Data processing method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a data processing method and device, electronic equipment and a storage medium, first configuration information is acquired, the first configuration information is used for representing at least one image content parameter of a to-be-generated sample image, and the image content parameter is used for indicating the style of a content object in the sample image; creating a data processing task according to the first configuration information and the target image material, the data processing task being used for rendering the target image material into at least two frames of sample images having different styles of content objects based on the image content parameters; and executing the data processing task to obtain training data corresponding to the target image material, the training data including a sample image and an image tag corresponding to the sample image. The first configuration information used for controlling the image content parameters of the to-be-generated sample image is acquired, and the target image material is combined to generate the training data, so that the content of the sample image in the training data is controllable, and the data quality is improved.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of artificial intelligence technology, and in particular, to a data processing method, device, electronic device, and storage medium. Background Art

[0002] Currently, more and more artificial intelligence (AI) algorithms and models are being proposed and applied. Through trained AI models, a variety of implementation functions have been realized, greatly improving production efficiency.

[0003] Among them, the training process for AI models is a key step to ensure the quality of the model. In this process, a large amount of high-quality training data is required to ensure the training effect of the model. In the prior art, taking the model training process of image processing as an example, the training data of image processing models is usually obtained through network collection or manual drawing.

[0004] However, the training data obtained based on the solutions in the prior art have problems such as poor image quality, low image consistency, and low acquisition efficiency. Summary of the invention

[0005] The embodiments of the present disclosure provide a data processing method, device, electronic device and storage medium to overcome the problems of poor image quality, low image consistency and low acquisition efficiency in training data.

[0006] In a first aspect, an embodiment of the present disclosure provides a data processing method, including:

[0007] Acquire first configuration information, where the first configuration information is used to characterize at least one image content parameter of a sample image to be generated, where the image content parameter is used to indicate the style of a content object in the sample image; create a data processing task based on the first configuration information and a target image material, where the data processing task is used to render the target image material into at least two frames of sample images with content objects of different styles based on the image content parameter; execute the data processing task to obtain training data corresponding to the target image material, where the training data includes the sample image and an image label corresponding to the sample image.

[0008] In a second aspect, an embodiment of the present disclosure provides a data processing device, including:

[0009] A configuration module, configured to obtain first configuration information, wherein the first configuration information is used to characterize at least one image content parameter of a sample image to be generated, wherein the image content parameter is used to indicate a style of a content object in the sample image;

[0010] A creation module, configured to create a data processing task according to the first configuration information and the target image material, wherein the data processing task is configured to render the target image material into at least two frames of sample images of content objects with different styles based on the image content parameters;

[0011] An execution module is used to execute the data processing task to obtain training data corresponding to the target image material, wherein the training data includes the sample image and an image label corresponding to the sample image.

[0012] In a third aspect, an embodiment of the present disclosure provides an electronic device, including: a processor and a memory;

[0013] The memory stores computer-executable instructions;

[0014] The processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the data processing method described in the first aspect and various possible designs of the first aspect.

[0015] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, in which computer execution instructions are stored. When a processor executes the computer execution instructions, the data processing method described in the first aspect and various possible designs of the first aspect is implemented.

[0016] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including a computer program, which, when executed by a processor, implements the data processing method described in the first aspect and various possible designs of the first aspect.

[0017] The data processing method, device, electronic device and storage medium provided in this embodiment obtain first configuration information, the first configuration information is used to characterize at least one image content parameter of the sample image to be generated, and the image content parameter is used to indicate the style of the content object in the sample image; according to the first configuration information and the target image material, a data processing task is created, and the data processing task is used to render the target image material into at least two frames of sample images with content objects of different styles based on the image content parameter; the data processing task is executed to obtain training data corresponding to the target image material, and the training data includes the sample image and the image label corresponding to the sample image. By obtaining the first configuration information for controlling the image content parameter of the sample image to be generated, combining the target image material, creating a data processing task, and by executing the data processing task, the target image material is rendered into a sample image of a content object with different styles, thereby obtaining training data, and on the basis of realizing the rapid generation of training data and improving the collection efficiency of training data, the content of the sample image in the training data is controllable, so that each sample image has high content consistency and high image quality, and solves the problems of poor image quality and low image consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0019] Figure 1 An application scenario diagram of the data processing method provided by the embodiment of the present disclosure;

[0020] Figure 2 Schematic diagram of the data processing method provided in the embodiment of the present disclosure Figure 1 ;

[0021] Figure 3 for Figure 2 A flowchart of a specific implementation method of step S102 in the illustrated embodiment;

[0022] Figure 4 A schematic diagram of a training data generation process provided by an embodiment of the present disclosure;

[0023] Figure 5 Schematic diagram of the data processing method provided in the embodiment of the present disclosure Figure 2 ;

[0024] Figure 6A schematic diagram of a mapping relationship between distribution information and generated training data provided in an embodiment of the present disclosure;

[0025] Figure 7 for Figure 5 A flowchart of a specific implementation method of step S205 in the illustrated embodiment;

[0026] Figure 8 A schematic diagram of a process of converting template parameters through a parameter interpreter provided in an embodiment of the present disclosure;

[0027] Fig. 9 A structural block diagram of a data processing device provided in an embodiment of the present disclosure;

[0028] Fig.10 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure;

[0029] Fig.11 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0031] It should be noted that the user information and data involved in this disclosure (including but not limited to data used for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0032] The application scenarios of the embodiments of the present disclosure are explained below:

[0033] Figure 1 This is an application scenario diagram of the data processing method provided in the embodiment of the present disclosure. The data processing method provided in the embodiment of the present disclosure can be applied to the scenario of image generation, and more specifically, can be applied to the application scenario of generating training data for the AI ​​model to be trained. The execution subject of this embodiment can be a data server or model server running the above-mentioned data server for generating training data, or other electronic devices with similar functions, such as terminal devices. Figure 1As shown in , taking a data server as an example, a designer user uploads the drawn image material to the material asset library, and the data server obtains the image material from the material asset library. The image material includes, for example, a two-dimensional image, a three-dimensional model or other material engineering files. Afterwards, the data server responds to the execution instruction sent by the terminal device, executes the data processing method provided in this embodiment, and generates a plurality of sample images with image labels, i.e., training data, according to the image material. And the training data is sent to the data warehouse. Afterwards, the training server used to train the model obtains the training data generated by the above-mentioned data server from the data warehouse, and trains the model based on the training data. Among them, the above-mentioned material asset library and data warehouse can be deployed inside the data server or outside the data server. The data server and the training server can be located in the same server cluster or distributed in different server clusters. The deployment method of the above-mentioned server and the functional unit in the server is only exemplary, and can be set according to needs, which is not limited here.

[0034] In the prior art, taking the model training process of image processing as an example, since the cost of manually drawing training images is too high, the prior art usually obtains image samples through network collection, and then uses them as training data to train the image processing model. However, the training data collected through the network, on the one hand, has low data acquisition efficiency, and on the other hand, since the image samples collected through the network are all generated images with fixed content, it is impossible to accurately control the image content, so it is difficult to obtain a large number of image samples that meet precise requirements. At the same time, there are problems with large differences in style and style of image content, resulting in poor content consistency of image samples, which affects the effect of model training. That is, the training data obtained based on the solution in the prior art has problems such as poor image quality, low image consistency, and low acquisition efficiency.

[0035] The embodiments of the present disclosure provide a data processing method to solve the above problems.

[0036] refer to Figure 2 , Figure 2 Schematic diagram of the data processing method provided in the embodiment of the present disclosure Figure 1 The method of this embodiment can be applied in a server or a terminal device. The data processing method includes:

[0037] Step S101: Acquire first configuration information, where the first configuration information is used to characterize at least one image content parameter of a sample image to be generated, and the image content parameter is used to indicate the style of a content object in the sample image.

[0038] For example, reference Figure 1In the application scenario schematic diagram shown, the data server (hereinafter referred to as the server) can obtain the first configuration information based on the first user instruction sent by the terminal device, wherein the first characterizes at least one image content parameter of the sample image to be generated, and the image content parameter is used to indicate the style of the content object in the sample image. For example, the first configuration information includes the image content parameter para_1 and the image content parameter para_2. Specifically, the image content parameter para_1 is used to indicate the size of the "eye" in the sample image to be generated later; the image content parameter para_2 is used to indicate the illumination angle of the sample image to be generated later. Among them, the "eye" in the sample image indicated by the image content parameter para_1 and the "illumination" of the sample image indicated by the image content parameter para_1 are the content object. In addition, the content object can also include other implementation forms, such as the camera angle of the sample image, the action of the character in the sample image, etc. Exemplarily, the image content parameter includes at least one of the following: the camera angle of the sample image, the expression and / or action of the character in the sample image, the illumination angle of the sample image, and the illumination color of the sample image. In one possible implementation, all factors affecting the content of the sample image to be generated can be defined as content objects, that is, through the image content parameters represented by the first configuration information, content control of multiple dimensions of the sample image to be generated can be achieved, so that it can show different content.

[0039] Further, in a possible implementation, the first configuration information includes at least one image content parameter and a parameter value interval corresponding to the image content parameter; the parameter value interval is used to control the style variation range of the content object in the sample image. For example, the first configuration information includes an image content parameter para_1 and a parameter value interval z_1=[10,15] corresponding to the image content parameter. The image content parameter para_1 represents the camera angle corresponding to the image to be generated, and the parameter value interval z=[10,15] represents a specific camera angle value, for example, 10 to 15 degrees to the left of the front. The specific setting method of the parameter value interval can be set as needed and is not limited here. In the steps of this embodiment, the style variation range of a content object is determined by the image content parameter contained in the first configuration information and the corresponding parameter value interval, so that in the subsequent steps, based on the style variation range, a target number of sample images of content objects with different styles can be generated, for example, based on the image content parameter para_1 and the corresponding parameter value interval z_1 in the above example, 10 portrait images with camera angles between 10 and 15 degrees are generated. The target quantity may be included in the configuration value of the first configuration information, or may be a globally fixed value without restriction.

[0040] Of course, in another possible implementation, the first configuration information may not include the parameter value interval. In this case, the parameter value interval may be a preset maximum range based on the content object corresponding to the image content parameter, which will not be further described here.

[0041] Step S102: creating a data processing task according to the first configuration information and the target image material, where the data processing task is used to render the target image material into at least two frames of sample images with content objects of different styles based on image content parameters.

[0042] Exemplarily, after obtaining the first configuration information, an image material for generating training data, i.e., a target image material, is further obtained, wherein the image material is, for example, a two-dimensional image, a three-dimensional model, or other material engineering files generated based on digital content creation (DCC) software. The target image material can be specified by the user, i.e., for example, the user sends a request to the server so that the server references the target image material from the material asset library. Afterwards, the server creates a data processing task based on the target image material and in combination with the first configuration information obtained in the previous step, so that the data processing task can achieve the purpose of parameterizing the target image material, thereby generating sample images of content objects with different styles.

[0043] In one possible implementation, Figure 3 As shown, the specific implementation of step S102 includes:

[0044] Step S1021: Acquire a first task template, where the first task template is used to generate at least two frames of sample images of content objects with different styles based on image material.

[0045] Step S1022: configuring the first task template according to the first configuration information to generate a second task template.

[0046] Step S1023: Create a data processing task based on the second task template.

[0047] Exemplarily, the first task template is a programmatic encapsulation for generating at least two frames of sample images with content objects of different styles based on image materials. More specifically, the first task template can be implemented in the form of a code file. The code file corresponding to the first task template contains the code for implementing the execution logic of the above-mentioned task template. Furthermore, the first task template is a parameterized template with a parameterized interface. By calling the template interface of the first task template, the configuration of the first task template can be realized, thereby changing the specific template parameters in the first task template to achieve the purpose of controlling the content of the sample image generated by the first task template. Specifically, in this embodiment, according to the first configuration information, the template interface of the first task template is called to configure the first task template, so that the content of the first configuration information is injected into the first task template, thereby generating a second task template, so that the second task template contains the image content parameters of the style of the content object in the sample image indicated by the above-mentioned, thereby achieving the content control of the image sample finally generated.

[0048] Afterwards, based on the second task template, task creation is performed to generate data processing tasks, and then the data processing tasks are pushed into the task queue for sequential processing. In a possible implementation, the server sends the data processing tasks to the task center, and the task center executes the data processing tasks. The task center can be a functional unit deployed inside or outside the server.

[0049] Optionally, before step S1021, the method further includes:

[0050] Step S1020A: In response to a second user instruction, obtain second configuration information, where the second configuration information is used to indicate a parameterized pipeline unit developed based on a continuous integration and continuous delivery tool, and the parameterized pipeline unit is used to process image materials based on image content parameters to generate corresponding sample images.

[0051] Step S1020B: Create a first task template based on the second configuration information.

[0052] Exemplarily, in this embodiment, before calling the first task template, a step of creating a first task target is also included. Specifically, the server can obtain second configuration information by receiving a second user instruction sent by a terminal device, wherein the second configuration information is used to indicate a parameterized pipeline unit developed based on a continuous integration and continuous delivery (CI / CD) tool, wherein the parameterized pipeline unit is a functional unit designed based on a parameterized mode for image rendering, and the parameterized pipeline unit can be carried by a code file and has an open parameter interface, through which the parameter interface, the parameterized pipeline unit can be parameterized. The parameterized pipeline unit can be pre-generated and stored in the server, or outside the server (such as a material asset library), and the second configuration information obtained by the second user instruction is used to indicate a parameterized pipeline unit, and then the first task template is created by binding the parameterized pipeline unit to a blank template that does not include processing logic.

[0053] In the steps of this embodiment, a first task template is created through a second user instruction sent by the terminal device, and then the first task template is configured based on the second configuration information generated by the first user instruction to obtain a second task template, thereby completing the creation of the data processing task. In this process, the user can select the parameterized pipeline unit through the second user instruction to determine the logic for processing the target image material, so that the training data finally generated can meet the personalized needs of the user, improve the information accuracy of the training data, and improve the training effect of the model.

[0054] Step S103: Execute a data processing task to obtain training data corresponding to the target image material, where the training data includes a sample image and an image label corresponding to the sample image.

[0055] Exemplarily, further, after the server generates the data processing task, the data processing tasks in the data processing queue are processed in sequence to realize the execution of the data processing tasks generated in the above steps. Specifically, after executing the above data processing task, the server can first create a corresponding rendering project, which contains the above processing logic and image content parameters for rendering the target image material; then, rendering is performed based on the rendering project to obtain multiple frames of sample images with content objects of different styles generated based on the target image material, such as 10 frames of sample images in which the "character's mouth" in the target image material changes from small to large. At the same time, during the rendering process, the content objects in the sample images are identified and labeled, and corresponding image labels are generated, thereby obtaining multiple groups of sample images and corresponding image labels stored in pairs, that is, the training data corresponding to the target image material.

[0056] Figure 4A schematic diagram of a training data generation process provided by an embodiment of the present disclosure is shown below. Figure 4 The above-mentioned steps are further described with reference to Figure 4 As shown, first, the server receives the first user instruction sent by the terminal device, obtains the first configuration information Info_1, and obtains the target image material from the material asset library. The target image material is, for example, the three-dimensional model P1 shown in the figure; then, based on the first configuration information Info_1 and the three-dimensional model P1, the server creates a corresponding data processing task Task_1, and sends the data processing task Task_1 to the task queue located in the task center for queuing processing. After the task center executes the data processing task Task_1, it generates multiple sample images, such as the image P1_1, image P1_2, image P1_3 (shown as P1_1, P1_2, P1_3) shown in the figure, and corresponding image tags, such as the tag tag_1, tag tag_2, tag tag_3 (shown as tag_1, tag_2, tag_3) shown in the figure. Among them, tag tag_1 = {mouth size: 1}, tag_2 = {mouth size: 2}; tag_3 = {mouth size: 3}. The training data is then returned to the server, thereby obtaining corresponding training data generated based on the target image material.

[0057] In this embodiment, by acquiring first configuration information, the first configuration information is used to characterize at least one image content parameter of the sample image to be generated, and the image content parameter is used to indicate the style of the content object in the sample image; according to the first configuration information and the target image material, a data processing task is created, and the data processing task is used to render the target image material into at least two frames of sample images with content objects of different styles based on the image content parameter; the data processing task is executed to obtain training data corresponding to the target image material, and the training data includes the sample image and the image label corresponding to the sample image. By acquiring first configuration information for controlling the image content parameter of the sample image to be generated, a data processing task is created in combination with the target image material, and by executing the data processing task, the target image material is rendered into sample images of content objects of different styles, thereby obtaining training data, and on the basis of realizing rapid generation of training data and improving the collection efficiency of training data, the content of the sample images in the training data is controllable, so that each sample image has high content consistency and high image quality, and solves the problems of poor image quality and low image consistency.

[0058] refer to Figure 5 , Figure 5 Schematic diagram of the data processing method provided in the embodiment of the present disclosure Figure 2 In this embodiment Figure 2On the basis of the illustrated embodiment, step S101 and step S103 are further refined, and the data processing method includes:

[0059] Step S201: receiving a first user instruction, where the first user instruction is used to indicate at least one target image content parameter for generating a sample image.

[0060] Step S202: obtaining distribution information corresponding to the target image content parameter, where the distribution information is used to characterize distribution characteristics of parameter values ​​corresponding to the target image content parameter.

[0061] Step S203: Generate first configuration information according to the target image content parameters and distribution information.

[0062] Exemplarily, after receiving the first user instruction sent by the terminal device, the server determines the target image content parameters indicated by the first user instruction, such as the camera angle, the expression of the person in the sample image, etc. After that, the distribution information corresponding to the above target image content parameters is obtained, and the distribution information refers to the distribution characteristics of the parameter values ​​corresponding to the target image content parameters. Based on the above example, for example, the distribution of angle values ​​corresponding to the camera angle, the style distribution of the expression of the person in the sample image. In the application scenario of generating training data for the AI ​​model to be trained, in order to improve the training effect of the AI ​​model, the style of the content object in the sample image used for training is more densely concentrated in the "common area", such as "front-facing face of the person", "smiling face of the person", etc., and a larger number of sample images are generated; while in the case of "side and rear-facing face of the person", a smaller number of sample images are generated, thereby improving the data quality of the training data. After that, according to the target image content parameters and the distribution information, the first configuration information is generated, so that the distribution of the parameter values ​​corresponding to the target image content parameters represented by the first configuration information can be consistent with the distribution information, thereby realizing the control of the number of sample images with different styles in the training data generated subsequently.

[0063] Figure 6 A schematic diagram of a mapping relationship between distribution information and generated training data provided in an embodiment of the present disclosure, referring to Figure 6As shown, the first configuration information is generated by the distribution information dis_1 and the image content parameter para_1, wherein the content object indicated by the image content parameter para_1 is "the size of the person's eyes", and the corresponding size value range is [10,20] (preset unit). The distribution information dis_1 represents the distribution of the above size value within the value range, for example, it satisfies the Gaussian distribution centered on the size value = 15. Afterwards, based on the first configuration information and the target image material, a data processing task is created and executed, and the obtained training data Data_1 containing multiple (for example, 10) sample images is obtained. In the training data Data_1, the "size of the person's eyes" displayed by each sample image meets the distribution characteristics represented by the above distribution information dis_1, so that the size of the person's eyes in most (for example, as shown in the figure, N_1 = 6) sample images in the above training data are concentrated in [p1, p2]. More specifically, for example, p1 = 13, p2 = 17, that is, the size of the person's eyes in 60% of the sample images is between 13 and 17. The sample images with the object eye size greater than 17 or the object eye size less than 13 only account for a minority of the training data (for example, as shown in the figure, N_2=2, N_3=2).

[0064] Further, exemplarily, the distribution information may be determined based on a user instruction sent by a terminal device, or may be a fixed setting. The distribution characteristics represented by the distribution information include at least a Gaussian distribution. Of course, the distribution information may also represent other types of distributions, which may be specifically set based on model training needs and user requirements. In this embodiment, by acquiring and configuring the distribution information corresponding to the target image content parameters, and generating the corresponding first configuration information, the sample images of content objects with different styles in the subsequent training data generated based on the first configuration information meet the restrictions of specific distribution characteristics, thereby improving the data quality of the training data and the efficiency of model training.

[0065] Step S204: creating a data processing task according to the first configuration information and the target image material, where the data processing task is used to render the target image material into at least two frames of sample images with content objects of different styles based on image content parameters.

[0066] Step S205: In response to the execution of the data processing task, a rendering project is created.

[0067] Exemplarily, after the data processing task is created, it will be added to the task queue and wait for execution, which will not be described in detail here. Afterwards, after the data processing task is executed, in response to the execution of the data processing task, the task logic in the data processing task is called to create a rendering project, wherein the rendering project is a project item for image rendering and generating output images, which contains the execution logic corresponding to the processing link for image processing, and the corresponding control parameters, i.e., rendering parameters. The implementation method of creating a rendering project is, for example, to realize the creation of a rendering project by calling the functional interface (API) provided by the digital content creation software for creating the project. The specific implementation method is determined based on the functional interface of the digital content creation software, and will not be described one by one here.

[0068] Further, illustratively, if Figure 7 As shown, the specific implementation of step S205 includes:

[0069] Step S2051: Obtain the task template corresponding to the data processing task.

[0070] Step S2052: calling the parameter interpreter corresponding to the task template, converting the template parameters in the task template, and obtaining rendering parameters.

[0071] Step S2053: calling the engineering interface of the digital content creation software based on the rendering parameters to create a rendering project.

[0072] Exemplarily, the data processing task includes a task template for generating at least two frames of sample images of content objects with different styles based on image materials, i.e., the second task template in the above-mentioned embodiment steps, or template information for obtaining the task template. After the task template corresponding to the data processing task activity is activated, the corresponding parameter interpreter is called from the task template to convert the template parameters in the task template. Among them, the task template is generated based on the parameterized pipeline unit, which includes template parameters for controlling the data processing logic corresponding to the task template. The data processing task needs to be executed based on the target digital content creation software, so the template parameters need to be converted accordingly to obtain parameters that can be used to render images using the target digital content creation software, i.e., rendering parameters.

[0073] Figure 8 A schematic diagram of a process of converting template parameters through a parameter interpreter provided in an embodiment of the present disclosure, such as Figure 8As shown, exemplarily, after the server executes the data processing task, it first obtains the template parameters in the task template based on the data processing task. The template parameters include parameters for the "camera angle", such as {roll=70; yaw=12; pitch=65} as shown in the figure. Among them, roll, yaw, and pitch are roll angle, yaw angle, and pitch angle, respectively. Afterwards, the server converts the template parameters by calling the parameter interpreter corresponding to the task template to obtain rendering parameters that match the target digital content creation software sw_1, such as in the figure, including:

[0074] {setAttr"my_camera.rotateX"65;

[0075] setAttr"my_camera.rotateY"12;

[0076] setAttr"my_camera.rotateZ"70}.

[0077] Based on the above example, the role of the parameter interpreter is equivalent to converting the template parameters in a custom format generated based on the user's instructions into instruction parameters that can be recognized and executed by a specific digital content creation software, and then using the digital content creation software to render the target image material based on the style of the content object represented by the template parameters to obtain a sample image. Among them, the specific implementation form of the rendering parameters is related to the target digital content creation software, so the above rendering parameters are only examples and will not be repeated here one by one.

[0078] Furthermore, in a possible implementation, after obtaining the rendering parameters, the engineering interface of the corresponding target digital content creation software can be called based on the rendering parameters to create a rendering project. The rendering project will be directly submitted to the server cluster for distributed rendering. The server cluster performs parameterized rendering based on the rendering parameters contained in the rendering project and the target image material, thereby generating multiple sample images of content objects with different styles. Afterwards, the sample images are further labeled through other services to obtain image labels corresponding to the sample images, thereby obtaining training data.

[0079] Of course, in another possible implementation, after step S205, the following is further included:

[0080] Step S206: calling the parameter interpreter corresponding to the task template to obtain the labeling rule information in the task template, where the labeling rule information is used to generate an image label corresponding to the sample image.

[0081] Step S207: Submit the rendering project and labeling rule information to the server cluster for distributed rendering to obtain training data.

[0082] Exemplarily, in another possible implementation, a parameter interpreter may be further used to obtain the labeling rule information in the task template, and the content object in the sample image is labeled based on the labeling rule information to generate an image label corresponding to the sample image. The labeling rule information may be a program code for describing the labeling rule, or identification information indicating the content object, which is not specifically limited here. Afterwards, the rendering project and the labeling rule information are submitted to the server cluster as a whole for distributed rendering. The server cluster performs parameterized rendering based on the rendering parameters contained in the rendering project and the target image material, thereby generating multiple sample images with content objects of different styles. In the process, different content objects in the sample image are labeled, such as the camera angle of the sample image, the expression and / or action of the person in the sample image, the illumination angle of the sample image, the illumination color of the sample image, etc., that is, the content represented by the image content parameter, so as to obtain the image label corresponding to the sample image. The server cluster may also be referred to as a "rendering farm", which is a device dedicated to image rendering, and can achieve the beneficial effect of improving image rendering efficiency. Of course, in other possible implementations, the server may directly perform rendering to obtain training data including sample images and image labels corresponding to the sample images.

[0083] Optionally, before step S2051, the method further includes:

[0084] Step S2050A: Obtain a packaged image of the target digital content creation software corresponding to the task template.

[0085] Step S2050B: construct a parameter interpreter corresponding to the task template according to the packaged image.

[0086] Furthermore, before using the parameter interpreter to perform the above-mentioned parameter conversion, it also includes obtaining a packaged image of the target digital content creation software corresponding to the task template, and using the package to construct a parameter interpreter corresponding to the task template, so that the obtained parameter interpreter can convert the template parameters into instruction parameters of a specific digital content creation software (i.e., the target digital content creation software), and then perform image rendering based on the specific digital content creation software.

[0087] Optionally, the method of this embodiment further includes:

[0088] Step S208: Display a task list, which at least includes the task execution status of the data processing task.

[0089] Exemplarily, when the executor of this embodiment is a server, the server sends task information corresponding to the task queue to the terminal device to display a task list, and the task list includes at least the task execution status of the data processing task. That is, after the server creates the data processing task, the task list is displayed through the terminal device communicating with it, so as to show the task execution status of the data processing task created above, wherein the task execution status includes, for example, information such as "processing", "completed", "not started" or the percentage of task execution.

[0090] In another possible situation, when the execution subject of this embodiment is a terminal device, the task execution status of the data processing task created above is directly displayed in the display interface of the terminal device. The specific implementation method is similar and will not be repeated here.

[0091] Step S209: In response to a trigger instruction for the data processing task, display at least one frame of sample image in the training data.

[0092] Exemplarily, when the execution subject of this embodiment is a server, after generating the training data, the server returns the prompt information that the training data has been generated to the terminal device for display. After that, the user sends a trigger instruction to the server by operating the terminal device. Based on the trigger instruction, the server sends at least one frame of sample image corresponding to the trigger instruction to the terminal device for display, thereby realizing the preview of the generated sample image on the terminal device side; after that, the user can further adjust the first configuration information according to the preview of the sample image to change the content of the generated sample image, thereby further improving and optimizing the quality of the training data. Finally, after the user has finished the preview and determined that the training data does not need to be further modified, the user can send the corresponding generation instruction to the server through the terminal device. In response to the generation instruction, the server sends the training data to the data warehouse for model training.

[0093] In another possible situation, when the executor of this embodiment is a terminal device, the terminal device responds to a trigger instruction for a data processing task once and displays at least one frame of sample image in the training data in the display interface. The specific implementation method is similar and will not be repeated here.

[0094] In this embodiment, the implementation method of step S204 is the same as that of the present disclosure. Figure 2 The implementation method of step S102 in the illustrated embodiment is the same and will not be described in detail here.

[0095] Corresponding to the data processing method of the above embodiment, Fig. 9 This is a structural block diagram of a data processing device provided by an embodiment of the present disclosure. For ease of explanation, only the parts related to the embodiment of the present disclosure are shown.

[0096] Reference Fig. 9 , the data processing device 3 comprises:

[0097] A configuration module 31 is used to obtain first configuration information, where the first configuration information is used to characterize at least one image content parameter of a sample image to be generated, where the image content parameter is used to indicate a style of a content object in the sample image;

[0098] A creation module 32, configured to create a data processing task according to the first configuration information and the target image material, wherein the data processing task is configured to render the target image material into at least two frames of sample images of content objects with different styles based on image content parameters;

[0099] The execution module 33 is used to execute the data processing task to obtain the training data corresponding to the target image material, and the training data includes the sample image and the image label corresponding to the sample image.

[0100] In one embodiment of the present disclosure, the configuration module 31 is specifically used to: receive a first user instruction, the first user instruction is used to indicate at least one target image content parameter for generating a sample image; obtain distribution information corresponding to the target image content parameter, the distribution information is used to characterize the distribution characteristics of the parameter values ​​corresponding to the target image content parameter; generate first configuration information according to the target image content parameter and the distribution information.

[0101] In one embodiment of the present disclosure, the distribution characteristics represented by the distribution information at least include Gaussian distribution.

[0102] In one embodiment of the present disclosure, the first configuration information includes at least one image content parameter and a parameter value interval corresponding to the image content parameter; the parameter value interval is used to control the style variation range of the content object in the sample image.

[0103] In one embodiment of the present disclosure, the image content parameters include at least one of the following: a camera angle of the sample image, expressions and / or actions of characters in the sample image, a lighting angle of the sample image, and a lighting color of the sample image.

[0104] In one embodiment of the present disclosure, the execution module 33 is specifically used to: create a rendering project in response to the execution of the data processing task; submit the rendering project to the server cluster for distributed rendering to obtain training data.

[0105] In one embodiment of the present disclosure, when creating a rendering project in response to the execution of a data processing task, the execution module 33 is specifically used to: obtain a task template corresponding to the data processing task; call a parameter interpreter corresponding to the task template to convert the template parameters in the task template to obtain rendering parameters; and call the engineering interface of the digital content creation software based on the rendering parameters to create a rendering project.

[0106] In one embodiment of the present disclosure, the execution module 33 is further used to: call the parameter interpreter corresponding to the task template to obtain the labeling rule information in the task template, and the labeling rule information is used to generate the image label corresponding to the sample image; when the execution module 33 submits the rendering project to the server cluster for distributed rendering to obtain training data, it is specifically used to: submit the rendering project and the labeling rule information to the server cluster for distributed rendering to obtain training data.

[0107] In one embodiment of the present disclosure, the execution module 33 is further used to: obtain a packaged image of the target digital content creation software corresponding to the task template; and construct a parameter interpreter corresponding to the task template according to the packaged image.

[0108] In one embodiment of the present disclosure, a creation module 32 is specifically used to: obtain a first task template, the first task template is used to generate at least two frames of sample images of content objects with different styles based on image material; configure the first task template according to first configuration information to generate a second task template; and create a data processing task based on the second task template.

[0109] In one embodiment of the present disclosure, before obtaining the first task template, the configuration module 31 is further used to: in response to a second user instruction, obtain second configuration information, the second configuration information is used to indicate a parameterized pipeline unit developed based on continuous integration and continuous delivery tools, the parameterized pipeline unit is used to process image materials based on image content parameters to generate corresponding sample images; based on the second configuration information, create the first task template.

[0110] In one embodiment of the present disclosure, the configuration module 31 is also used for at least one of the following: displaying a task list, the task list at least including the task execution status of the data processing task; and displaying at least one frame of sample image in the training data in response to a trigger instruction for the data processing task.

[0111] The configuration module 31, the creation module 32 and the execution module 33 are connected in sequence. The data processing device 3 provided in this embodiment can execute the technical solution of the above method embodiment, and its implementation principle and technical effect are similar, which will not be described in detail in this embodiment.

[0112] Fig.10 A schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure is shown in FIG. Fig.10 As shown, the electronic device 4 includes:

[0113] A processor 41, and a memory 42 communicatively connected to the processor 41;

[0114] The memory 42 stores computer executable instructions;

[0115] The processor 41 executes the computer execution instructions stored in the memory 42 to implement the following Figure 2-Figure 8 The data processing method in the illustrated embodiment.

[0116] Optionally, the processor 41 and the memory 42 are connected via a bus 43 .

[0117] For related instructions, please refer to Figure 2-Figure 8 The relevant descriptions and effects corresponding to the steps in the corresponding embodiments can be understood, and no further elaboration is made here.

[0118] The present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, which are used to implement the present invention when executed by a processor. Figure 2-Figure 8 The data processing method provided in any one of the corresponding embodiments.

[0119] In order to implement the above embodiment, the embodiment of the present disclosure also provides an electronic device.

[0120] refer to Fig.11 , which shows a schematic diagram of the structure of an electronic device 900 suitable for implementing the embodiment of the present disclosure, and the electronic device 900 may be a terminal device or a server. The terminal device may include but is not limited to mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, personal digital assistants (PDAs), tablet computers (Portable Android Devices, PADs), portable multimedia players (PMPs), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig.11 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0121] like Fig.11As shown, the electronic device 900 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 902 or a program loaded from a storage device 908 to a random access memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the electronic device 900 are also stored. The processing device 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0122] Typically, the following devices may be connected to the I / O interface 905: input devices 906 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 907 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 908 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 909. The communication device 909 may allow the electronic device 900 to communicate with other devices wirelessly or by wire to exchange data. Although Fig.11 The electronic device 900 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0123] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 909, or installed from a storage device 908, or installed from a ROM 902. When the computer program is executed by the processing device 901, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0124] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0125] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0126] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device executes the method shown in the above embodiment.

[0127] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0128] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0129] The units involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a unit does not limit the unit itself in some cases. For example, the first acquisition unit may also be described as a "unit for acquiring at least two Internet Protocol addresses".

[0130] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0131] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0132] In a first aspect, according to one or more embodiments of the present disclosure, a data processing method is provided, comprising:

[0133] Acquire first configuration information, where the first configuration information is used to characterize at least one image content parameter of a sample image to be generated, where the image content parameter is used to indicate the style of a content object in the sample image; create a data processing task based on the first configuration information and a target image material, where the data processing task is used to render the target image material into at least two frames of sample images with content objects of different styles based on the image content parameter; execute the data processing task to obtain training data corresponding to the target image material, where the training data includes the sample image and an image label corresponding to the sample image.

[0134] According to one or more embodiments of the present disclosure, obtaining the first configuration information includes: receiving a first user instruction, the first user instruction is used to indicate at least one target image content parameter for generating the sample image; obtaining distribution information corresponding to the target image content parameter, the distribution information is used to characterize distribution characteristics of parameter values ​​corresponding to the target image content parameter; generating the first configuration information according to the target image content parameter and the distribution information.

[0135] According to one or more embodiments of the present disclosure, the distribution characteristics represented by the distribution information at least include Gaussian distribution.

[0136] According to one or more embodiments of the present disclosure, the first configuration information includes at least one image content parameter and a parameter value interval corresponding to the image content parameter; the parameter value interval is used to control the style variation range of the content object in the sample image.

[0137] According to one or more embodiments of the present disclosure, the image content parameters include at least one of the following: a camera angle of the sample image, expressions and / or actions of characters in the sample image, a lighting angle of the sample image, and a lighting color of the sample image.

[0138] According to one or more embodiments of the present disclosure, the data processing task is executed to obtain the training data corresponding to the target image material, including: creating a rendering project in response to the execution of the data processing task; submitting the rendering project to a server cluster for distributed rendering to obtain the training data.

[0139] According to one or more embodiments of the present disclosure, creating a rendering project in response to the execution of the data processing task includes: obtaining a task template corresponding to the data processing task; calling a parameter interpreter corresponding to the task template to convert the template parameters in the task template to obtain rendering parameters; and calling an engineering interface of a digital content creation software based on the rendering parameters to create a rendering project.

[0140] According to one or more embodiments of the present disclosure, the method also includes: calling a parameter interpreter corresponding to the task template to obtain labeling rule information in the task template, wherein the labeling rule information is used to generate an image label corresponding to the sample image; submitting the rendering project to a server cluster for distributed rendering to obtain the training data includes: submitting the rendering project and the labeling rule information to a server cluster for distributed rendering to obtain the training data.

[0141] According to one or more embodiments of the present disclosure, the method further includes: obtaining a packaged image of the target digital content creation software corresponding to the task template; and constructing a parameter interpreter corresponding to the task template based on the packaged image.

[0142] According to one or more embodiments of the present disclosure, creating a data processing task based on the first configuration information and the target image material includes: obtaining a first task template, the first task template being used to generate at least two frames of sample images of content objects with different styles based on the image material; configuring the first task template according to the first configuration information to generate a second task template; and creating a data processing task based on the second task template.

[0143] According to one or more embodiments of the present disclosure, before obtaining the first task template, it also includes: in response to a second user instruction, obtaining second configuration information, the second configuration information is used to indicate a parameterized pipeline unit developed based on a continuous integration and continuous delivery tool, the parameterized pipeline unit is used to process image materials based on image content parameters to generate corresponding sample images; based on the second configuration information, creating the first task template.

[0144] According to one or more embodiments of the present disclosure, the method further includes at least one of the following: displaying a task list, wherein the task list includes at least a task execution status of the data processing task; and displaying at least one frame of sample image in the training data in response to a trigger instruction for the data processing task.

[0145] In a second aspect, according to one or more embodiments of the present disclosure, there is provided a data processing device, comprising:

[0146] A configuration module, configured to obtain first configuration information, wherein the first configuration information is used to characterize at least one image content parameter of a sample image to be generated, wherein the image content parameter is used to indicate a style of a content object in the sample image;

[0147] A creation module, configured to create a data processing task according to the first configuration information and the target image material, wherein the data processing task is configured to render the target image material into at least two frames of sample images of content objects with different styles based on the image content parameters;

[0148] An execution module is used to execute the data processing task to obtain training data corresponding to the target image material, wherein the training data includes the sample image and an image label corresponding to the sample image.

[0149] According to one or more embodiments of the present disclosure, the configuration module is specifically used to: receive a first user instruction, where the first user instruction is used to indicate at least one target image content parameter for generating the sample image; obtain distribution information corresponding to the target image content parameter, where the distribution information is used to characterize distribution characteristics of parameter values ​​corresponding to the target image content parameter; and generate the first configuration information based on the target image content parameter and the distribution information.

[0150] According to one or more embodiments of the present disclosure, the distribution characteristics represented by the distribution information at least include Gaussian distribution.

[0151] According to one or more embodiments of the present disclosure, the first configuration information includes at least one image content parameter and a parameter value interval corresponding to the image content parameter; the parameter value interval is used to control the style variation range of the content object in the sample image.

[0152] According to one or more embodiments of the present disclosure, the image content parameters include at least one of the following: a camera angle of the sample image, expressions and / or actions of characters in the sample image, a lighting angle of the sample image, and a lighting color of the sample image.

[0153] According to one or more embodiments of the present disclosure, the execution module is specifically used to: create a rendering project in response to the execution of the data processing task; submit the rendering project to a server cluster for distributed rendering to obtain the training data.

[0154] According to one or more embodiments of the present disclosure, when the execution module creates a rendering project in response to the execution of the data processing task, it is specifically used to: obtain a task template corresponding to the data processing task; call a parameter interpreter corresponding to the task template to convert the template parameters in the task template to obtain rendering parameters; and call the engineering interface of the digital content creation software based on the rendering parameters to create a rendering project.

[0155] According to one or more embodiments of the present disclosure, the execution module is further used to: call the parameter interpreter corresponding to the task template to obtain the labeling rule information in the task template, and the labeling rule information is used to generate the image label corresponding to the sample image; when the execution module submits the rendering project to the server cluster for distributed rendering to obtain the training data, the execution module is specifically used to: submit the rendering project and the labeling rule information to the server cluster for distributed rendering to obtain the training data.

[0156] According to one or more embodiments of the present disclosure, the execution module is further used to: obtain a packaged image of the target digital content creation software corresponding to the task template; and construct a parameter interpreter corresponding to the task template based on the packaged image.

[0157] According to one or more embodiments of the present disclosure, the creation module is specifically used to: obtain a first task template, the first task template is used to generate at least two frames of sample images of content objects with different styles based on image material; configure the first task template according to the first configuration information to generate a second task template; and create a data processing task based on the second task template.

[0158] According to one or more embodiments of the present disclosure, before obtaining the first task template, the configuration module 31 is further used to: obtain second configuration information in response to a second user instruction, wherein the second configuration information is used to indicate a parameterized pipeline unit developed based on a continuous integration and continuous delivery tool, and the parameterized pipeline unit is used to process image materials based on image content parameters to generate corresponding sample images; and create the first task template based on the second configuration information.

[0159] According to one or more embodiments of the present disclosure, the configuration module is also used for at least one of the following: displaying a task list, wherein the task list includes at least the task execution status of the data processing task; and displaying at least one frame of sample image in the training data in response to a trigger instruction for the data processing task.

[0160] In a third aspect, according to one or more embodiments of the present disclosure, there is provided an electronic device, comprising: at least one processor and a memory;

[0161] The memory stores computer-executable instructions;

[0162] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the data processing method described in the first aspect and various possible designs of the first aspect.

[0163] In a fourth aspect, according to one or more embodiments of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer execution instructions. When a processor executes the computer execution instructions, the data processing method described in the first aspect and various possible designs of the first aspect is implemented.

[0164] In a fifth aspect, according to one or more embodiments of the present disclosure, a computer program product is provided, including a computer program, which, when executed by a processor, implements the data processing method described in the first aspect and various possible designs of the first aspect.

[0165] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0166] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0167] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

Claims

1. A data processing method, characterized in that: include: Acquire first configuration information, where the first configuration information is used to characterize at least one image content parameter of a sample image to be generated, where the image content parameter is used to indicate a style of a content object in the sample image; Creating a data processing task according to the first configuration information and the target image material, wherein the data processing task is used to render the target image material into at least two frames of sample images of content objects with different styles based on the image content parameters; The data processing task is executed to obtain training data corresponding to the target image material, wherein the training data includes the sample image and an image label corresponding to the sample image.

2. The method according to claim 1, characterized in that The obtaining of the first configuration information includes: receiving a first user instruction, wherein the first user instruction is used to indicate at least one target image content parameter for generating the sample image; Acquire distribution information corresponding to the target image content parameter, where the distribution information is used to characterize distribution characteristics of parameter values ​​corresponding to the target image content parameter; The first configuration information is generated according to the target image content parameter and the distribution information.

3. The method according to claim 2, characterized in that The distribution characteristics represented by the distribution information at least include Gaussian distribution.

4. The method according to claim 1, characterized in that: The first configuration information includes at least one image content parameter and a parameter value interval corresponding to the image content parameter; the parameter value interval is used to control the style variation range of the content object in the sample image.

5. The method according to claim 1, characterized in that The image content parameters include at least one of the following: The camera angle of the sample image, the expression and / or action of the person in the sample image, the lighting angle of the sample image, and the lighting color of the sample image.

6. The method according to claim 1, characterized in that Executing the data processing task to obtain training data corresponding to the target image material includes: In response to the execution of the data processing task, creating a rendering project; The rendering project is submitted to a server cluster for distributed rendering to obtain the training data.

7. The method according to claim 6, characterized in that The creating a rendering project in response to the execution of the data processing task includes: Get the task template corresponding to the data processing task; Calling a parameter interpreter corresponding to the task template to convert the template parameters in the task template to obtain rendering parameters; The engineering interface of the digital content creation software is called based on the rendering parameters to create a rendering project.

8. The method according to claim 7, characterized in that The method further comprises: Calling a parameter interpreter corresponding to the task template to obtain labeling rule information in the task template, where the labeling rule information is used to generate an image label corresponding to the sample image; Submitting the rendering project to a server cluster for distributed rendering to obtain the training data includes: The rendering project and the labeling rule information are submitted to a server cluster for distributed rendering to obtain the training data.

9. The method according to claim 7, characterized in that: The method further comprises: Obtaining a packaged image of a target digital content creation software corresponding to the task template; According to the packaged image, a parameter interpreter corresponding to the task template is constructed.

10. The method according to claim 1, characterized in that The step of creating a data processing task according to the first configuration information and the target image material includes: Acquire a first task template, where the first task template is used to generate at least two frames of sample images of content objects with different styles based on image material; Configure the first task template according to the first configuration information to generate a second task template; A data processing task is created based on the second task template.

11. The method according to claim 10, characterized in that Before obtaining the first task template, the method further includes: In response to a second user instruction, second configuration information is obtained, where the second configuration information is used to indicate a parameterized pipeline unit developed based on a continuous integration and continuous delivery tool, where the parameterized pipeline unit is used to process an image material based on an image content parameter to generate a corresponding sample image; Based on the second configuration information, the first task template is created.

12. The method according to claim 1, characterized in that The method further comprises at least one of the following: Display a task list, wherein the task list at least includes the task execution status of the data processing task; In response to a trigger instruction for the data processing task, at least one frame of sample image in the training data is displayed.

13. A data processing device, characterized in that: include: A configuration module, configured to obtain first configuration information, wherein the first configuration information is used to characterize at least one image content parameter of a sample image to be generated, wherein the image content parameter is used to indicate a style of a content object in the sample image; A creation module, configured to create a data processing task according to the first configuration information and the target image material, wherein the data processing task is configured to render the target image material into at least two frames of sample images of content objects with different styles based on the image content parameters; An execution module is used to execute the data processing task to obtain training data corresponding to the target image material, wherein the training data includes the sample image and an image label corresponding to the sample image.

14. An electronic device, characterized in that: include: Processor and memory; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the data processing method according to any one of claims 1 to 12.

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the processor executes the computer-executable instructions, the data processing method according to any one of claims 1 to 12 is implemented.

Images