存储器， 快闪存储器， 磁盘或光盘。 电源组件 806为电子设备 800的各种组件提供电力。 电源组件 806可以包括电源管理系 统， 一个或多个电源， 及其他与为电子设备 800生成、 管理和分配电力相关联的组件。 多媒体组件 808包括在所述电子设备 800和用户之间的提供一个输出接口的屏幕。 在一 些实施例中， 屏幕可以包括液晶显示器（ LCD）和触摸面板（ TP）。 如果屏幕包括触摸面 板， 屏幕可以被实现为触摸屏， 以接收来自用户的输入信号。 触摸面板包括一个或多个触 摸传感器以感测触摸、 滑动和触摸面板上的手势。 所述触摸传感器可以不仅感测触摸或滑 动动作的边界， 而且还检测与所述触摸或滑动操作相关的持续时间和压力。 在一些实施例 中， 多媒体组件 808包括一个前置摄像头和 /或后置摄像头。 当电子设备 800处于操作模式， 如拍摄模式或视频模式时， 前置摄像头和 /或后置摄像头可以接收外部的多媒体数据。 每个 前置摄像头和后置摄像头可以是一个固定的光学透镜系统或具有焦距和光学变焦能力。 音 频组件 810被配置为输出和 /或输入音频信号。 例如， 音频组件 810包括一个麦克风 （ MIC）, 当电子设备 800处于操作模式， 如呼叫模式、 记录模式和语音识别模式时， 麦克 风被配置为接收外部音频信号。 所接收的音频信号可以被进一步存储在存储器 804或经由 通信组件 816发送。在一些实施例中，音频组件 810还包括一个扬声器，用于输出音频信号。 CROSS-REFERENCE TO RELATED APPLICATIONS Priority, the entire content of this Chinese patent application is incorporated herein by reference. FIELD OF THE DISCLOSURE The present disclosure relates to the field of computer technology, and in particular, to a method and device for network training and pedestrian re-identification, a storage medium, and a computer program. 2. Description of the Related Art Pedestrian re-identification (Person Re-identification), also known as pedestrian re-identification, is a technology that uses computer vision technology to determine whether there is a specific pedestrian in an image or a video sequence. At present, pedestrian re-identification technology has been widely used in many fields and industries, such as intelligent video detection, intelligent security and so on. Since the person re-identification technology involves private data such as face, human body, and personal identity in the process of processing images or video frame sequences, there is an urgent need for a person re-identification method that can avoid leakage of private data. SUMMARY Embodiments of the present disclosure propose a network training and pedestrian re-identification method and device, a storage medium, and a technical solution for a computer program. According to an aspect of the embodiments of the present disclosure, a network training method is provided. The method is applied to a cloud server, where the cloud server includes a first person re-identification network, and the method includes: sending first network parameters corresponding to the first pedestrian re-identification network; receiving second network parameters returned by the multiple edge servers, wherein, for any of the edge servers, the edge servers include A second person re-identification network, an identity classification network and a local image dataset, the second person re-identification network and the first person re-identification network have the same network structure, and the second network parameter is the edge obtained after the end server trains the second pedestrian re-identification network according to the local image data set, the identity classification network and the first network parameters; Two network parameters, updating the first pedestrian re-identification network to obtain the updated first pedestrian re-identification network. In a possible implementation manner, the first pedestrian re-identification network is updated according to the second network parameters returned by the multiple edge servers to obtain the updated first row The person re-identification network includes: receiving weights corresponding to the second network parameters returned by the plurality of edge servers, wherein, for any of the edge servers, the weights corresponding to the second network parameters are the Determined by the edge server according to the second pedestrian re-identification network before training and the second pedestrian re-identification network after training; according to the weights corresponding to the second network parameters returned by the multiple edge servers , performing a weighted average on the second network parameters returned by the plurality of edge servers to obtain the updated first network parameters; according to the updated first network parameters, the first pedestrian The re-identification network is updated to obtain the updated first pedestrian re-identification network. In a possible implementation manner, the method further includes: sending a shared image data set to the multiple edge servers; receiving pseudo tags returned by the multiple edge servers, wherein for any of the edge servers an end server, where the pseudo label is generated by the side server according to the shared image dataset and the trained second pedestrian re-identification network; according to the shared image dataset and the multiple side servers For the returned pseudo-label, the updated first pedestrian re-identification network is trained to obtain the trained first pedestrian re-identification network. In a possible implementation manner, the returning according to the shared image data set and the multiple side servers training the updated first pedestrian re-identification network, and obtaining the trained first pedestrian re-identification network, comprising: determining, according to the pseudo-tags returned by the multiple edge servers, Average pseudo-label; train the updated first person re-identification network according to the shared image data set and the average pseudo-label, to obtain the trained first person re-identification network. According to an aspect of the embodiments of the present disclosure, a network training method is provided, the method is applied to an edge server, and the edge server includes a second person re-identification network, an identity classification network and a local image dataset, so The method includes: receiving a first network parameter corresponding to a first pedestrian re-identification network sent by a cloud server, wherein the first pedestrian-re-identification network and the second pedestrian-re-identification network have the same network structure; The local image data set, the identity classification network and the first network parameters are used to train the second pedestrian re-identification network, and the trained second pedestrian re-identification network is obtained, wherein the first pedestrian re-identification network is obtained. The two-person re-identification network corresponds to the second network parameter; and the second network parameter is sent to the cloud server. In a possible implementation manner, the second pedestrian re-identification network is trained according to the local image data set, the identity classification network and the first network parameters to obtain the trained The second person re-identification network includes: training the second person re-identification network and the identity classification network according to the local image data set and the first network parameters, to obtain the trained second person re-identification network. A person re-identification network and the trained identity classification network. In a possible implementation manner, the method further includes: storing the trained identity classification network in the edge server. In a possible implementation manner, the local image data set includes image data corresponding to multiple identities; and the dimension of the identity classification network is related to the number of the multiple identities. In a possible implementation manner, the method further includes: receiving a shared image data set sent by the cloud server; generating a pseudo-label according to the shared image data set and the trained second person re-identification network ; Send the pseudo tag to the cloud server. In a possible implementation manner, the method further includes: determining a first feature vector according to the second pedestrian re-identification network before training and the local image data set, and determining a first feature vector according to the second pedestrian re-identification network after training and the local image data set; The two-person re-identification network and the local image data set determine a second feature vector; determine the cosine distance between the first feature vector and the second feature vector; determine the second feature vector according to the cosine distance the weight corresponding to the network parameter; sending the weight corresponding to the second network parameter to the cloud server. In a possible implementation manner, the edge server is an image acquisition device; and the local image data set is acquired according to the image acquisition device. In a possible implementation manner, the edge server is connected to at least one image acquisition device, and the edge server and the at least one image acquisition device are located in the same geographical area; the local image data set is the The edge server is obtained from the at least one image acquisition device. According to an aspect of the embodiments of the present disclosure, a pedestrian re-identification method is provided, including: performing pedestrian re-identification processing on at least one frame of images to be identified obtained within a target geographic area through a target pedestrian re-identification network, and determining the pedestrian The re-identification result; wherein, the target pedestrian re-identification network is obtained by training the above-mentioned network training method. In a possible implementation manner, the target person re-identification network is an updated first person re-identification network or a trained first person re-identification network. In a possible implementation manner, in the case that an edge server is included within the target geographical area, and the edge server includes a trained second pedestrian re-identification network, the target pedestrian The re-identification network is the second person re-identification network after training. According to an aspect of the embodiments of the present disclosure, a network training device is provided, the network training device is applied to a cloud server, the cloud server includes a first person re-identification network, and the device includes: a sending part, which is is configured to send the first network parameters corresponding to the first pedestrian re-identification network to the plurality of edge servers; the receiving part is configured to receive the second network parameters returned by the multiple edge servers, wherein for any a described edge server, The edge server includes a second person re-identification network, an identity classification network and a local image dataset, the second person re-identification network and the first person re-identification network have the same network structure, and the first person re-identification network has the same network structure. The second network parameter is obtained by the side server after training the second person re-identification network according to the local image data set, the identity classification network and the first network parameter; the update part is configured as According to the second network parameters returned by the multiple edge servers, the first pedestrian re-identification network is updated to obtain the updated first pedestrian re-identification network. According to an aspect of the embodiments of the present disclosure, a network training apparatus is provided, the apparatus is applied to an edge server, and the edge server includes a second pedestrian re-identification network, an identity classification network and a local image dataset, so The apparatus includes: a receiving part configured to receive a first network parameter corresponding to a first pedestrian re-identification network sent by a cloud server, wherein the first pedestrian-re-identification network and the second pedestrian-re-identification network have the same network structure; the network training part is configured to train the second pedestrian re-identification network according to the local image data set, the identity classification network and the first network parameters, and obtain the trained the second pedestrian re-identification network, wherein the second pedestrian re-identification network corresponds to a second network parameter; the sending part is configured to send the second network parameter to the cloud server according to an embodiment of the present disclosure. In an aspect, a pedestrian re-identification device is provided, comprising: a pedestrian re-identification part configured to perform pedestrian re-identification processing on at least one frame of images to be identified obtained within a target geographical area through a target pedestrian re-identification network, and determine Pedestrian re-identification result; wherein, the target pedestrian re-identification network is obtained by training the above-mentioned network training method. According to an aspect of the embodiments of the present disclosure, there is provided an electronic device, comprising: a processor; a memory configured to store instructions executable by the processor; wherein the processor is configured to invoke the instructions stored in the memory, to perform the above method. According to an aspect of the embodiments of the present disclosure, there is provided a computer-readable storage medium having computer program instructions stored thereon, the computer program instructions implementing the foregoing method when executed by a processor. According to an aspect of the embodiments of the present disclosure, a computer program is provided, including computer-readable code, when the computer-readable code is executed in an electronic device, the processor in the electronic device implements the foregoing method when executed. In the embodiment of the present disclosure, in the cloud server including the first pedestrian re-identification network, the first network parameters corresponding to the first pedestrian re-identification network are sent to multiple edge servers, and the multiple edge servers are received. The returned second network parameter, wherein, for any edge server, the edge server includes a second person re-identification network, an identity classification network and a local image dataset that have the same network structure as the first person re-identification network, The second network parameter is obtained by the edge server after training the second pedestrian re-identification network according to the local image data set, the identity classification network and the first network parameters, and then according to the second network parameters returned by the multiple edge servers, to The first person re-identification network is updated, and the updated first person re-identification network is obtained. The cloud server combines multiple side servers to train the pedestrian re-identification network. During the training process, the image data set is still saved in the side server and does not need to be uploaded to the cloud server, so that the pedestrian re-identification network can be effectively trained while protecting the data. privacy. It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not limiting of the embodiments of the present disclosure. Other features and aspects of embodiments of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description, serve to explain the technical solutions of the embodiments of the present disclosure. Fig. 1 shows a flow chart of a network training method according to an embodiment of the present disclosure; Fig. 2 shows a structural diagram of an exemplary network training provided according to an embodiment of the present disclosure; Fig. 3 shows an embodiment according to the present disclosure A flowchart of a network training method according to an embodiment of the present disclosure; FIG. 4 shows an exemplary schematic diagram of determining the weight of a second network parameter provided according to an embodiment of the present disclosure; FIG. 5 shows an exemplary schematic diagram provided according to an embodiment of the present disclosure. A cloud server-side server network structure Fig. 6 shows an exemplary network structure diagram of a cloud server-side server-terminal device provided according to an embodiment of the present disclosure; Fig. 7 shows an exemplary network training provided according to an embodiment of the present disclosure 8 shows a block diagram of a network training apparatus according to an embodiment of the present disclosure; FIG. 9 shows a block diagram of a network training apparatus according to an embodiment of the present disclosure; A block diagram of an electronic device; FIG. 11 shows a block diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numbers in the figures denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise indicated. The word "exemplary" is used exclusively herein to mean "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments. The term "and/or" in this article is only an association relationship to describe associated objects, which means that there can be three kinds of relationships, for example, A and/or B, which can mean: A exists alone, A and B exist simultaneously, and A and B exist independently B these three cases. In addition, the term "at least one" herein refers to any one of multiple or any combination of at least two of multiple, for example, including at least one of A, B, and C, may mean including from A, B, and C. Any one or more elements selected from the set of B and C. In addition, in order to better illustrate the embodiments of the present disclosure, numerous specific details are given in the following detailed description. It should be understood by those skilled in the art that the embodiments of the present disclosure can also be implemented without certain specific details. In some instances, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the gist of the embodiments of the present disclosure. FIG. 1 shows a flowchart of a network training method according to an embodiment of the present disclosure. The network training method can be executed by a cloud server, and the cloud server includes a first pedestrian re-identification network. In some possible implementations, the network training method may be implemented by a cloud server invoking computer-readable instructions stored in a memory. As shown in FIG. 1 , the method may include: In step S11, sending first network parameters corresponding to the first pedestrian re-identification network to a plurality of edge servers. In step S12, the second network parameters returned by multiple edge servers are received, wherein, for any edge server, the edge server includes a second pedestrian re-identification network, an identity classification network and a local image data set, the second pedestrian The re-identification network and the first person re-identification network have the same network structure, and the second network parameters are obtained after the edge server trains the second person re-identification network according to the local image data set, the identity classification network and the first network parameters. of. In step S13, the first pedestrian re-identification network is updated according to the second network parameters returned by the multiple edge servers to obtain an updated first pedestrian re-identification network. The cloud server combines multiple side servers to train the pedestrian re-identification network. During the training process, the image data set is still saved in the side server and does not need to be uploaded to the cloud server, so that the pedestrian re-identification network can be effectively trained while protecting the data. privacy. In addition, since there is no need to upload the image dataset to the cloud server, communication bandwidth can be effectively saved. When the cloud server combines multiple side servers to train the pedestrian re-identification network, the cloud server can combine multiple side servers to perform network training based on the federated learning algorithm. For example, a person re-identification network is jointly trained by multiple communities, and each community is set up with a side server. Through the federated learning algorithm, the image data set (the image data set collected by the image acquisition equipment set in the community or nearby) still remains It is stored in the community (local side server) and does not need to be uploaded to other communities (other side servers), thus protecting data privacy. In practical applications, due to the different data volumes of local image datasets in different edge servers, different edge servers have different data sizes. The data between the end servers is heterogeneous. When the traditional federated learning algorithm uses multiple side servers for network training, the weights of the second network parameters obtained by the network training in the side servers are set according to the amount of data in different side servers. However, since the amount of data cannot directly reflect the training effect of network training, the cloud server updates the first pedestrian re-identification network by using the weights of the second network parameters obtained in the side server based on this weight determination method. , which will lead to lower accuracy of the updated first person re-identification network. In a possible implementation manner, the first pedestrian re-identification network is updated according to the second network parameters returned by multiple edge servers, and the updated first pedestrian re-identification network is obtained, including: receiving multiple The weight corresponding to the second network parameter returned by the edge server, wherein, for any edge server, the weight corresponding to the second network parameter is the weight of the edge server according to the second pedestrian re-identification network before training and the second pedestrian weight after training. It is determined by identifying the network; according to the weights corresponding to the second network parameters returned by the multiple edge servers, weighted average is performed on the second network parameters returned by the multiple edge servers to obtain the updated first network parameters; according to the updated first network parameters; The first network parameters are updated, and the first pedestrian re-identification network is updated to obtain the updated first pedestrian re-identification network. Since the weight of the second network parameter sent by the side server is determined by the side server according to the second pedestrian re-identification network before training and the second pedestrian re-identification network after training, that is, the weight of the second network parameter It is determined according to the training effect of the edge server, so that the cloud server performs a weighted average of the second network parameters returned by the multiple edge servers according to the weights corresponding to the second network parameters returned by each edge server to obtain a higher-precision network parameter. The updated first network parameters, and then after the first pedestrian re-identification network is updated according to the updated first network parameters, the accuracy of the updated first pedestrian re-identification network is effectively improved. In practical applications, since the local image datasets in different edge servers are collected in different scenarios (lighting, angle), the data between different edge servers is heterogeneous, which leads to The performance of the trained second person re-identification network obtained by training the local image data set, the identity classification network and the first network parameters is better than the updated first person re-identification obtained by training the cloud server in conjunction with multiple side servers. The internet. Therefore, based on the knowledge evaporation algorithm, the second person re-identification network trained in each side server can be used as the teacher network, the updated first person re-identification network in the cloud server can be used as the student network, and the teacher network can be used to The student network is trained (using the updated second person re-identification network to train the updated first-person re-identification network) to improve the stability and convergence of the first-person re-identification network training process. In a possible implementation manner, the method further includes: sending a shared image data set to multiple edge servers; receiving pseudo tags returned by the multiple edge servers, wherein, for any edge server, the pseudo tag is the edge server The server is generated according to the shared image data set and the trained second person re-identification network; according to the shared image data set and the pseudo-labels returned by multiple side servers, the updated first person re-identification network is trained to obtain The first person re-identification network after training. The cloud server receives the pseudo-label returned by each side server. Since the pseudo-label is generated by the side-end server according to the shared image data set and the trained second person re-identification network, the pseudo-label can be used to represent the trained second person re-identification network. The network characteristics of the person re-identification network, therefore, according to the shared image data set and the pseudo-labels returned by multiple side servers, the updated first person re-identification network is trained, which is equivalent to synthesizing the network characteristics of each side server. The updated first person re-identification network is trained, so that the stability and convergence of the training process of the first person re-identification network can be effectively improved. The shared image dataset refers to an image dataset that both the cloud server and each side server can use for network training. FIG. 2 shows a structural diagram of an exemplary network training provided according to an embodiment of the present disclosure. As shown in FIG. 2 , the second person re-identification network trained in the multiple edge servers constitutes a teacher network 1, a teacher network 2, and a teacher network N, where N is the number of multiple edge servers, and N>1 . The updated first person re-identification network in the cloud server constitutes the student network. The teacher network 1 uses the shared image data set to generate pseudo-labels and sends the pseudo-labels to the cloud server; the teacher network 2 uses the shared image data set to generate pseudo-labels, and sends the pseudo-labels to the cloud server; Image datasets generate pseudo-labels, and send pseudo-labels to
Figure imgf000008_0001
The pedestrian re-identification network is trained to obtain a trained second pedestrian re-identification network corresponding to the second network parameters, and then the second network parameters are sent to the cloud server. In a possible implementation manner, the edge server is connected to at least one image acquisition device, and the edge server and the at least one image acquisition device are located in the same geographical area; the local image data set is obtained by the edge server from the at least one image acquisition device. obtained. In the case where at least one image acquisition device is provided within the same geographical area, an edge server may be provided within the geographical area, and in this case, the at least one image acquisition device does not need to have storage capability and computing power. The edge server is connected to each image acquisition device, and further acquires images from each image acquisition device to construct a local image data set. The edge server receives the first network parameters corresponding to the first pedestrian re-identification network sent by the cloud server, and trains the second pedestrian re-identification network according to the local image data set and the first network parameters, and obtains a parameter corresponding to the second network parameters. The trained second person re-identification network, and then sends the second network parameters to the cloud server. In a possible implementation manner, the second person re-identification network is trained according to the local image data set, the identity classification network and the first network parameters, and the trained second person re-identification network is obtained, including: according to the local image The data set and the first network parameters are used to train the second person re-identification network and the identity classification network, and the trained second person re-identification network and the trained identity classification network are obtained. In a possible implementation manner, the local image dataset includes image data corresponding to multiple identities; the dimension of the identity classification network is related to the number of multiple identities. Since the trained person re-identification network is a network for identifying images, therefore, in the process of training the person re-identification network, a local image dataset including image data corresponding to multiple identities needs to be used, and the identity Classification network, the dimension of the identity classification network is related to the number of multiple identities included in the local image dataset. For example, if the local image dataset includes image data corresponding to 100 identities, the dimension of the identity classification network is 100. That is, the identity classification network includes 100 different identity classes. The side-end server constructs the local second person re-identification network and the identity classification network as a combined network, and uses the first network parameters received from the cloud server and the local image data set to train the combined network, and then obtains the combined network after training network, wherein the trained combined network includes a trained second person re-identification network and a trained identity classification network, and the trained second person re-identification network corresponds to a second network parameter. Further, the edge server sends the second network parameter to the cloud server. Since the first person re-identification network and the second person re-identification network have the same network structure, the first person re-identification network can be updated by using the second network parameters. In a possible implementation manner, the method further includes: storing the trained identity classification network in the edge server. Since the first person re-identification network trained in the cloud server does not need to use the classifier network in the actual pedestrian re-identification process, in order to save the communication bandwidth and ensure that the cloud is used in the joint training process based on the federated learning algorithm The network structure between the server and the side server is consistent. The side server only sends the second network parameters corresponding to the trained second person re-identification network to the cloud server, and stores the trained identity classification network on the side server. local. In a possible implementation manner, the method further includes: receiving a shared image data set sent by the cloud server; generating a pseudo-label according to the shared image data set and the trained second person re-identification network; sending the pseudo-label to the cloud server . Still taking the above FIG. 2 as an example, as shown in FIG. 2, the side server receives the shared image data set sent by the cloud server, and uses the shared image data set and the locally trained second person re-identification network to generate pseudo labels, and then the edge The terminal server sends a pseudo-label to the cloud server, because the pseudo-label can be used to represent the network characteristics of the trained second pedestrian re-identification network, so that the cloud server can re-identify the updated first pedestrian in the cloud server according to the pseudo-label. The first person re-identification network after training is obtained after the recognition network performs network training, so that the network performance of the first person re-identification network after training is closer to that of the second person re-identification network after training in the side server. Thus, the stability and convergence of the first person re-identification network training process are effectively improved.
Figure imgf000010_0001
In the geographic area (for example, the same community, or the same company), the edge server A obtains images from the image acquisition device 1 and the image acquisition device 2 respectively to construct a local image dataset. The side server B is connected to the terminal device 3, the terminal device 4 and the terminal device 5, and the terminal device 3, the terminal device 4 and the terminal device 5 are image acquisition devices (the image acquisition device 3, the image acquisition device 4 and the image acquisition device 5, for example , the image acquisition device is a camera), the edge server B, the image acquisition device 3, the image acquisition device 4, and the image acquisition device 5 are set in the same geographical area (for example, the same community, or the same company), and the edge server B is Image capture device 3, image capture device 4, and image capture device 5 acquire images to construct a local image dataset. The cloud server combines two side servers (side server A and side server B) to train the pedestrian re-identification network. During the training process, the image data set is still stored locally on each side server without uploading to the cloud server. Data privacy is preserved while effectively training a person re-identification network. According to the above discussion, the present disclosure proposes two training architectures combining federated learning and person re-identification: a cloud-edge architecture and a device-edge-cloud architecture. Cloud-edge architecture: The cloud server communicates directly with the smart cameras, and the cloud coordinates multiple smart cameras to train at the same time. The smart camera caches pictures on the edge, and deletes and cleans them regularly to reduce the storage pressure on the edge server. And this architecture requires smart cameras to have certain computing power, storage and communication capabilities. Cloud-edge-end architecture: The edge gateway (that is, the above-mentioned edge-end server) is connected to multiple smart cameras, the cloud server is connected to multiple edge gateways, and the pedestrian re-identification training image is transmitted from the smart camera to the edge gateway, and cached in the edge gateway, the edge gateway Federated learning training with cloud servers. During this process, the data remains at the edge gateway, and data privacy can still be protected. Among them, typical application scenarios, such as multiple communities jointly training a pedestrian re-identification model, each community has an edge gateway connected to multiple smart cameras, through federated learning, the data is still retained in the community, not transmitted to other Community or cloud server to protect data privacy. In a possible implementation manner, when the cloud server cooperates with multiple side servers to train the pedestrian re-identification network, the multiple side servers may also be partly image acquisition devices (for example, smart cameras) that communicate directly with the cloud server. ), part of which is an edge server connected to at least one image acquisition device, which is not specifically limited in this embodiment of the present disclosure. FIG. 7 shows a structural diagram of an exemplary network training provided according to an embodiment of the present disclosure. As shown in FIG. 7 , the cloud server can communicate with multiple side servers (side server 1, side server 2, >... , side server N), and the first line included in the cloud server The person re-identification network and the second person re-identification network included in the edge server have the same network structure. Each edge server also includes a local image dataset and an identity classification network. The cloud server sends the first network parameters corresponding to the first pedestrian re-identification network to the multiple edge servers, and each edge server uses the local image data set and the identity classification network to re-identify the second pedestrian after receiving the first network parameters. The network is trained to obtain a trained second person re-identification network corresponding to the second network parameters, and a trained identity classification network. In order to ensure that the network structure jointly trained by the cloud server and each side server is consistent, each side server only sends the second network parameters corresponding to the trained second person re-identification network to the cloud server. The cloud server updates the first pedestrian re-identification network according to the received second network parameters returned by the multiple edge servers, and obtains the updated first pedestrian re-identification network. Then, the first network parameters corresponding to the updated first pedestrian re-identification network are sent to multiple side servers for cyclic training, until the updated recognition accuracy of the first pedestrian re-identification network in the cloud server reaches the threshold, or the cycle is repeated. When the number of training reaches the preset number, the training ends. The general federated learning algorithm (Federated Averaging, FedAvg) requires that the multi-party models (person re-identification deep learning model, ie the above-mentioned person re-identification network) to be synchronized must be exactly the same. The classifier layer of the deep learning model for person re-identification (that is, the above-mentioned identity classification network) depends on how many different pedestrians each party's data contains, so the classifier layers of the multi-party models participating in training may be different, resulting in participation in federated learning. Multi-party models may be different, so the federated learning algorithm FedAvg is not applicable in the above application scenarios. As can be seen from the above content, since the federated learning algorithm is improved in the present disclosure, that is, the models of multiple parties participating in the federated learning are allowed to be partially different, so the federated learning can be better applied to the training of person re-identification. Since the data volume of the local image datasets in different edge servers is different, the difference between different edge servers data is heterogeneous. When combining multiple edge servers to train the first person re-identification network, in order to reduce the impact of data heterogeneity on the accuracy of the updated first person re-identification network, each edge server can use a The weight determination method of the training effect is used to determine the weight of the second network parameter corresponding to the trained second pedestrian re-identification network, so that the cloud server combines the second network parameters returned by each side server to carry out the first pedestrian re-identification network. After the update, the updated first person re-identification network with higher accuracy is obtained. The specific steps of the weight determination method based on the training effect are as described in the relevant parts of the foregoing embodiments, and are not repeated here. Since the local image datasets in different side servers are collected under different scenes (lighting, angle), the data between different side servers is heterogeneous, which leads to each side server according to the local image dataset. The performance of the trained second person re-identification network obtained by training with the first network parameters is better than that of the updated first person re-identification network trained by the cloud server in conjunction with multiple side servers. In order to improve the stability and convergence of the training process of the first person re-identification network, the knowledge evaporation algorithm can be used, based on the updated second person re-identification network and shared image data set in each side server, and the cloud server The updated first person re-identification network is trained, thereby effectively improving the stability and convergence of the training process of the first person re-identification network. The specific training process based on the knowledge steaming library algorithm is described in the relevant part of the above-mentioned embodiment, and details are not repeated here. Based on the above content, it can be seen that the present disclosure proposes a method of using a knowledge library, taking the local model of multiple parties participating in federated learning as the teacher model, and the model of the cloud server as the student model, and using the knowledge library method to better integrate the knowledge of the teacher model It is passed to the student model, thereby improving the stability and convergence of the model training. In the process of training the pedestrian re-identification network based on the network structure shown in FIG. 7 , the weight determination method based on the training effect and the network training based on the knowledge evaporation algorithm can be used separately, or can be used in combination. Embodiments of the present disclosure There is no specific limitation on this. In an application scenario, for example, when multiple companies or institutions want to jointly train a person re-identification network to improve the robustness of the trained person re-identification network, in order to avoid aggregating data from multiple parties into the same For the problem of data privacy leakage generated on a server, the pedestrian re-identification network can be jointly trained based on the network structure shown in FIG. The cloud server communicates directly, and the data is still stored locally during the training process, and there is no need to upload it to the cloud server, so that the pedestrian re-identification network can be obtained through effective training in the cloud server, while protecting the data privacy of multiple companies or institutions. In an application scenario, for example, company A provides company B with a pedestrian re-identification network training service. If the image data of each image acquisition device (for example, a smart camera) of company B is uploaded to company A, it will generate Data privacy leaks. At this time, company A can jointly train the pedestrian re-identification network based on the network structure shown in Figure 7, company A can be used as a cloud server, and each image acquisition device in company B can be used as multiple side servers. The data is still stored locally in company B and does not need to be uploaded to company A, so that the person re-identification network can be obtained through effective training in company B while protecting the data privacy of company A. Embodiments of the present disclosure also provide a pedestrian re-identification method. The pedestrian re-identification method can be executed by a terminal device or other processing device, wherein the terminal device can be an image acquisition device (eg, a smart camera), user equipment (User Equipment, UE), mobile device, user terminal, terminal, cellular phone , cordless phones, Personal Digital Assistant (PDA), handheld devices, computing devices, in-vehicle devices, wearable devices, etc. Other processing devices may be servers or cloud servers, or the like. In some possible implementations, the pedestrian re-identification method may be implemented by the processor calling computer-readable instructions stored in the memory. The method may include: performing pedestrian re-identification processing on at least one frame of images to be identified obtained within the target geographical area through a target pedestrian re-identification network, and determining a pedestrian re-identification result; wherein, the target pedestrian re-identification network adopts the foregoing embodiment. The network training method is trained. The target pedestrian re-identification network may perform pedestrian re-identification processing on at least one frame of the to-be-identified image within the target geographic area, and determine whether there is a characteristic pedestrian in the at least one frame of the to-be-identified image. In a possible implementation, the target person re-identification network is the updated first person re-identification network or the training The first pedestrian re-identification network after training. Since the updated first person re-identification network or the trained first person re-identification network in the cloud server is universal, that is, it can be applied to any application scenario, therefore, the updated first person re-identification network in the cloud server can be used. The pedestrian re-identification network or the trained first pedestrian re-identification network realizes the pedestrian re-identification processing of at least one frame of the image to be recognized obtained within the target geographic area, so as to obtain the pedestrian re-identification result. In a possible implementation manner, in the case that an edge server is included within the target geographic area, and the edge server includes a trained second pedestrian re-identification network, the target pedestrian re-identification network is the trained second pedestrian re-identification network. Pedestrian Re-Identification Network. Combining the above embodiments of the network training method for the cloud server and the edge server, it can be seen that since the local image data sets in different edge servers are collected in different scenarios (lighting, angle), the data between different edge servers is It has heterogeneity. The trained second person re-identification network obtained by different side servers according to the local image data set is personalized and more suitable for local scenes, which leads to the trained second person re-identification in each side server. The performance of the network is better than the updated first person re-identification network trained by the cloud server combined with multiple side servers. Therefore, in the case where the side server is included in the target geographical area, and the second pedestrian re-identification network after training is included in the side server, the trained second pedestrian that is more suitable for the local scene of the target geographical area can be used. The re-identification network performs pedestrian re-identification processing on at least one frame of the image to be identified, so as to improve the accuracy of the processing result. In the present disclosure, after one deployment, further training iterations can be performed according to the data generated by the edge server, and a low-cost model can be continuously updated and updated. It can be understood that the above method embodiments mentioned in the embodiments of the present disclosure can be combined with each other to form a combined embodiment without violating the principle and logic. Those skilled in the art can understand that, in the above method in the specific implementation manner, the specific execution order of each step should be determined by its function and possible internal logic. In addition, the embodiments of the present disclosure also provide a network training/pedestrian re-identification apparatus, electronic equipment, computer-readable storage media, and programs, all of which can be used to implement any of the network training/pedestrian re-identification methods provided by the embodiments of the present disclosure, Corresponding technical solutions and descriptions and refer to the corresponding records in the method section, which will not be repeated. FIG. 8 shows a block diagram of a network training apparatus according to an embodiment of the present disclosure. The network training device is applied to a cloud server, and the cloud server includes a first pedestrian re-identification network. As shown in FIG. 8, the apparatus 80 includes: a sending part 81, configured to send a first network parameter corresponding to the first pedestrian re-identification network to a plurality of edge servers; a receiving part 82, configured to receive a plurality of edge servers The second network parameter returned by the server, wherein, for any edge server, the edge server includes a second pedestrian re-identification network, an identity classification network and a local image dataset, the second pedestrian re-identification network and the first pedestrian re-identification network The networks have the same network structure, and the second network parameters are obtained by the edge server after training the second pedestrian re-identification network according to the local image data set, the identity classification network and the first network parameters; the updating part 83 is configured to The second network parameters returned by the multiple edge servers update the first pedestrian re-identification network to obtain the updated first pedestrian re-identification network. In a possible implementation manner, the updating part 83 includes: a receiving sub-part, configured to receive weights corresponding to the second network parameters returned by multiple edge servers, wherein, for any edge server, the second network parameter The corresponding weight is determined by the edge server according to the second pedestrian re-identification network before training and the second pedestrian re-identification network after training; The weights corresponding to the two network parameters are weighted and averaged on the second network parameters returned by the multiple edge servers to obtain the updated first network parameters; the second update sub-section is configured to, according to the updated first network parameters, The first pedestrian re-identification network is updated to obtain the updated first pedestrian re-identification network. In a possible implementation manner, the sending part 81 is further configured to send the shared image data set to multiple edge servers; the receiving part 82 is further configured to receive pseudo tags returned by multiple edge servers, wherein, For any side server, the pseudo-label is generated by the side server according to the shared image data set and the trained second person re-identification network; the apparatus 80 further includes: a network training part configured to The pseudo-label returned by the edge server is used to train the updated first pedestrian re-identification network, and the trained first pedestrian re-identification network is obtained. In a possible implementation manner, the network training part is further configured to: determine an average pseudo-label according to pseudo-labels returned by multiple edge servers; The person re-identification network is trained, and the first person re-identification network after training is obtained. FIG. 9 shows a block diagram of a network training apparatus according to an embodiment of the present disclosure. The network training device is applied to a side server, and the side server includes a second person re-identification network, an identity classification network and a local image data set. As shown in FIG. 9, the apparatus 90 includes: a receiving part 91, configured to receive the first network parameter corresponding to the first pedestrian re-identification network sent by the cloud server, wherein the first pedestrian-re-identification network and the second pedestrian-re-identification network The recognition network has the same network structure; the network training part 92 is configured to train the second person re-identification network according to the local image data set, the identity classification network and the first network parameters to obtain the trained second person re-identification network network, wherein the second pedestrian re-identification network corresponds to the second network parameter; the sending part 93 is configured to send the second network parameter to the cloud server. In a possible implementation manner, the network training part 92 is further configured to: train the second person re-identification network and the identity classification network according to the local image data set and the first network parameters, to obtain a trained second person re-identification network and an identity classification network. Person re-identification network and trained identity classification network. In a possible manner, the apparatus 90 further includes: a storage part configured to store the trained identity classification network in the edge server. In a possible manner, the local image dataset includes image data corresponding to multiple identities; the dimension of the identity classification network is related to the number of multiple identities. In a possible manner, the receiving part 91 is further configured to receive the shared image data set sent by the cloud server; the apparatus 90 further includes: a pseudo-label generating part, configured to The two-person re-identification network generates a pseudo-label; the sending part 93 is further configured to send the pseudo-label to the cloud server. In a possible manner, the apparatus 90 further includes: a first determining part configured to determine a first feature vector according to the second person re-identification network before training and the local image data set, and The pedestrian re-identification network and the local image data set determine a second feature vector; a second determination part is configured to determine a cosine distance between the first feature vector and the second feature vector; a third determination part is configured to determine the cosine distance according to the cosine The distance is used to determine the weight corresponding to the second network parameter; the sending part 93 is further configured to send the weight corresponding to the second network parameter to the cloud server. In a possible manner, the side server is an image acquisition device; the local image data set is acquired according to the image acquisition device. In a possible manner, the edge server is connected to at least one image acquisition device, and the edge server and the at least one image acquisition device are located in the same geographical area; the local image data set is obtained by the edge server from the at least one image acquisition device owned. Embodiments of the present disclosure further provide a pedestrian re-identification device, including: a pedestrian re-identification part configured to perform pedestrian re-identification processing on at least one frame of an image to be identified obtained within a target geographic area through a target pedestrian re-identification network , and determine the pedestrian re-identification result; wherein, the target pedestrian re-identification network is trained by the above-mentioned network training method. In a possible implementation manner, the target person re-identification network is an updated first person re-identification network or a trained first person re-identification network. In a possible implementation manner, in the case that an edge server is included within the target geographic area, and the edge server includes a trained second pedestrian re-identification network, the target pedestrian re-identification network is the trained second pedestrian re-identification network. Pedestrian Re-Identification Network. In some embodiments, the functions or included parts of the apparatus for network training/person re-identification provided by the embodiments of the present disclosure may be configured to execute the methods described in the above method embodiments, and the specific implementation may refer to the above method embodiments The description of , for brevity, is not repeated here. In the embodiments of the present disclosure and other embodiments, a "part" may be a part of a circuit, a part of a processor, a part of a program or software, etc., of course, a unit, a module, or a non-modular form. Embodiments of the present disclosure further provide a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the foregoing method is implemented. The computer-readable storage medium may be a non-volatile computer-readable storage medium. An embodiment of the present disclosure further provides an electronic device, including: a processor; a memory configured to store instructions executable by the processor; wherein the processor is configured to invoke the instructions stored in the memory to execute the above method. Embodiments of the present disclosure also provide a computer program product, including computer-readable codes, and when the computer-readable codes are run on a device, a processor in the device executes the network training for implementing the network training provided in any of the above embodiments /Directive for pedestrian re-identification methods. Embodiments of the present disclosure further provide another computer program product configured to store computer-readable instructions, which, when executed, cause the computer to perform the operations of the network training/person re-identification method provided by any of the foregoing embodiments. The electronic device may be provided as a terminal, server or other form of device. FIG. 10 shows a block diagram of an electronic device according to an embodiment of the present disclosure. As shown in FIG. 10, the electronic device 800 may be an image capture device (eg, a smart camera), a mobile phone, a computer, a digital broadcasting terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant Wait for the terminal. 10 , the electronic device 800 may include one or more of the following components: a processing component 802, a memory 804, a power supply component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, a sensor component 814 , and the communication component 816. The processing component 802 generally controls the overall operation of the electronic device 800, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 802 may include one or more processors 820 to execute instructions to perform all or part of the steps of the methods described above. Additionally, processing component 802 may include one or more modules to facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802. Memory 804 is configured to store various types of data to support operation at electronic device 800 . Examples of such data include instructions for any application or method operating on electronic device 800, contact data, phonebook data, messages, pictures, videos, and the like. Memory 804 may be implemented by any type of volatile or non-volatile storage device or combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Magnetic Disk or Optical Disk. Power supply assembly 806 provides power to various components of electronic device 800 . Power supply components 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 800 . Multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user. In a In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may not only sense the boundaries of a touch or swipe action, but also detect the duration and pressure associated with the touch or swipe action. In some embodiments, the multimedia component 808 includes a front-facing camera and/or a rear-facing camera. When the electronic device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each of the front and rear cameras can be a fixed optical lens system or have focal length and optical zoom capability. Audio component 810 is configured to output and/or input audio signals. For example, the audio component 810 includes a microphone (MIC) that is configured to receive external audio signals when the electronic device 800 is in an operating mode, such as a calling mode, a recording mode, and a voice recognition mode. The received audio signal may be further stored in memory 804 or transmitted via communication component 816 . In some embodiments, audio component 810 also includes a speaker for outputting audio signals.
1/ O接口 812为处理组件 802和外围接口模块之间提供接口， 上述外围接口模块可以是 键盘， 点击轮， 按钮等。 这些按钮可包括但不限于： 主页按钮、 音量按钮、 启动按钮和锁 定按钮。 传感 器组件 814包括一个或多个传感器， 用于为电子设备 800提供各个方面的状态评估。 例如， 传感器组件 814可以检测到电子设备 800的打开 /关闭状态， 组件的相对定位， 例如所 述组件为电子设备 800的显示器和小键盘， 传感器组件 814还可以检测电子设备 800或电子 设备 800 —个组件的位置改变， 用户与电子设备 800接触的存在或不存在， 电子设备 800方 位或加速 /减速和电子设备 800的温度变化。 传感器组件 814可以包括接近传感器， 被配置用 来在没有任何的物理接触时检测附近物体的存在 。 传感器组件 814还可以包括光传感器， 如互补金属氧化物半导体 （CMOS）或电荷耦合装置 （CCD） 图像传感器， 用于在成像应 用中使用。在一些实施例中， 该传感器组件 814还可以包括加速度传感器， 陀螺仪传感器， 磁传感器， 压力传感器或温度传感器。 通信组件 816被配置为便于电子设备 800和其他设备之间有线或无线方式的通信。 电子 设备 800可以接入基于通信标准的无线网络，如无线网络（ WiFi）,第二代移动通信技术（ 2G） 或第三代移动通信技术 （3G）, 或它们的组合。 在一个示例性实施例中， 通信组件 816经由 广播信道接收来自外部广播管理系统的广播信号或广播相关信息。 在一个示例性实施例中, 所述通信组件 816还包括近场通信（NFC）模块， 以促进短程通信。 例如， 在 NFC模块可基 于射频识别（ RFID）技札 红外数据协会（ IrDA）技札超宽带（ UWB）技札蓝牙（ BT） 技术和其他技术来实现。 在 示例性实施例中， 电子设备 800可以被一个或多个应用专用集成电路（ ASIC）、 数字 信号处理器 （DSP）、 数字信号处理设备（DSPD）、 可编程逻辑器件 （PLD）、 现场可编程 门阵列（FPGA）、控制器、微控制器、微处理器或其他电子元件实现， 用于执行上述方法。 在 示例性实施例中， 还提供了一种非易失性计算机可读存储介质， 例如包括计算机程 序指令的存储器 804 , 上述计算机程序指令可由电子设备 800的处理器 820执行以完成上述 方法。 图 11示出根据本公开实施例的一种电子设备的框图。 如图 11所示， 电子设备 1900可以 被提供为一服务器。 参照图 11 , 电子设备 1900包括处理组件 1922, 其进一步包括一个或多 个处理器， 以及由存储器 1932所代表的存储器资源， 用于存储可由处理组件 1922的执行的 指令， 例如应用程序。 存储器 1932中存储的应用程序可以包括一个或一个以上的每一个对 应于一组指令的模块。 此外， 处理组件 1922被配置为执行指令， 以执行上述方法。 电子设备 1900还可以包括一个电源组件 1926被配置为执行电子设备 1900的电源管理， 一个有线或无线网络接口 1950被配置为将电子设备 1900连接到网络，和一个输入输出（ I/O） 接口 1958。 电子设备 1900可以操作基于存储在存储器 1932的操作系统， 例如微软月艮务器操 作系统 （Windows Server™）, 苹果公司推出的基于图形用户界面操作系统（Mac OS X™）, 多用户多进程的计算机操作系统（ Unix™）,自由和开放原代码的类 Unix操作系统（ Linux™）, 开放原代码的类 Unix操作系统（ FreeBSD™）或类似。 在 示例性实施例中， 还提供了一种非易失性计算机可读存储介质， 例如包括计算机程 序指令的存储器 1932, 上述计算机程序指令可由电子设备 1900的处理组件 1922执行以完成 上述方法。 本公开 实施例可以是系统、 方法和 /或计算机程序产品。 计算机程序产品可以包括计算 机可读存储介质， 其上载有用于使处理器实现本公开实施例的各个方面的计算机可读程序 指令。 计 算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。 计算机可读存储介质例如可以是 （但不限于） 电存储设备、 磁存储设备、 光存储设备、 电 磁存储设备、 半导体存储设备或者上述的任意合适的组合。 计算机可读存储介质的更具体 的例子 （非穷举的列表） 包括： 便携式计算机盘、 硬盘、 随机存取存储器 （RAM）、 只读 存储器（ ROM）、可擦式可编程只读存储器（ EPROM或闪存）、静态随机存取存储器（ SRAM）、 便携式压缩盘只读存储器 （CD-ROM）、 数字多功能盘（DVD）、 记忆棒、 软盘、 机械编码 设备、 例如其上存储有指令的打孔卡或凹槽内凸起结构、 以及上述的任意合适的组合。 这 里所使用的计算机可读存储介质不被解释为瞬时信号本身， 诸如无线电波或者其他自由传 播的电磁波、 通过波导或其他传输媒介传播的电磁波（例如， 通过光纤电缆的光脉冲）、 或者通过电线传输的电信号。 这里 所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算 /处理 设备， 或者通过网络、 例如因特网、 局域网、 广域网和 /或无线网下载到外部计算机或外部 存储设备。 网络可以包括铜传输电缆、 光纤传输、 无线传输、 路由器、 防火墙、 交换机、 网关计算机和 /或边缘服务器。 每个计算 /处理设备中的网络适配卡或者网络接口从网络接 收计算机可读程序指令， 并转发该计算机可读程序指令， 以供存储在各个计算 /处理设备中 的计算机可读存储介质中。 用 于执行本公开实施例操作的计算机程序指令可以是汇编指令、 指令集架构 （ISA） 指令、 机器指令、 机器相关指令、 微代码、 固件指令、 状态设置数据、 或者以一种或多种 编程语言的任意组合编写的源代码或目标代码， 所述编程语言包括面向对象的编程语言一 诸如 Smalltalk. C++等， 以及常规的过程式编程语言一诸如 “C”语言或类似的编程语言。 计 算机可读程序指令可以完全地在用户计算机上执行、 部分地在用户计算机上执行、 作为一 个独立的软件包执行、 部分在用户计算机上部分在远程计算机上执行、 或者完全在远程计 算机或服务器上执行。 在涉及远程计算机的情形中， 远程计算机可以通过任意种类的网络 — 包括局域网（LAN）或广域网（WAN）一连接到用户计算机， 或者， 可以连接到外部计算机 （例如利用因特网服务提供商来通过因特网连接）。 在一些实施例中， 通过利用计算机可 读程序指令的状态信息来个性化定制电子电路， 例如可编程逻辑电路、 现场可编程门阵列 （ FPGA）或可编程逻辑阵列 （PLA）, 该电子电路可以执行计算机可读程序指令， 从而实 现本公开实施例的各个方面。 这里参照根据本公开实施例的方法、 装置（系统）和计算机程序产品的流程图和 /或框 图描述了本公开实施例的各个方面。 应当理解， 流程图和 /或框图的每个方框以及流程图和 /或框图中各方框的组合， 都可以由计算机可读程序指令实现。 这些计算机可读程序指令可以提供给通用计算机、 专用计算机或其它可编程数据处理 装置的处理器， 从而生产出一种机器， 使得这些指令在通过计算机或其它可编程数据处理 装置的处理器执行时， 产生了实现流程图和 /或框图中的一个或多个方框中规定的功能 /动 作的装置。 也可以把这些计算机可读程序指令存储在计算机可读存储介质中， 这些指令使 得计算机、 可编程数据处理装置和 /或其他设备以特定方式工作， 从而， 存储有指令的计算 机可读介质则包括一个制造品，其包括实现流程图和 /或框图中的一个或多个方框中规定的 功能 /动作的各个方面的指令。 也可 以把计算机可读程序指令加载到计算机、 其它可编程数据处理装置、 或其它设备 上， 使得在计算机、 其它可编程数据处理装置或其它设备上执行一系列操作步骤， 以产生 计算机实现的过程， 从而使得在计算机、 其它可编程数据处理装置、 或其它设备上执行的 指令实现流程图和 /或框图中的一个或多个方框中规定的功能 /动作。 附图中的流程图和框图显示了根据本公开实施例的多个实施例的系统、 方法和计算机 程序产品的可能实现的体系架构、 功能和操作。 在这点上， 流程图或框图中的每个方框可 以代表一个模块、 程序段或指令的一部分， 所述模块、 程序段或指令的一部分包含一个或 多个用于实现规定的逻辑功能的可执行指令。 在有些作为替换的实现中， 方框中所标注的 功能也可以以不同于附图中所标注的顺序发生。 例如， 两个连续的方框实际上可以基本并 行地执行， 它们有时也可以按相反的顺序执行， 这依所涉及的功能而定。 也要注意的是, 框图和 /或流程图中的每个方框、 以及框图和 /或流程图中的方框的组合， 可以用执行规定 的功能或动作的专用的基于硬件的系统来实现， 或者可以用专用硬件与计算机指令的组合 来实现。 该计算机程序产品可以具体通过硬件、 软件或其结合的方式实现。 在一个可选实施例 中， 所述计算机程序产品具体体现为计算机存储介质， 在另一个可选实施例中， 计算机程 序产品具体体现为软件产品， 例如软件开发 & (Software Development Kit, SDK)等等。 以上已经描述了本公开的各实施例， 上述说明是示例性的， 并非穷尽性的， 并且也不 限于所披露的各实施例。 在不偏离所说明的各实施例的范围和精神的情况下， 对于本技术 领域的普通技术人员来说许多修改和变更都是显而易见的。 本文中所用术语的选择， 旨在 最好地解释各实施例的原理、 实际应用或对市场中的技术的改进， 或者使本技术领域的其 它普通技术人员能理解本文披露的各实施例。 工业实用性 本公 开实施例涉及一种网络训练、 行人重识别方法及装置、 存储介质、 计算机程序， 云端服务器中包括第一行人重识别网络， 所述方法包括： 向多个边端服务器发送第一行人 重识别网络对应的第一网络参数； 接收多个边端服务器返回的第二网络参数， 其中， 针对 任一边端服务器， 边端服务器中包括第二行人重识别网络、 身份分类网络和本地图像数据 集， 第二行人重识别网络和第一行人重识别网络具有相同的网络结构， 第二网络参数是边 端服务器根据本地图像数据集、 身份分类网络和第一网络参数对第二行人重识别网络进行 训练之后得到的； 根据多个边端服务器返回的第二网络参数， 对第一行人重识别网络进行 更新， 得到更新后的第一行人重识别网络。 由于云端服务器联合多个边端服务器对行人重 识别网络进行训练， 训练过程中图像数据集仍然保存在边端服务器中， 无需上传至云端服 务器， 从而可以在有效训练行人重识别网络的同时保护数据隐私性。 The I/O interface 812 provides an interface between the processing component 802 and a peripheral interface module, and the above-mentioned peripheral interface module may be a keyboard, a click wheel, a button, and the like. These buttons may include, but are not limited to: home button, volume buttons, start button, and lock button. Sensor assembly 814 includes one or more sensors for providing status assessments of various aspects of electronic device 800 . For example, the sensor component 814 can detect the open/closed state of the electronic device 800, the relative positioning of the components, for example, the components are the display and the keypad of the electronic device 800, and the sensor component 814 can also detect the electronic device 800 or the electronic device 800— Changes in the positions of components, presence or absence of user contact with the electronic device 800 , orientation or acceleration/deceleration of the electronic device 800 , and changes in the temperature of the electronic device 800 . Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact. Sensor assembly 814 may also include a light sensor, such as a complementary metal oxide semiconductor (CMOS) or charge coupled device (CCD) image sensor, for use in imaging applications. In some embodiments, the sensor component 814 may further include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor or a temperature sensor. Communication component 816 is configured to facilitate wired or wireless communication between electronic device 800 and other devices. The electronic device 800 may access a wireless network based on a communication standard, such as a wireless network (WiFi), a second generation mobile communication technology (2G) or a third generation mobile communication technology (3G), or a combination thereof. In one exemplary embodiment, the communication component 816 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies. In an exemplary embodiment, electronic device 800 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable A programmed gate array (FPGA), controller, microcontroller, microprocessor or other electronic component implementation is used to perform the above method. In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 804 including computer program instructions that can be executed by the processor 820 of the electronic device 800 to complete the above method. FIG. 11 shows a block diagram of an electronic device according to an embodiment of the present disclosure. As shown in FIG. 11, the electronic device 1900 may be provided as a server. 11, the electronic device 1900 includes a processing component 1922, which further includes one or more processors, and memory resources represented by memory 1932 for storing instructions executable by the processing component 1922, such as application programs. An application program stored in memory 1932 may include one or more modules, each corresponding to a set of instructions. Additionally, the processing component 1922 is configured to execute instructions to perform the above-described methods. The electronic device 1900 may also include a power supply assembly 1926 configured to perform power management of the electronic device 1900, a wired or wireless network interface 1950 configured to connect the electronic device 1900 to a network, and an input output (I/O) interface 1958 . Electronic device 1900 may operate based on an operating system stored in memory 1932, such as Microsoft's server operating system (Windows Server™), Apple's graphical user interface-based operating system (Mac OS X™), A multi-user, multi-process computer operating system (Unix™), a free and open source Unix-like operating system (Linux™), an open source Unix-like operating system (FreeBSD™) or similar. In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 comprising computer program instructions executable by the processing component 1922 of the electronic device 1900 to accomplish the above method. Embodiments of the present disclosure may be systems, methods and/or computer program products. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of the embodiments of the present disclosure. A computer-readable storage medium may be a tangible device that can hold and store instructions for use by the instruction execution device. The computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the above. More specific examples (non-exhaustive list) of computer readable storage media include: portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM) or flash memory), static random access memory (SRAM), portable compact disk read only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanically coded device, such as a printer on which instructions are stored Hole cards or raised structures in grooves, and any suitable combination of the above. Computer-readable storage media, as used herein, are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (eg, light pulses through fiber optic cables), or through electrical wires transmitted electrical signals. The computer readable program instructions described herein can be downloaded to various computing/processing devices from a computer readable storage medium, or to an external computer or external storage device over a network such as the Internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. The network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network, and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing/processing device . Computer program instructions for performing the operations of embodiments of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting data, or programmed in one or more Source or object code written in any combination of languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as the "C" language or similar programming languages. The computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server implement. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider to connect). In some embodiments, by utilizing state information of computer readable program instructions to personalize an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit may The computer readable program instructions are executed to implement various aspects of the embodiments of the present disclosure. Aspects of embodiments of the present disclosure are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions. These computer readable program instructions can be provided to the processor of a general purpose computer, special purpose computer or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device , resulting in means for implementing the functions/acts specified in one or more blocks of the flowchart and/or block diagrams. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions make a computer, a programmable data processing apparatus and/or other equipment work in a specific manner, so that the computer-readable medium storing the instructions includes An article of manufacture that includes implementing one or more of the blocks specified in the flowchart and/or block diagram Instructions for various aspects of function/action. Computer-readable program instructions can also be loaded into a computer, other programmable data processing apparatus, or other equipment, so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other equipment to produce a computer-implemented process. , thereby causing instructions executed on a computer, other programmable data processing apparatus, or other device to implement the functions/acts specified in one or more blocks of the flowcharts and/or block diagrams. The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, program segment or part of an instruction, the module, program segment or part of the instruction including one or more functions for implementing the specified logic function. executable instructions. In some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It is also noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented by dedicated hardware-based systems that perform the specified functions or actions , or can be implemented using a combination of dedicated hardware and computer instructions. The computer program product can be specifically implemented by hardware, software or a combination thereof. In an optional embodiment, the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as Software Development Kit (SDK), etc. Wait. Various embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and not limited to the disclosed embodiments. Numerous modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles of the various embodiments, practical application or improvement of technology in the market, or to enable others of ordinary skill in the art to understand the various embodiments disclosed herein. Industrial Applicability The embodiments of the present disclosure relate to a network training and pedestrian re-identification method and device, a storage medium, and a computer program. The cloud server includes a first pedestrian-re-identification network, and the method includes: sending a message to a plurality of edge servers. First network parameters corresponding to the first pedestrian re-identification network; receiving second network parameters returned by multiple edge servers, wherein, for any edge server, the edge server includes a second pedestrian re-identification network, an identity classification network and the local image data set, the second pedestrian re-identification network and the first pedestrian re-identification network have the same network structure, and the second network parameter is the edge server according to the local image data set, the identity classification network and the first network parameters. Obtained after the two-person re-identification network is trained; according to the second network parameters returned by the multiple edge servers, the first person-re-identification network is updated to obtain the updated first-person-re-identification network. Since the cloud server combines multiple side servers to train the pedestrian re-identification network, the image data set is still stored in the side server during the training process, and does not need to be uploaded to the cloud server, so that the pedestrian re-identification network can be effectively trained while protecting the data. privacy.

Claims

权利要 求书 claims
1、 一种网络训练方法， 所述方法应用于云端服务器， 所述云端服务器中包括第一行 人重识别网络， 所述方法包括： 向多个边端服务器发送所述第一行人重识别网络对应的第一网络参数； 接收所述 多个边端服务器返回的第二网络参数， 其中， 针对任一所述边端服务器， 所 述边端服务器中包括第二行人重识别网络、 身份分类网络和本地图像数据集， 所述第二行 人重识别网络和所述第一行人重识别网络具有相同的网络结构， 所述第二网络参数是所述 边端服务器根据所述本地图像数据集、 所述身份分类网络和所述第一网络参数对所述第二 行人重识别网络进行训练之后得到的； 根据所述 多个边端服务器返回的所述第二网络参数， 对所述第一行人重识别网络进行 更新， 得到更新后的所述第一行人重识别网络。 1. A network training method, the method is applied to a cloud server, and the cloud server includes a first pedestrian re-identification network, and the method includes: sending the first pedestrian re-identification to a plurality of edge servers The first network parameter corresponding to the network; receiving the second network parameter returned by the plurality of edge servers, wherein, for any of the edge servers, the edge server includes a second pedestrian re-identification network, identity classification network and a local image data set, the second pedestrian re-identification network and the first pedestrian re-identification network have the same network structure, and the second network parameter is the edge server according to the local image data set , obtained after the identity classification network and the first network parameters are trained on the second pedestrian re-identification network; according to the second network parameters returned by the multiple edge servers, the first The pedestrian re-identification network is updated to obtain the updated first pedestrian re-identification network.
2、 根据权利要求 1所述的方法， 其中， 所述根据所述多个边端服务器返回的所述第二 网络参数，对所述第一行人重识别网络进行更新，得到更新后的所述第一行人重识别网络 , 包括： 接收所述 多个边端服务器返回的所述第二网络参数对应的权重， 其中， 针对任一所述 边端服务器， 所述第二网络参数对应的权重是所述边端服务器根据训练前的所述第二行人 重识别网络和训练后的所述第二行人重识别网络确定得到的； 根据所述 多个边端服务器返回的所述第二网络参数对应的权重， 对所述多个边端服务 器返回的所述第二网络参数进行加权平均， 得到更新后的所述第一网络参数； 根据更新后 的所述第一网络参数， 对所述第一行人重识别网络进行更新， 得到更新后 的所述第一行人重识别网络。 2. The method according to claim 1, wherein the first pedestrian re-identification network is updated according to the second network parameters returned by the plurality of edge servers to obtain the updated all The first pedestrian re-identification network includes: receiving weights corresponding to the second network parameters returned by the plurality of edge servers, wherein, for any of the edge servers, the second network parameter corresponds to The weight is determined by the edge server according to the second pedestrian re-identification network before training and the second pedestrian re-identification network after training; according to the second network returned by the multiple edge servers The weights corresponding to the parameters are weighted and averaged on the second network parameters returned by the plurality of edge servers to obtain the updated first network parameters; according to the updated first network parameters, the The first pedestrian re-identification network is updated to obtain the updated first pedestrian re-identification network.
3、 根据权利要求 1或 2所述的方法， 其中， 所述方法还包括： 向所述多个边端服务器发送共享图像数据集； 接收所述 多个边端服务器返回的伪标签， 其中， 针对任一所述边端服务器， 所述伪标 签是所述边端服务器根据所述共 享图像数据集以及训练后的所述第二行人重识别网络生 成的； 根据所述共 享图像数据集和所述多个边端服务器返回的伪标签， 对更新后的所述第一 行人重识别网络进行训练， 得到训练后的所述第一行人重识别网络。 3. The method according to claim 1 or 2, wherein the method further comprises: sending a shared image data set to the multiple edge servers; receiving pseudo tags returned by the multiple edge servers, wherein, For any of the side servers, the pseudo labels are generated by the side server according to the shared image data set and the trained second person re-identification network; The pseudo labels returned by the plurality of edge servers are used to train the updated first person re-identification network, and the trained first person re-identification network is obtained.
4、 根据权利要求 3所述的方法， 其中， 所述根据所述共享图像数据集和所述多个边端 服务器返回的伪标签， 对更新后的所述第一行人重识别网络进行训练， 得到训练后的所述 第一行人重识别网络， 包括： 根据所述 多个边端服务器返回的伪标签， 确定平均伪标签； 根据所述共 享图像数据集和所述平均伪标签， 对更新后的所述第一行人重识别网络进 行训练， 得到训练后的所述第一行人重识别网络。 4. The method according to claim 3, wherein the updated first person re-identification network is trained according to the shared image data set and the pseudo labels returned by the plurality of edge servers , obtaining the trained first person re-identification network, including: determining an average pseudo-label according to the pseudo-labels returned by the plurality of edge servers; according to the shared image dataset and the average pseudo-label, determining The updated first person re-identification network is trained to obtain the trained first person re-identification network.
5、 一种网络训练方法， 其中， 所述方法应用于边端服务器， 所述边端服务器中包括 第二行人重识别网络、 身份分类网络和本地图像数据集， 所述方法包括： 接收云端服务器发送的第一行人重识别网络对应的第一网络参数， 其中， 所述第一行 人重识别网络和所述第二行人重识别网络具有相同的网络结构； 根据所述本地 图像数据集、 所述身份分类网络和所述第一网络参数 , 对所述第二行人 重识别网络进行训练， 得到训练后的所述第二行人重识别网络， 其中， 所述第二行人重识 别网络对应第二网络参数； 向所述云端服务器发送所述第二网络参数。 5. A network training method, wherein the method is applied to a side server, and the side server includes a second person re-identification network, an identity classification network and a local image dataset, the method comprising: receiving a cloud server Send the first network parameters corresponding to the first pedestrian re-identification network, wherein the first pedestrian re-identification network and the second pedestrian re-identification network have the same network structure; according to the local image data set, The identity classification network and the first network parameters are used to train the second pedestrian re-identification network to obtain the trained second pedestrian re-identification network, wherein the second pedestrian re-identification network corresponds to the first pedestrian re-identification network. two network parameters; sending the second network parameters to the cloud server.
6、 根据权利要求 5所述的方法， 其中， 所述根据所述本地图像数据集、 所述身份分类 网络和所述第一网络参数， 对所述第二行人重识别网络进行训练， 得到训练后的所述第二 行人重识别网络， 包括： 根据所述本地 图像数据集和所述第一网络参数 , 对所述第二行人重识别网络和所述身 份分类网络进行训练， 得到训练后的所述第二行人重识别网络和训练后的所述身份分类网 络。 6. The method according to claim 5, wherein the second pedestrian re-identification network is trained according to the local image data set, the identity classification network and the first network parameters to obtain the training after the second The pedestrian re-identification network includes: training the second pedestrian re-identification network and the identity classification network according to the local image data set and the first network parameters, to obtain the trained second pedestrian re-identification network. A recognition network and the trained identity classification network.
7、 根据权利要求 6所述的方法， 其中， 所述方法还包括： 将训练后的所述身份分类网络存储在所述边端服务器中。 7. The method according to claim 6, wherein the method further comprises: storing the trained identity classification network in the edge server.
8、 根据权利要求 6或 7所述的方法， 其中， 所述本地图像数据集中包括多个身份对应 的图像数据； 所述身份分类网络的维度与所述多个身份的个数相关。 8. The method according to claim 6 or 7, wherein the local image data set includes image data corresponding to multiple identities; the dimension of the identity classification network is related to the number of the multiple identities.
9、 根据权利要求 5至 8中任一项所述的方法， 其中， 所述方法还包括： 接收所述云端服务器发送的共享图像数据集； 根据所述共 享图像数据集和训练后的所述第二行人重识别网络， 生成伪标签； 向所述云端服务器发送所述伪标签。 9. The method according to any one of claims 5 to 8, wherein the method further comprises: receiving a shared image data set sent by the cloud server; The second pedestrian re-identification network generates a pseudo-label; and sends the pseudo-label to the cloud server.
10、 根据权利要求 5至 9任一项所述的方法， 其中， 所述方法还包括： 才艮据训练前的所述第二行人重识别网络和所述本地图像数据集确定第一特征向量， 以 及根据训练后的所述第二行人重识别网络和所述本地图像数据集， 确定第二特征向量； 确定所述第一特征 向量和所述第二特征向量之间的余弦距离； 根据所述余弦距 离， 确定所述第二网络参数对应的权重； 向所述云端服务器发送所述第二网络参数对应的权重。 10. The method according to any one of claims 5 to 9, wherein the method further comprises: determining a first feature vector according to the second person re-identification network before training and the local image data set , and determine a second feature vector according to the trained second person re-identification network and the local image data set; determine the cosine distance between the first feature vector and the second feature vector; the cosine distance, determining the weight corresponding to the second network parameter; and sending the weight corresponding to the second network parameter to the cloud server.
11、根据权利要求 5至 10中任一项所述的方法，其中，所述边端服务器为图像采集设备； 所述本地图像数据集是根据所述图像采集设备采集得到的。 11. The method according to any one of claims 5 to 10, wherein the edge server is an image acquisition device; and the local image data set is acquired according to the image acquisition device.
12、 根据权利要求 5至 10中任一项所述的方法， 其中， 所述边端服务器与至少一个图 像采集设备连接， 所述边端服务器和所述至少一个图像采集设备位于相同地理区域范围； 所述本地图像数据集是所述边端服务器从所述至少一个图像采集设备中获取得到的。 12. The method according to any one of claims 5 to 10, wherein the edge server is connected to at least one image acquisition device, and the edge server and the at least one image acquisition device are located in the same geographical area ; the local image data set is obtained by the edge server from the at least one image acquisition device.
13、 一种行人重识别方法， 包括： 通 过目标行人重识别网络对在目标地理区域范围内获取到的至少一帧待识别 图像进 行行人重识别处理， 确定行人重识别结果； 其 中， 所述目标行人重识别网络采用权利要求 1至 12中任一项所述的网络训练方法训 练得到。 13. A pedestrian re-identification method, comprising: performing pedestrian re-identification processing on at least one frame of images to be identified obtained within a target geographic area through a target pedestrian re-identification network, and determining a pedestrian re-identification result; wherein the target The pedestrian re-identification network is trained by using the network training method described in any one of claims 1 to 12.
14、 根据权利要求 13所述的方法， 其中， 所述目标行人重识别网络为更新后的第一行 人重识别网络或训练后的第一行人重识别网络。 14. The method according to claim 13, wherein the target pedestrian re-identification network is an updated first pedestrian re-identification network or a trained first pedestrian re-identification network.
15、根据权利要求 13所述的方法，其中，在所述目标地理区域范围内包括边端服务器， 且所述边端服务器中包括训练后的第二行人重识别网络的情况下， 所述目标行人重识别网 络为训练后的第二行人重识别网络。 15. The method according to claim 13, wherein, in the case that an edge server is included within the target geographic area, and the edge server includes a trained second person re-identification network, the target The person re-identification network is the second person re-identification network after training.
16、 一种网络训练装置， 所述装置应用于云端服务器， 所述云端服务器中包括第一行 人重识别网络， 所述装置包括： 发送部分 ， 被配置为向多个边端服务器发送所述第一行人重识别网络对应的第一网络 参数； 接收部分 ， 被配置为接收所述多个边端服务器返回的第二网络参数， 其中， 针对任一 所述边端服务器， 所述边端服务器中包括第二行人重识别网络、 身份分类网络和本地图像 数据集， 所述第二行人重识别网络和所述第一行人重识别网络具有相同的网络结构， 所述 第二网络参数是所述边端服务器根据所述本地图像数据集、 所述身份分类网络和所述第一 网络参数对所述第二行人重识别网络进行训练之后得到的； 更新部分 ， 被配置为根据所述多个边端服务器返回的所述第二网络参数， 对所述第一 行人重识别网络进行更新， 得到更新后的所述第一行人重识别网络。 16. A network training device, the device is applied to a cloud server, the cloud server includes a first pedestrian re-identification network, and the device includes: a sending part configured to send the the first network parameter corresponding to the first pedestrian re-identification network; the receiving part is configured to receive the second network parameter returned by the plurality of edge servers, wherein, for any of the edge servers, the edge server The server includes a second person re-identification network, an identity classification network and a local image data set, the second person re-identification network and the first person re-identification network have the same network structure, and the second network parameter is obtained by the edge server after training the second pedestrian re-identification network according to the local image data set, the identity classification network and the first network parameters; the update part is configured to The second network parameter returned by the edge server updates the first pedestrian re-identification network to obtain the updated first pedestrian re-identification network.
17、 一种网络训练装置， 所述装置应用于边端服务器， 所述边端服务器中包括第二行 19 人重识别网络、 身份分类网络和本地图像数据集， 所述装置包括： 接收部分 ， 被配置为接收云端服务器发送的第一行人重识别网络对应的第一网络参数, 其中， 所述第一行人重识别网络和所述第二行人重识别网络具有相同的网络结构； 网络训练部分， 被配置为根据所述本地图像数据集、 所述身份分类网络和所述第一网 络参数， 对所述第二行人重识别网络进行训练， 得到训练后的所述第二行人重识别网络， 其中， 所述第二行人重识别网络对应第二网络参数； 发送部分 ， 被配置为向所述云端服务器发送所述第二网络参数。 17. A network training device, the device is applied to an edge server, and the edge server includes a second line 19 A person re-identification network, an identity classification network and a local image data set, the apparatus includes: a receiving part, configured to receive a first network parameter corresponding to the first pedestrian re-identification network sent by a cloud server, wherein the first The pedestrian re-identification network and the second pedestrian re-identification network have the same network structure; the network training part is configured to, according to the local image data set, the identity classification network and the first network parameters, The second pedestrian re-identification network is trained, and the trained second pedestrian re-identification network is obtained, wherein the second pedestrian re-identification network corresponds to the second network parameter; the sending part is configured to send to the cloud The server sends the second network parameter.
18、 一种行人重识别装置， 包括： 行人重识别部分， 被配置为通过目标行人重识别网络对在目标地理区域范围内获取到 的至少一帧待识别图像进行行人重识别处理， 确定行人重识别结果； 其中， 所述目标行人 重识别网络采用权利要求 1至 12中任一项所述的网络训练方法训练得到。 18. A pedestrian re-identification device, comprising: a pedestrian re-identification part, configured to perform pedestrian re-identification processing on at least one frame of images to be recognized obtained within a target geographical area through a target pedestrian re-identification network, and determine the pedestrian re-identification. The identification result; wherein, the target pedestrian re-identification network is obtained by training the network training method according to any one of claims 1 to 12.
19、 一种电子设备， 包括： 处理器； 被配置为存储处理器可执行指令的存储器； 其 中， 所述处理器被配置为调用所述存储器存储的指令， 以执行权利要求 1至 15中任 意一项所述的方法。 19. An electronic device, comprising: a processor; a memory configured to store instructions executable by the processor; wherein the processor is configured to invoke the instructions stored in the memory to execute any of claims 1 to 15 one of the methods described.
20、 一种计算机可读存储介质， 其上存储有计算机程序指令， 所述计算机程序指令被 处理器执行时实现权利要求 1至 15中任意一项所述的方法。 20. A computer-readable storage medium having computer program instructions stored thereon, the computer program instructions implementing the method of any one of claims 1 to 15 when executed by a processor.
21、 一种计算机程序， 包括计算机可读代码， 当所述计算机可读代码在电子设备中运 行时， 所述电子设备中的处理器执行时实现权利要求 1至 15中任意一项所述的方法。 21. A computer program, comprising computer-readable codes, when the computer-readable codes are executed in an electronic device, the processor in the electronic device implements any one of claims 1 to 15 when executed. method.
PCT/IB2020/060047 2020-08-25 2020-10-27 Network training method and apparatus, person re-identification method and apparatus, storage medium, and computer program WO2022043741A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010864291.4 2020-08-25
CN202010864291.4A CN112001321B (en) 2020-08-25 2020-08-25 Network training method, pedestrian re-identification method, device, electronic equipment and storage medium

Publications (1)

Publication Number Publication Date
WO2022043741A1 true WO2022043741A1 (en) 2022-03-03

Family

ID=73471079

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IB2020/060047 WO2022043741A1 (en) 2020-08-25 2020-10-27 Network training method and apparatus, person re-identification method and apparatus, storage medium, and computer program

Country Status (3)

Country Link
CN (1) CN112001321B (en)
TW (1) TWI769635B (en)
WO (1) WO2022043741A1 (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114743243A (en) * 2022-04-06 2022-07-12 平安科技（深圳）有限公司 Human face recognition method, device, equipment and storage medium based on artificial intelligence
CN115022316A (en) * 2022-05-20 2022-09-06 阿里巴巴（中国）有限公司 End cloud cooperative data processing system, method, equipment and computer storage medium
CN115310130A (en) * 2022-08-15 2022-11-08 南京航空航天大学 Multi-site medical data analysis method and system based on federal learning
CN115601791A (en) * 2022-11-10 2023-01-13 江南大学（Cn） Unsupervised pedestrian re-identification method based on Multiformer and outlier sample re-distribution
CN117851838A (en) * 2024-03-07 2024-04-09 广州大学 Identification method of heterogeneous data sources in collaborative learning process

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112507893A (en) * 2020-12-14 2021-03-16 华南理工大学 Distributed unsupervised pedestrian re-identification method based on edge calculation
CN112906857B (en) * 2021-01-21 2024-03-19 商汤国际私人有限公司 Network training method and device, electronic equipment and storage medium
CN112861695B (en) * 2021-02-02 2023-10-24 北京大学 Pedestrian identity re-identification method and device, electronic equipment and storage medium
CN112906677B (en) * 2021-05-06 2021-08-03 南京信息工程大学 Pedestrian target detection and re-identification method based on improved SSD (solid State disk) network
CN113205863B (en) * 2021-06-04 2022-03-25 广西师范大学 Training method of individualized model based on distillation semi-supervised federal learning
CN113326938A (en) * 2021-06-21 2021-08-31 商汤国际私人有限公司 Network training method, pedestrian re-identification method, network training device, pedestrian re-identification device, electronic equipment and storage medium
CN113326939A (en) * 2021-06-21 2021-08-31 商汤国际私人有限公司 Network training method, pedestrian re-identification method, network training device, pedestrian re-identification device, electronic equipment and storage medium
CN113792606B (en) * 2021-08-18 2024-04-26 清华大学 Low-cost self-supervision pedestrian re-identification model construction method based on multi-target tracking
CN113807369A (en) * 2021-09-26 2021-12-17 北京市商汤科技开发有限公司 Target re-identification method and device, electronic equipment and storage medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107851213A (en) * 2015-07-22 2018-03-27 高通股份有限公司 Shift learning in neutral net
CN108230296A (en) * 2017-11-30 2018-06-29 腾讯科技（深圳）有限公司 The recognition methods of characteristics of image and device, storage medium, electronic device
CN109993300A (en) * 2017-12-29 2019-07-09 华为技术有限公司 A kind of training method and device of neural network model
EP3528179A1 (en) * 2018-02-15 2019-08-21 Koninklijke Philips N.V. Training a neural network
CN110490058A (en) * 2019-07-09 2019-11-22 北京迈格威科技有限公司 Training method, device, system and the computer-readable medium of pedestrian detection model
CN111126108A (en) * 2018-10-31 2020-05-08 北京市商汤科技开发有限公司 Training method and device of image detection model and image detection method and device

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090112455A1 (en) * 2007-10-24 2009-04-30 Yahoo! Inc. Method and system for rendering simplified point finding maps
WO2017024242A1 (en) * 2015-08-05 2017-02-09 Equifax Inc. Model integration tool
CN107563327B (en) * 2017-08-31 2021-07-20 武汉大学 Pedestrian re-identification method and system based on self-walking feedback
CN110795477A (en) * 2019-09-20 2020-02-14 平安科技（深圳）有限公司 Data training method, device and system
CN110825900A (en) * 2019-11-07 2020-02-21 重庆紫光华山智安科技有限公司 Training method of feature reconstruction layer, reconstruction method of image features and related device
CN110956202B (en) * 2019-11-13 2023-08-01 重庆大学 Image training method, system, medium and intelligent device based on distributed learning
CN111291611A (en) * 2019-12-20 2020-06-16 长沙千视通智能科技有限公司 Pedestrian re-identification method and device based on Bayesian query expansion
CN111107094B (en) * 2019-12-25 2022-05-20 青岛大学 Lightweight ground-oriented medical Internet of things big data sharing system
CN111241580B (en) * 2020-01-09 2022-08-09 广州大学 Trusted execution environment-based federated learning method
CN111401281B (en) * 2020-03-23 2022-06-21 山东师范大学 Unsupervised pedestrian re-identification method and system based on deep clustering and sample learning

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107851213A (en) * 2015-07-22 2018-03-27 高通股份有限公司 Shift learning in neutral net
CN108230296A (en) * 2017-11-30 2018-06-29 腾讯科技（深圳）有限公司 The recognition methods of characteristics of image and device, storage medium, electronic device
CN109993300A (en) * 2017-12-29 2019-07-09 华为技术有限公司 A kind of training method and device of neural network model
EP3528179A1 (en) * 2018-02-15 2019-08-21 Koninklijke Philips N.V. Training a neural network
CN111126108A (en) * 2018-10-31 2020-05-08 北京市商汤科技开发有限公司 Training method and device of image detection model and image detection method and device
CN110490058A (en) * 2019-07-09 2019-11-22 北京迈格威科技有限公司 Training method, device, system and the computer-readable medium of pedestrian detection model

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114743243A (en) * 2022-04-06 2022-07-12 平安科技（深圳）有限公司 Human face recognition method, device, equipment and storage medium based on artificial intelligence
CN114743243B (en) * 2022-04-06 2024-05-31 平安科技（深圳）有限公司 Human face recognition method, device, equipment and storage medium based on artificial intelligence
CN115022316A (en) * 2022-05-20 2022-09-06 阿里巴巴（中国）有限公司 End cloud cooperative data processing system, method, equipment and computer storage medium
CN115022316B (en) * 2022-05-20 2023-08-11 阿里巴巴（中国）有限公司 End cloud collaborative data processing system, method, equipment and computer storage medium
CN115310130A (en) * 2022-08-15 2022-11-08 南京航空航天大学 Multi-site medical data analysis method and system based on federal learning
CN115310130B (en) * 2022-08-15 2023-11-17 南京航空航天大学 Multi-site medical data analysis method and system based on federal learning
CN115601791A (en) * 2022-11-10 2023-01-13 江南大学（Cn） Unsupervised pedestrian re-identification method based on Multiformer and outlier sample re-distribution
CN115601791B (en) * 2022-11-10 2023-05-02 江南大学 Unsupervised pedestrian re-identification method based on multi-former and outlier sample re-distribution
CN117851838A (en) * 2024-03-07 2024-04-09 广州大学 Identification method of heterogeneous data sources in collaborative learning process

Also Published As

Publication number Publication date
CN112001321B (en) 2024-06-14
TW202209151A (en) 2022-03-01
TWI769635B (en) 2022-07-01
CN112001321A (en) 2020-11-27

Similar Documents

Publication Publication Date Title
WO2022043741A1 (en) Network training method and apparatus, person re-identification method and apparatus, storage medium, and computer program
WO2021196401A1 (en) Image reconstruction method and apparatus, electronic device and storage medium
WO2020135127A1 (en) Pedestrian recognition method and device
WO2021031609A1 (en) Living body detection method and device, electronic apparatus and storage medium
TWI767596B (en) Scene depth and camera motion prediction method, electronic equipment and computer readable storage medium
WO2022188305A1 (en) Information presentation method and apparatus, and electronic device, storage medium and computer program
TWI767217B (en) Coordinate system alignment method and apparatus, electronic device and computer-readable storage medium
WO2021036382A1 (en) Image processing method and apparatus, electronic device and storage medium
CN111241887B (en) Target object key point identification method and device, electronic equipment and storage medium
TWI778313B (en) Method and electronic equipment for image processing and storage medium thereof
TWI757668B (en) Network optimization method and device, image processing method and device, storage medium
CN111563138B (en) Positioning method and device, electronic equipment and storage medium
WO2022151686A1 (en) Scene image display method and apparatus, device, storage medium, program and product
WO2022134475A1 (en) Point cloud map construction method and apparatus, electronic device, storage medium and program
WO2022179013A1 (en) Object positioning method and apparatus, electronic device, storage medium, and program
WO2022193507A1 (en) Image processing method and apparatus, device, storage medium, program, and program product
TW202205127A (en) Target detection method, electronic equipment and computer-readable storage medium
WO2023024439A1 (en) Behavior recognition method and apparatus, electronic device and storage medium
CN113506229B (en) Neural network training and image generating method and device
TWI770531B (en) Face recognition method, electronic device and storage medium thereof
TWI751593B (en) Network training method and device, image processing method and device, electronic equipment, computer readable storage medium and computer program
CN113283343A (en) Crowd positioning method and device, electronic equipment and storage medium
CN113506324B (en) Image processing method and device, electronic equipment and storage medium
US20210383580A1 (en) Method, apparatus and system for anchor sharing, electronic device and storage medium
WO2022110801A1 (en) Data processing method and apparatus, electronic device, and storage medium

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20951311

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20951311

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 522432273

Country of ref document: SA