Image data processing method and device, electronic device, and storage medium

Abstract

The image data processing method and device, electronic device, and storage medium proposed in this application relate to the field of image processing technology and financial technology. The method is applied to a security platform, and includes: storing sample images in an archive database and generating an image storage path; generating image training data based on image structured data and image storage path, and storing the image training data in a distributed database; obtaining image training data from a distributed database based on a model training request, extracting sample images from an archive database based on the image storage path in the image training data, and training a preset initial image processing model based on the image training data and sample images in a secure execution space to obtain a target image processing model; obtaining a target image, and performing image processing on the target image through the target image processing model. The embodiment of the present application can perform model training under the premise of ensuring data security and ensure the effectiveness of model training.

Description

Technical Field

[0001] The present application relates to the fields of image processing technology and financial technology, and in particular to an image data processing method and device, electronic equipment, and storage medium. Background Art

[0002] Currently, many scenarios require image data, including both images and image-related data. For example, in the FinTech field, image data is required for training new models and optimizing existing models. FinTech models can include document image processing models and facial image processing models.

[0003] There are multiple channels for acquiring image data in related technologies. For example, one channel first obtains authorization from the data provider, then operators obtain the image data from the data provider, and then share the image data with developers. However, for security reasons, some image data does not want to be publicly available in plain text. To this end, qualified desensitization tools are often required to desensitize the image data, and then use the desensitized image data for model training. However, desensitization significantly affects the quality of the image data and can seriously affect the effectiveness of model training. Summary of the Invention

[0004] The main purpose of the embodiments of the present application is to propose an image data processing method and device, electronic device, and storage medium that can perform model training while ensuring data security and ensure the effectiveness of model training.

[0005] To achieve the above objectives, a first aspect of an embodiment of the present application provides an image data processing method, which is applied to a security platform. The method includes:

[0006] Acquire a sample image and image structured data from a service platform end; wherein the image structured data is obtained by the service platform end identifying the sample image;

[0007] Storing the sample image in a preset archiving database and generating an image storage path;

[0008] generating image training data based on the image structured data and the image storage path, and storing the image training data in a preset distributed database; wherein the image training data includes the image storage path, and the security level of the distributed database is higher than the security level of the archive database;

[0009] Receiving a model training request from a development platform end, wherein the development platform end does not have permission to access the business platform end;

[0010] Obtaining the image training data from the distributed database according to the model training request, extracting the sample image from the archive database according to the image storage path in the image training data, and storing the image training data and the sample image in a preset secure execution space; and training a preset initial image processing model in the secure execution space according to the image training data and the sample image to obtain a target image processing model; wherein the secure execution space is not accessible from outside the secure execution space;

[0011] A target image is acquired, and image processing is performed on the target image using the target image processing model.

[0012] In some embodiments, the secure execution space includes a secure database and a secure host; storing the image training data and the sample images in a preset secure execution space, and training the initial image processing model in the secure execution space based on the image training data and the sample images to obtain a target image processing model includes:

[0013] Storing the image training data and the sample image in the secure database; wherein the secure database is not accessible by hosts other than the secure host;

[0014] extracting image training data and the sample image from the security database via the security host;

[0015] The initial image processing model is trained by the secure host using the image training data and the sample image to obtain the target image processing model.

[0016] In some embodiments, training the initial image processing model using the image training data and the sample image by the secure host to obtain the target image processing model includes:

[0017] Obtaining image processing resources from a preset resource providing host through the security host; the security host does not have the authority to be accessed by hosts other than the resource providing host, and the resource providing host does not have the authority to be accessed by hosts other than the security host;

[0018] The initial image processing model is trained by the secure host with the image processing resources, the image training data and the sample images to obtain the target image processing model.

[0019] In some embodiments, before receiving the model training request from the development platform, the method further includes:

[0020] Receive permission application requests from the development platform;

[0021] Sending the permission application request to a preset permission management terminal, which approves the permission application request and sends the approval result to the security platform terminal; wherein the approval result indicates whether the development platform terminal is authorized or not;

[0022] If the approval result indicates that the development platform is authorized, the identification of the development platform is sent to the security host, and the security host marks the development platform as authorized. The security host is used to receive the model training request from the authorized development platform.

[0023] In some embodiments, after generating image training data based on the image structured data and the image storage path, and storing the image training data in a preset distributed database, the method further includes:

[0024] Receive image data query requests through a preset secure cloud desktop;

[0025] Obtaining the sample image from the archive database according to the image data query request, extracting the image training data from the distributed database, and transmitting the sample image and the image training data to the secure cloud desktop;

[0026] The sample image and the image training data are displayed through the secure cloud desktop; wherein the secure cloud desktop does not have permission for download operations.

[0027] In some embodiments, obtaining the sample image from the archive database according to the image data query request, extracting the image training data from the distributed database, and transmitting the sample image and the image training data to the secure cloud desktop includes:

[0028] Exporting the image data query request from the secure cloud desktop through a preset visual front end, and transmitting the image data query request to a preset trusted intelligent platform;

[0029] Obtaining the sample image from the archive database and extracting the image training data from the distributed database through the trusted intelligent platform using the image data query request, and transmitting the sample image and the image training data to the visualization front end;

[0030] Through the visualization front end, the sample image and the image training data are respectively desensitized, and the desensitized sample image and the desensitized image training data are transmitted to the secure cloud desktop;

[0031] Among them, the secure cloud desktop does not have the authority to communicate directly with the trusted intelligent platform.

[0032] In some embodiments, the image data query request is sent by the annotation platform;

[0033] After displaying the sample image and the image training data through the secure cloud desktop, the method further includes:

[0034] Receiving, through the secure cloud desktop, a labeling instruction for the sample image from the labeling platform;

[0035] Exporting the annotation instructions from the secure cloud desktop through a preset visual front end and transmitting them to a preset trusted intelligent platform;

[0036] Annotating the sample image using the annotation instruction via the trusted intelligent platform to obtain image annotation data;

[0037] The image annotation data is stored in the distributed database through the trusted intelligent platform.

[0038] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides an image data processing device, which is applied to a security platform, and includes:

[0039] A data acquisition module, configured to acquire a sample image and image structured data from a service platform; wherein the image structured data is obtained by the service platform identifying the sample image;

[0040] An image archiving module, configured to store the sample image in a preset archiving database and generate an image storage path;

[0041] a data generation module, configured to generate image training data based on the image structured data and the image storage path, and store the image training data in a preset distributed database; wherein the image training data includes the image storage path, and the security level of the distributed database is higher than the security level of the archive database;

[0042] A request receiving module, configured to receive a model training request from a development platform, wherein the development platform does not have permission to access the business platform;

[0043] a model training module, configured to obtain the image training data from the distributed database according to the model training request, extract the sample images from the archive database according to the image storage path in the image training data, and store the image training data and the sample images in a preset secure execution space; and train a preset initial image processing model in the secure execution space according to the image training data and the sample images to obtain a target image processing model; wherein the secure execution space is not accessible from outside the secure execution space;

[0044] The image processing module is used to obtain a target image and perform image processing on the target image through the target image processing model.

[0045] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the image data processing method described in the first aspect when executing the computer program.

[0046] To achieve the above-mentioned purpose, the fourth aspect of the embodiments of the present application proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the image data processing method described in the first aspect above.

[0047] The image data processing method and apparatus, electronic device, and storage medium proposed in this application first establish a communication architecture comprising a business platform, a development platform, and a security platform, with the development platform prohibited from accessing the business platform. The business platform serves as the source of data production. The security platform can obtain sample images and structured image data from the business platform, then store the sample images in an archive database and store image training data containing the structured image data in a distributed database. The distributed database has a higher security level than the archive database. This not only achieves decoupled storage of sample images and image training data, but also ensures the security of the image training data. When the development platform requires model training, it can send a model training request to the security platform. Based on the model training request, the security platform obtains image training data and sample images from the distributed database and the archive database, respectively. Then, within the secure execution space, the security platform trains an initial image processing model based on the image training data and sample images to obtain a target image processing model. After obtaining the target image, the target image is processed using the target image processing model. The secure execution space is not accessible from outside the secure execution space, thus providing high security. In this way, the development platform can meet the model training requirements without obtaining the target image and image training data. During the model training process, not only can the target image and image training data be used to train the model, but these data will not flow out of the secure execution space, and the data security is relatively high. Moreover, these data have not been desensitized, and the data integrity is good, so the model training effect is better. In summary, this application can perform model training under the premise of ensuring data security and ensure the model training effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is a flowchart of the image data processing method provided in an embodiment of the present application;

[0049] Figure 2 is a flowchart of an image data processing method provided by another embodiment of the present application;

[0050] Figure 3 yes Figure 1 Flowchart of step 105 in FIG.

[0051] Figure 4 yes Figure 3 Flowchart of step 303 in FIG.

[0052] Figure 5 is a flowchart of an image data processing method provided by another embodiment of the present application;

[0053] Figure 6 yes Figure 5 Flowchart of step 502 in FIG.

[0054] Figure 7 is a flowchart of an image data processing method provided by another embodiment of the present application;

[0055] Figure 8 This is a block diagram of the module structure of the image data processing device provided in an embodiment of the present application;

[0056] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0058] It should be noted that although the device schematics illustrate functional module divisions and the flowcharts illustrate logical sequences, in certain circumstances, the steps shown or described may be performed in a sequence that differs from the module divisions in the device or the sequence in the flowcharts. The terms "first," "second," and so on, in the specification, claims, and drawings, are used to distinguish similar items and are not necessarily used to describe a specific sequence or precedence.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0060] Currently, many scenarios require the use of image data, including both images and image-related data. There are multiple channels for acquiring image data for related technologies. Channel 1 involves obtaining authorization from the data provider, followed by data collection by operations personnel, and then sharing the image data files with developers. The disadvantages of this channel 1 are: 1. Sensitive data cannot be transferred outside the production environment unless it is desensitized; otherwise, there is no guarantee that sensitive information will not be leaked; 2. Qualified desensitization tools are not available; 3. Desensitization can severely impact algorithm training effectiveness; 4. The process is lengthy, requires frequent manual intervention, and is slow. Channel 2 involves crawling the web to acquire data. The disadvantages of this channel 2 are: 1. The acquired data quality is low and insufficient to meet requirements; 2. The crawler's scope is limited, resulting in insufficient data richness and dimensionality; 3. Crawling data related to specialized businesses, such as medical and financial invoices, cannot be acquired through crawling. Channel 3: Publicly available open source datasets. The disadvantages of this channel 3 are: 1. Authorization is required, and use cannot exceed the authorized scope. Development and testing environments cannot effectively detect developers / R&D personnel, who may exceed their authorized scope.

[0061] Based on this, the main purpose of the embodiments of the present application is to propose an image data processing method and an image data processing device, an electronic device, and a storage medium, aiming to obtain and use image data with the greatest freedom on the basis of ensuring the security of image data.

[0062] The image data processing method provided in the embodiments of the present application is applied to the server side, and can also be software running on the terminal or server side. The server side can be configured as an independent physical server, or as a server cluster or distributed system composed of multiple physical servers, or as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the image data processing method, etc., but is not limited to the above forms.

[0063] The present application can be used in many general or special computer system environments or configurations. For example: server computers, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, network PCs, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0064] The embodiments of the present application provide an image data processing method, an image data processing device, an electronic device, and a storage medium, which are specifically illustrated by the following embodiments. First, the image data processing method in the embodiments of the present application is described.

[0065] Figure 1 This is an optional flowchart of the image data processing method provided in an embodiment of the present application. The image data processing method is applied to the security platform end and may include but is not limited to steps 101 to 106.

[0066] Step 101: Obtain sample images and image structured data from the service platform.

[0067] Step 102: Store the sample image in a preset archiving database and generate an image storage path;

[0068] Step 103: Generate image training data based on the image structured data and the image storage path, and store the image training data in a preset distributed database;

[0069] Step 104: receiving a model training request from the development platform;

[0070] Step 105: Obtain image training data from a distributed database according to the model training request, extract sample images from an archive database according to the image storage path in the image training data, and store the image training data and sample images in a preset secure execution space. Within the secure execution space, train a preset initial image processing model based on the image training data and sample images to obtain a target image processing model.

[0071] Step 106: Acquire a target image and perform image processing on the target image using a target image processing model.

[0072] Steps 101 to 106 illustrated in the embodiment of this application are performed by first constructing a communication architecture comprising a business platform, a development platform, and a security platform, with the development platform prohibited from accessing the business platform. The business platform serves as the source of data production. The security platform can obtain sample images and structured image data from the business platform, then store the sample images in an archive database and store image training data containing the structured image data in a distributed database. The distributed database has a higher security level than the archive database. This not only achieves decoupled storage of sample images and image training data, but also ensures the security of the image training data. When the development platform requires model training, it can send a model training request to the security platform. Based on the model training request, the security platform obtains image training data and sample images from the distributed database and the archive database, respectively. Then, within the secure execution space, the security platform trains the initial image processing model based on the image training data and sample images to obtain the target image processing model. After obtaining the target image, the target image is processed using the target image processing model. The secure execution space is not accessible from outside the secure execution space, thus providing a high level of security. In this way, the development platform can meet the model training requirements without obtaining the target image and image training data. During the model training process, not only can the target image and image training data be used to train the model, but these data will not flow out of the secure execution space, and the data security is relatively high. Moreover, these data have not been desensitized, and the data integrity is good, so the model training effect is better. In summary, this application can perform model training under the premise of ensuring data security and ensure the model training effect.

[0073] Before introducing step 101 to step 106 in detail, the communication architecture provided by the embodiment of the present application is first introduced. In one embodiment, the communication architecture includes a business platform side, a development platform side, and a security platform side.

[0074] The business platform is a terminal / server that provides business services. It includes a business service platform and a business database. Users can upload target images to the business service platform by calling the business service interface. The business service platform then uses the recognition algorithm / recognition model to identify the target image and obtain structured image data.

[0075] The development platform side is a terminal / server side for business development. The development platform side can build an initial image processing model by writing program code. When the development platform side needs to perform model training on the initial image processing model, the three channels mentioned above can be used to obtain image-related data for model training. In an embodiment of the present application, the development platform side is prohibited from accessing the business platform side, and the development platform side can access the security platform side. For example, the development platform side generates a model training request based on the initial image processing model and sends the model training request to the security platform side.

[0076] The security platform side is a cluster of devices that owns and uses image data. The security platform side can obtain image data from the business platform side, specifically including target images and image structured data. After receiving a model training request, the security platform side can use the image data to train the initial image processing model. Specifically, in the process of using image data, image data leakage may occur, so the security platform side is provided with a secure execution space, which prohibits access from outside the secure execution space. Within the secure execution space, the security platform side trains the initial image processing model based on the image training data and the target image to obtain the target image processing model. The security platform side includes a trusted execution space.

[0077] In step 101 of some embodiments, the security platform obtains a sample image and image structured data from the business platform. The image structured data is obtained by the business platform through recognition of the sample image. The business platform has been introduced above and will not be repeated here. The sample image can be referred to as unstructured image data. Unstructured image data is the image to be recognized by an algorithm such as OCR, such as an image of a document or a PDF contract. Image structured data is the JSON format of the image unstructured data obtained after algorithmic discrimination and recognition, such as the content of the recognized image.

[0078] In one embodiment, step 101 may include: according to the acquisition cycle pre-set by the security platform end, each acquisition cycle passes, sending an image data acquisition request to the business platform end, and receiving a sample image and image structured data sent by the business platform end based on the image data acquisition request.

[0079] In step 102 of some embodiments, the security platform stores the sample image in a preset archiving database and generates an image storage path. The archiving database is a database capable of storing images. For example, the archiving database may be a network storage server (NAS). The image storage path indicates the path where the target image is stored in the archiving database.

[0080] In step 103 of some embodiments, the security platform generates image training data based on the image structured data and the image storage path, and stores the image training data in a pre-set distributed database. The image training data includes the image storage path. The image training data also includes the image structured data. The distributed database has a higher security level than the archive database. The distributed database can be an Elasticsearch database.

[0081] In one embodiment, the process of storing the image training data in a pre-set distributed database may include: encrypting the image training data using a data encryption algorithm to obtain encrypted image training data; and storing the encrypted image training data in the distributed database. Subsequent steps require first obtaining the encrypted image training data from the distributed database, and then decrypting the encrypted image training data using a data decryption algorithm to obtain the image training data. The data encryption algorithm and the data decryption algorithm form a pair of symmetric encryption and decryption algorithms.

[0082] In one embodiment, referring to Figure 2 Before step 104, the image data processing method provided in the embodiment of the present application may further include:

[0083] Step 201: receiving a permission request from a development platform;

[0084] Step 202: Send the permission application request to the preset permission management terminal, which will review and approve the permission application request and send the approval result to the security platform terminal;

[0085] In step 203, if the approval result indicates that the development platform is authorized, the identification of the development platform is sent to the security host, which marks the development platform as authorized. The security host is used to receive model training requests from the authorized development platform.

[0086] The permission management terminal is a terminal / server that approves permission requests. The approval result can indicate whether the development platform is authorized or not. For example, if a user requires model training permissions, the development platform sends a permission request to the permission management terminal. The permission management terminal then sends the permission request to the direct supervisor, department head, and information security department for approval. Once approved, the user's development platform obtains the corresponding permissions.

[0087] The benefit of the embodiment of the above steps 201 to 203 is that, with the help of the permission management end, the security platform end can mark whether the development platform end has the model training permission, thereby realizing secure communication interconnection.

[0088] In step 104 of some embodiments, the security platform receives a model training request from the development platform. The development platform does not have access to the business platform. The development platform cannot obtain sample images and image structured data from the business platform. A model training request is a request for model training.

[0089] In step 105 of some embodiments, the secure platform obtains image training data from a distributed database based on the model training request, extracts sample images from an archive database based on the image storage path in the image training data, and stores the image training data and sample images in a pre-set secure execution space. Within the secure execution space, the secure platform trains a pre-set initial image processing model based on the image training data and sample images to obtain a target image processing model. The secure execution space is not accessible from outside the secure execution space.

[0090] The secure execution space includes a secure database and a secure host. The secure database does not have access permissions from hosts other than the secure host. In one embodiment, referring to Figure 3 , step 105 may include:

[0091] Step 301, storing image training data and sample images in a secure database;

[0092] Step 302: extracting image training data and sample images from a secure database via a secure host;

[0093] Step 303: The initial image processing model is trained using the image training data and sample images via the secure host to obtain a target image processing model.

[0094] A trusted execution space (TES), also known as a trusted execution environment (TEE), is a hardware-based secure computing solution. Using hardware and software, a TES creates a secure zone within a server cluster, ensuring the confidentiality and integrity of loaded programs and data. A secure database is a container for storing data, such as a network storage server (NAS). A secure host, also known as a safe house server, is a device that can access the secure database and perform model training.

[0095] The benefit of the embodiment of the above steps 301 to 303 is that model training can be performed on the basis of ensuring data security, and the security host can access the security host at any time, which is more flexible.

[0096] In one embodiment, referring to Figure 4 , step 203 may include:

[0097] Step 401: Obtain image processing resources from a preset resource providing host via a secure host;

[0098] Step 402 : The initial image processing model is trained by the secure host using image processing resources, image training data, and sample images to obtain a target image processing model.

[0099] Specifically, a resource-providing host refers to a host that provides image processing-related resources. Resources provided by a resource-providing host may include collective PIPY repositories, collective YUM repositories, collective image repositories, etc. A secure host is not accessible by hosts other than the resource-providing host. A resource-providing host is not accessible by hosts other than the secure host.

[0100] The benefit of the embodiment of steps 401 to 402 is that the security host is allowed to obtain required resources from the resource providing host, but the security host and the resource providing host are not allowed to access other hosts, thereby reducing the possibility of data leakage while being able to obtain required resources.

[0101] In one embodiment, the process of training the initial image processing model using image training data and sample images to obtain a target image processing model may include: inputting the sample image into the initial image processing model for image processing to obtain sample image processing data; calculating loss data based on the sample image processing data and the image training data to obtain target loss data; and adjusting the parameters of the initial image processing model based on the target loss data to obtain the target image processing model. For example, the image training data includes the actual image category of the sample image, and the sample image processing data includes the predicted image category of the sample image. Based on the actual image category and the predicted image category, the target loss data can be calculated as the target category loss data. The parameters of the initial image processing model can be adjusted based on the target category loss data to obtain the target image processing model as a target image classification model.

[0102] The above-mentioned loss data can be calculated using error functions such as mean square error loss function and similarity function.

[0103] The security platform includes a security cloud desktop. In one embodiment, after step 103, refer to Figure 5 The image data processing method provided in this embodiment may further include:

[0104] Step 501: receiving an image data query request via a preset secure cloud desktop;

[0105] Step 502: Obtain sample images from the archive database and extract image training data from the distributed database according to the image data query request, and transmit the sample images and image training data to the secure cloud desktop;

[0106] Step 503: Display the sample images and image training data via the secure cloud desktop.

[0107] Specifically, secure cloud desktops do not have download permissions. Secure cloud desktops refer to webssh resource pools. Data and text on cloud desktops cannot be downloaded or copied locally. User behavior is monitored, and data can be manually cleared and reset regularly.

[0108] The benefit of the embodiment of the above steps 501 to 503 is that when there is an image data query request, data can be displayed through a secure cloud desktop that does not allow download operations, reducing the risk of data leakage and ensuring data security.

[0109] The security platform also includes a visual front end and a trusted intelligent platform. In one embodiment, refer to Figure 6 , step 502 may include:

[0110] Step 601: Export an image data query request from the secure cloud desktop through a preset visual front end, and transmit the image data query request to a preset trusted intelligent platform;

[0111] Step 602: Obtain sample images from the archive database and extract image training data from the distributed database through the trusted intelligent platform using an image data query request, and transmit the sample images and image training data to the visualization front end;

[0112] Step 603: Desensitize the sample images and the image training data respectively through the visualization front end, and transmit the desensitized sample images and the desensitized image training data to the secure cloud desktop.

[0113] Specifically, the secure cloud desktop does not have direct communication permissions with the trusted intelligent platform. The visual front-end and the trusted intelligent platform are both processing units capable of performing corresponding tasks. The visual front-end serves as an intermediate station for data forwarding between the secure cloud desktop and the trusted intelligent platform. The visual front-end is pre-configured with desensitization rules, which can be used to desensitize sample images and image training data.

[0114] The benefit of the embodiment of the above steps 601 to 603 is that it prevents sensitive information from leaking out, thereby further improving data security.

[0115] The image data query request is sent by the annotation platform. In one embodiment, refer to Figure 7 After step 503, the image data processing method provided by this embodiment may further include:

[0116] Step 701: receiving a labeling instruction for a sample image from a labeling platform via a secure cloud desktop;

[0117] Step 702: Export the annotation instructions from the secure cloud desktop through the preset visual front end and transmit them to the preset trusted intelligent platform;

[0118] Step 703: annotate the sample image using the annotation instruction via the trusted intelligent platform to obtain image annotation data;

[0119] Step 704: Store the image annotation data in a distributed database through the trusted intelligent platform.

[0120] Specifically, the annotation platform refers to the terminal / server used by annotators. Annotators can use the annotation platform to perform data annotation operations on a secure cloud desktop, making data annotation convenient and efficient.

[0121] The benefit of the embodiment of the above steps 701 to 704 is that the sample image can be secondary labeled based on the secure cloud desktop, which has high flexibility and security.

[0122] See also Figure 8 The embodiment of the present application also provides an image data processing device, which can implement the above-mentioned image data processing method. Figure 8 The module structure block diagram of the image data processing device provided in the embodiment of the present application includes: a data acquisition module 801, an image archiving module 802, a data generation module 803, a request receiving module 804, a model training module 805 and an image processing module 806. Among them, the data acquisition module 801 is used to obtain sample images and image structured data from the business platform end; wherein, the image structured data is obtained by the business platform end through recognition of the sample image; the image archiving module 802 is used to store the sample image in a preset archiving database and generate an image storage path; the data generation module 803 is used to generate image training data based on the image structured data and the image storage path, and store the image training data in a preset distributed database; wherein, the image training data includes the image storage path, and the security level of the distributed database is higher than the security level of the archiving database; the request receiving module 804 is used to receive the model training data from the development platform end. Request, wherein the development platform side does not have the authority to access the business platform side; the model training module 805 is used to obtain image training data from the distributed database according to the model training request, extract sample images from the archive database according to the image storage path in the image training data, and store the image training data and sample images in a preset secure execution space, and in the secure execution space, train the preset initial image processing model according to the image training data and the sample images to obtain the target image processing model; wherein the secure execution space does not have the authority to be accessed from outside the secure execution space; the image processing module 806 is used to obtain the target image and perform image processing on the target image through the target image processing model.

[0123] In another embodiment, the image data processing device also includes: a permission approval module, which is specifically used to: receive a permission application request from the development platform end; send the permission application request to a preset permission management end, and the permission management end approves the permission application request and sends the approval result to the security platform end; wherein the approval result indicates whether the development platform end is authorized or not; if the approval result indicates that the development platform end is authorized, the identification of the development platform end is sent to the security host, and the security host marks the development platform end as authorized, and the security host is used to receive the model training request from the authorized development platform end.

[0124] In another embodiment, the image data processing device also includes: a data query module, specifically used to: receive image data query requests through a preset secure cloud desktop; obtain sample images from the archive database according to the image data query request, and extract image training data from the distributed database, and transmit the sample images and image training data to the secure cloud desktop; display the sample images and image training data through the secure cloud desktop; wherein the secure cloud desktop does not have the authority for download operations.

[0125] In another embodiment, the image data processing device also includes: a data annotation module, which is specifically used to: receive annotation instructions for sample images from the annotation platform through a secure cloud desktop; export annotation instructions from the secure cloud desktop through a preset visual front end, and transmit them to a preset trusted intelligent platform; annotate the sample image with the annotation instructions through the trusted intelligent platform to obtain image annotation data; and store the image annotation data in a distributed database through the trusted intelligent platform.

[0126] It should be noted that the specific implementation of the image data processing device is basically the same as the specific embodiment of the above-mentioned image data processing method, and will not be repeated here.

[0127] The present application also provides an electronic device comprising: a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for enabling communication between the processor and the memory. When the program is executed by the processor, the aforementioned image data processing method is implemented. The electronic device may be any intelligent terminal, such as a tablet computer or an in-vehicle computer.

[0128] See also Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is shown. The electronic device includes:

[0129] The processor 901 can be implemented as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application.

[0130] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 902 and is called by the processor 901 to execute the image data processing method of the embodiments of this application.

[0131] Input / output interface 903, used to implement information input and output;

[0132] Communication interface 904, used to implement communication interaction between this device and other devices, which can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, Wi-Fi, Bluetooth, etc.);

[0133] Bus 905 , which transmits information between various components of the device (e.g., processor 901 , memory 902 , input / output interface 903 , and communication interface 904 );

[0134] The processor 901 , the memory 902 , the input / output interface 903 and the communication interface 904 are connected to each other in communication within the device via a bus 905 .

[0135] An embodiment of the present application also provides a storage medium, which is a computer-readable storage medium used for computer-readable storage. The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the above-mentioned image data processing method.

[0136] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0137] The embodiments described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0138] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0139] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0140] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0141] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0142] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0143] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0144] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0145] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0146] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions for enabling an electronic device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store programs, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0147] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A method for processing image data, characterized in that: Applied to a security platform, the method includes: Acquire a sample image and image structured data from a service platform end; wherein the image structured data is obtained by the service platform end identifying the sample image; Storing the sample image in a preset archiving database and generating an image storage path; generating image training data based on the image structured data and the image storage path, and storing the image training data in a preset distributed database; wherein the image training data includes the image storage path, and the security level of the distributed database is higher than the security level of the archive database; Receiving a model training request from a development platform end, wherein the development platform end does not have permission to access the business platform end; Obtaining the image training data from the distributed database according to the model training request, extracting the sample image from the archive database according to the image storage path in the image training data, and storing the image training data and the sample image in a preset secure execution space; and training a preset initial image processing model in the secure execution space according to the image training data and the sample image to obtain a target image processing model; wherein the secure execution space is not accessible from outside the secure execution space; Acquire a target image, and perform image processing on the target image using the target image processing model; The method further comprises: Receiving, via a preset secure cloud desktop, a labeling instruction for the sample image from the labeling platform; Exporting the annotation instructions from the secure cloud desktop through a preset visual front end and transmitting them to a preset trusted intelligent platform; Annotating the sample image using the annotation instruction via the trusted intelligent platform to obtain image annotation data; The image annotation data is stored in the distributed database through the trusted intelligent platform.

2. The method according to claim 1, characterized in that The secure execution space includes a secure database and a secure host; The storing the image training data and the sample image in a preset secure execution space, and training the initial image processing model in the secure execution space based on the image training data and the sample image to obtain a target image processing model, includes: Storing the image training data and the sample image in the secure database; the secure database is not accessible by hosts other than the secure host; extracting image training data and the sample image from the security database via the security host; The initial image processing model is trained by the secure host using the image training data and the sample image to obtain the target image processing model.

3. The method according to claim 2, characterized in that The step of training the initial image processing model using the image training data and the sample image to obtain the target image processing model includes: Obtaining image processing resources from a preset resource providing host through the security host; the security host does not have the authority to be accessed by hosts other than the resource providing host, and the resource providing host does not have the authority to be accessed by hosts other than the security host; The initial image processing model is trained by the secure host with the image processing resources, the image training data and the sample images to obtain the target image processing model.

4. The method according to claim 2, characterized in that Before receiving the model training request from the development platform, the method further includes: Receive permission application requests from the development platform; Sending the permission application request to a preset permission management terminal, which approves the permission application request and sends the approval result to the security platform terminal; wherein the approval result indicates whether the development platform terminal is authorized or not; If the approval result indicates that the development platform is authorized, the identification of the development platform is sent to the security host, and the security host marks the development platform as authorized. The security host is used to receive the model training request from the authorized development platform.

5. The method according to any one of claims 1 to 4, characterized in that After generating image training data according to the image structured data and the image storage path, and storing the image training data in a preset distributed database, the method further includes: receiving an image data query request through a preset secure cloud desktop, wherein the image data query request is sent by the annotation platform; Obtaining the sample image from the archive database according to the image data query request, extracting the image training data from the distributed database, and transmitting the sample image and the image training data to the secure cloud desktop; The sample image and the image training data are displayed through the secure cloud desktop; wherein the secure cloud desktop does not have permission for download operations.

6. The method according to claim 5, characterized in that The obtaining of the sample image from the archive database according to the image data query request, extracting the image training data from the distributed database, and transmitting the sample image and the image training data to the secure cloud desktop includes: Exporting the image data query request from the secure cloud desktop through a preset visual front end, and transmitting the image data query request to a preset trusted intelligent platform; Obtaining the sample image from the archive database and extracting the image training data from the distributed database through the trusted intelligent platform using the image data query request, and transmitting the sample image and the image training data to the visualization front end; Through the visualization front end, the sample image and the image training data are respectively desensitized, and the desensitized sample image and the desensitized image training data are transmitted to the secure cloud desktop; Among them, the secure cloud desktop does not have the authority to communicate directly with the trusted intelligent platform.

7. An image data processing device, characterized in that: Applied to a security platform, the device includes: A data acquisition module, configured to acquire a sample image and image structured data from a service platform; wherein the image structured data is obtained by the service platform identifying the sample image; An image archiving module, configured to store the sample image in a preset archiving database and generate an image storage path; a data generation module, configured to generate image training data based on the image structured data and the image storage path, and store the image training data in a preset distributed database; wherein the image training data includes the image storage path, and the security level of the distributed database is higher than the security level of the archive database; A data annotation module is configured to receive, via a preset secure cloud desktop, annotation instructions for the sample image from the annotation platform; export the annotation instructions from the secure cloud desktop via a preset visual front end and transmit them to a preset trusted intelligent platform; annotate the sample image with the annotation instructions via the trusted intelligent platform to obtain image annotation data; and store the image annotation data in the distributed database via the trusted intelligent platform; A request receiving module, configured to receive a model training request from a development platform, wherein the development platform does not have permission to access the business platform; a model training module, configured to obtain the image training data from the distributed database according to the model training request, extract the sample images from the archive database according to the image storage path in the image training data, and store the image training data and the sample images in a preset secure execution space; and train a preset initial image processing model in the secure execution space according to the image training data and the sample images to obtain a target image processing model; wherein the secure execution space is not accessible from outside the secure execution space; The image processing module is used to obtain a target image and perform image processing on the target image through the target image processing model.

8. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images