Method, device and equipment for generating a set of stickers, and storage medium

Abstract

The present disclosure relates to a method, device and equipment for generating a picture-in-picture sticker, and a storage medium. The method comprises: obtaining sticker description text input by a target object; analyzing the sticker description text to extract a plurality of key semantic information; dividing the plurality of key semantic information into at least two levels of respective corresponding hierarchical semantic information according to visual effects; converting the hierarchical semantic information of each level into description information for generating an image, and obtaining structured image description information based on the respective corresponding description information of the at least two levels; and generating a picture-in-picture sticker according to the structured image description information. The present disclosure realizes rich visual levels in the constructed sticker, and realizes the diversification of the picture-in-picture sticker.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to a method, apparatus, device and storage medium for generating image stickers. Background Technology

[0002] With the rapid development of generative artificial intelligence (AIGC) technology, the technology of generating images based on natural language is gradually being widely used in content creation, digital design, and sticker production. Especially in mobile content creation scenarios, users' demand for personalized, themed sets, and style-consistent visual sticker resources is constantly growing.

[0003] In related technologies, there are processes for generating stickers from image or text input, but the resulting sticker sets have a limited visual appeal. Summary of the Invention

[0004] This disclosure provides a method, apparatus, device, and storage medium for generating image set stickers, to at least solve the problem of monotonous visual representation in related technologies. The technical solution of this disclosure is as follows: According to a first aspect of the present disclosure, a method for generating image set stickers is provided, comprising: Obtain the sticker description text input by the target object; the sticker description text is the description information of the sticker set to be generated; The sticker description text is parsed to extract multiple key semantic information; and the multiple key semantic information is divided into at least two levels of hierarchical semantic information according to their visual function. The hierarchical semantic information of each level is transformed into descriptive information for generating images, and structured image description information is obtained based on the descriptive information corresponding to each of the at least two levels. The set of stickers is generated based on the structured image description information.

[0005] In one exemplary embodiment, generating the set of stickers based on the structured image description information includes: The structured image description information is converted into resource description information according to a preset resource description format; Based on the resource description information, a set of stickers generation task is constructed, and the set of stickers is generated according to the set of stickers generation task.

[0006] In one exemplary embodiment, dividing the plurality of key semantic information into at least two levels of hierarchical semantic information according to their visual function includes: Based on the visual function corresponding to each key semantic information, the main content representing the core theme and key details of the image is extracted from the multiple key semantic information to obtain the first level of semantic information; Secondary content representing the image background and / or supplementary image information is extracted from the multiple key semantic information to obtain second-level semantic information; The decorative elements representing the decorative elements of the image are extracted from the multiple key semantic information to obtain the third level of semantic information.

[0007] In one exemplary embodiment, the step of converting the hierarchical semantic information of each level into descriptive information for generating an image, and obtaining structured image descriptive information based on the descriptive information corresponding to each of the at least two levels, includes: The first-level semantic information is transformed into first-structured descriptive information according to a preset sentence framework; The second-level semantic information is transformed into second-structured descriptive information according to the preset sentence framework; The third-level semantic information is transformed into third-structured descriptive information according to the preset sentence framework; The structured image description information is determined based on the first structured description information, the second structured description information, and the third structured description information.

[0008] In one exemplary embodiment, converting the structured image description information into resource description information according to a preset resource description format includes: The structured image description information is verified according to a preset verification strategy to obtain the verification result; Based on the verification result, the structured image description information is completed according to the standardized fields to obtain the completed description information; The completed description information is converted into resource description information according to a preset resource description format.

[0009] In one exemplary embodiment, the resource description information includes description information corresponding to multiple levels of semantic information, and the method further includes: Generate a resource identifier corresponding to each description information according to the hierarchical semantic information corresponding to each description information in the resource description information; The resource list stores the correspondence between the resource descriptors and each descriptive information in the resource description information.

[0010] In one exemplary embodiment, the step of constructing a sticker set generation task based on the resource description information and generating the sticker set according to the sticker set generation task includes: The resource description information is parsed to obtain the first resource description information corresponding to the first level semantic information, the second resource description information corresponding to the second level semantic information, and the third resource description information corresponding to the third level semantic information; A first sticker generation task is generated based on the first resource description information, and a first-level sticker is generated based on the first sticker generation task. A second sticker generation task is generated based on the second resource description information, and a second-level sticker is generated based on the second sticker generation task. A third sticker generation task is generated based on the third resource description information, and a third-level sticker is generated based on the third sticker generation task. The first layer sticker, the second layer sticker, and the third layer sticker are merged to generate the set of stickers.

[0011] In one exemplary embodiment, the step of constructing a sticker set generation task based on the resource description information and generating the sticker set according to the sticker set generation task includes: Based on the resource description information, a generation task for each sticker image in the sticker set is constructed, and the sticker set generation task is constructed according to the generation task for each sticker image. Based on the generation task of each sticker image in the sticker set generation task and the text image model, generate the sticker image corresponding to each generation task. The set of stickers is constructed based on the sticker image corresponding to each generated task.

[0012] In one exemplary embodiment, constructing the set of stickers based on the sticker image corresponding to each generated task includes: Each sticker image is input into at least two stylization models for stylization processing, resulting in multiple stylized images corresponding to each sticker image; The multiple stylized images corresponding to each sticker image are combined to obtain the set of stickers.

[0013] In one exemplary embodiment, combining multiple stylized images corresponding to each sticker image to obtain the sticker set includes: Combine multiple stylized images corresponding to each sticker image to obtain the initial set of images; Perform background segmentation on each image in the initial set of images, remove the original background from each image, and generate sticker materials with transparent backgrounds for each image; Multiple sticker materials are subjected to sticker enhancement processing to obtain multiple enhanced stickers; the sticker enhancement processing includes adding at least one of stroke, shadow, and highlight; The set of stickers is constructed based on the multiple enhanced stickers.

[0014] In one exemplary embodiment, after constructing the set of stickers based on the plurality of enhanced stickers, the method further includes: Upload the set of stickers to the resource storage system and generate the access path for the set of stickers; Based on the access path, generation time, and version number of the image set stickers, the resource description information is updated to obtain updated resource description information.

[0015] In one exemplary embodiment, after constructing the set of stickers based on the plurality of enhanced stickers, the method further includes: Obtain the storage path, generation parameters, and stylization information of the sticker set; Based on the storage path, generation parameters, and stylization information of the image set stickers, generate image set description information; The correspondence between the sticker set and the description information of the sticker set is stored in a preset database.

[0016] In one exemplary embodiment, the method further includes: Obtain the image of the object to be processed; At least one sticker is selected from the set of stickers, and the at least one sticker is fused with the object image to generate a stylized object image.

[0017] According to a second aspect of the present disclosure, an apparatus for generating image set stickers is provided, comprising: The description text acquisition module is configured to acquire the sticker description text input by the target object; the sticker description text is the description information of the set of stickers to be generated; The semantic extraction module is configured to parse the sticker description text, extract multiple key semantic information, and divide the multiple key semantic information into at least two levels of hierarchical semantic information according to their visual function. The structured information acquisition module is configured to convert the hierarchical semantic information of each level into descriptive information for generating images, and to obtain structured image description information based on the descriptive information corresponding to each of the at least two levels. The sticker set generation module is configured to generate the sticker set based on the structured image description information.

[0018] In one exemplary embodiment, the sticker set generation module includes: The resource information conversion unit is configured to convert the structured image description information into resource description information according to a preset resource description format; The sticker generation unit is configured to perform a task to build a set of stickers based on the resource description information, and to generate the set of stickers according to the task.

[0019] In one exemplary embodiment, the semantic extraction module includes: The first semantic extraction unit is configured to perform visual functions corresponding to each key semantic information to extract the main content representing the core theme and key details of the image from the multiple key semantic information to obtain the first level semantic information. The second semantic extraction unit is configured to extract secondary content representing image background and / or image supplementary information from the plurality of key semantic information to obtain second-level semantic information. The third semantic extraction unit is configured to extract decorative content representing decorative elements of the image from the multiple key semantic information to obtain third-level semantic information.

[0020] In one exemplary embodiment, the structured information acquisition module is further configured to perform the following steps: converting the first-level semantic information into first structured description information according to a preset sentence framework; converting the second-level semantic information into second structured description information according to the preset sentence framework; converting the third-level semantic information into third structured description information according to the preset sentence framework; and determining the structured image description information based on the first structured description information, the second structured description information, and the third structured description information.

[0021] In one exemplary embodiment, the resource information conversion unit includes: The verification unit is configured to perform verification on the structured image description information according to a preset verification strategy and obtain a verification result. The completion unit is configured to perform completion processing on the structured image description information according to the standardized fields based on the verification result, so as to obtain completed description information; The resource information conversion unit is configured to convert the completed description information into resource description information according to a preset resource description format.

[0022] In one exemplary embodiment, the resource description information includes description information corresponding to multiple levels of semantic information, and the device further includes: The resource identifier generation module is configured to generate a resource identifier corresponding to each description information according to the hierarchical semantic information corresponding to each description information in the resource description information. The relational storage module is configured to store the correspondence between the resource descriptors and each descriptor in the resource description information in the resource list.

[0023] In one exemplary embodiment, the sticker generation unit includes: The parsing unit is configured to parse the resource description information to obtain first resource description information corresponding to the first level semantic information, second resource description information corresponding to the second level semantic information, and third resource description information corresponding to the third level semantic information. The first sticker generation unit is configured to perform a first sticker generation task based on the first resource description information, and to generate a first-level sticker based on the first sticker generation task. The second sticker generation unit is configured to perform a second sticker generation task based on the second resource description information, and to generate a second-level sticker based on the second sticker generation task. The third sticker generation unit is configured to perform a third sticker generation task based on the third resource description information, and to generate a third-level sticker based on the third sticker generation task. The fusion unit is configured to perform fusion of the first-layer sticker, the second-layer sticker, and the third-layer sticker to generate the set of stickers.

[0024] In one exemplary embodiment, the sticker generation unit includes: The task construction unit is configured to execute a task to generate each sticker image in the sticker set based on the resource description information, and to construct the sticker set generation task according to the generation task of each sticker image. The sticker image generation unit is configured to perform a generation task based on each sticker image in the sticker set generation task and the text image model, and generate a sticker image corresponding to each generation task. The sticker set construction unit is configured to construct the sticker set based on the sticker image corresponding to each generation task.

[0025] In one exemplary embodiment, the sticker set construction unit includes: The stylization processing subunit is configured to perform stylization processing on each sticker image by inputting at least two stylization models to obtain multiple stylized images corresponding to each sticker image; The sticker combination subunit is configured to combine multiple stylized images corresponding to each sticker image to obtain the set of stickers.

[0026] In one exemplary embodiment, the sticker combination subunit is further configured to perform the following: combine multiple stylized images corresponding to each sticker image to obtain an initial set of images; perform background segmentation processing on each image in the initial set of images to remove the original background in each image and generate sticker material with a transparent background corresponding to each image; perform sticker enhancement processing on the multiple sticker materials respectively to obtain multiple enhanced stickers; the sticker enhancement processing includes adding at least one of stroke, shadow, and highlight; and construct the set of stickers based on the multiple enhanced stickers.

[0027] In one exemplary embodiment, the apparatus further includes: The path generation module is configured to upload the set of stickers to the resource storage system and generate the access path for the set of stickers; The information update module is configured to update the resource description information based on the access path, generation time, and version number of the image sticker set, thereby obtaining updated resource description information.

[0028] In one exemplary embodiment, the apparatus further includes: The information acquisition module is configured to acquire the storage path, generation parameters, and stylization information of the image set stickers; The description information generation module is configured to generate description information for the set of stickers based on the storage path, generation parameters, and stylization information of the sticker set. The relationship storage module is configured to store the correspondence between the sticker set and the description information of the sticker set in a preset database.

[0029] In one exemplary embodiment, the apparatus further includes: The image acquisition module is configured to acquire images of the objects to be processed. The image fusion module is configured to perform image fusion processing on the at least one sticker selected from the set of stickers and the at least one sticker and the object image to be processed, thereby generating a stylized object image.

[0030] According to a third aspect of the present disclosure, an electronic device is provided, comprising: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the method for generating sticker sets as described above.

[0031] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided, wherein when instructions in the computer-readable storage medium are executed by an electronic device processor, the electronic device is able to perform the method for generating overlay stickers as described above.

[0032] According to a fifth aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the method for generating image stickers as described above.

[0033] The technical solutions provided by the embodiments of this disclosure have at least the following beneficial effects: This invention obtains sticker description text input from a target object; the sticker description text is the description information of the sticker set to be generated; the sticker description text is parsed to extract multiple key semantic information; the multiple key semantic information is divided into at least two levels of hierarchical semantic information according to visual function; the hierarchical semantic information of each level is then converted into description information for generating images, and structured image description information is obtained based on the description information corresponding to the at least two levels; the sticker set is generated according to the structured image description information. This invention extracts multiple levels of semantic information and then converts it into structured image description information to construct sticker sets, thereby achieving rich visual layers in the constructed stickers, avoiding the monotony of sticker visual expression, and realizing the diversification of sticker sets.

[0034] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0035] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0036] Figure 1 This is an application environment diagram illustrating a method for generating image stickers according to an exemplary embodiment.

[0037] Figure 2 This is a flowchart illustrating a method for generating a set of stickers according to an exemplary embodiment.

[0038] Figure 3 This is a flowchart illustrating a method for dividing a plurality of key semantic information into at least two levels of hierarchical semantic information according to their visual function, based on an exemplary embodiment.

[0039] Figure 4This is a flowchart illustrating a method for converting structured image description information into resource description information according to a preset resource description format, based on an exemplary embodiment.

[0040] Figure 5 This is a schematic diagram illustrating a method for converting sticker description text into structured image description information according to an exemplary embodiment.

[0041] Figure 6 This is a schematic diagram illustrating a method for constructing resource description information according to an exemplary embodiment.

[0042] Figure 7 This is a schematic diagram illustrating a method for generating a sticker image according to an exemplary embodiment.

[0043] Figure 8 This is a schematic diagram illustrating a method for generating a set of stickers based on multiple stylized models, according to an exemplary embodiment.

[0044] Figure 9 This is a block diagram illustrating an apparatus for generating sticker sets according to an exemplary embodiment.

[0045] Figure 10 This is a block diagram illustrating a server according to an exemplary embodiment.

[0046] Figure 11 This is a block diagram illustrating an electronic device for generating sticker sets according to an exemplary embodiment. Detailed Implementation

[0047] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0048] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0049] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data used for analysis, etc.) involved in this disclosure require user authorization or full authorization from all parties when the embodiments of this disclosure are applied to specific products or technologies. Furthermore, the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions.

[0050] The following problems typically exist in the sticker generation process of related technologies: 1. Sticker generation methods are often limited by different designers or tools, making it difficult to maintain a consistent overall style, resulting in a lack of visual uniformity and professionalism.

[0051] 2. Insufficient visual hierarchy: Traditional generation methods struggle to achieve rich visual hierarchy in stickers, resulting in a monotonous visual presentation that fails to meet users' demands for diverse and high-quality visual effects.

[0052] 3. Low generation efficiency: Existing methods are mostly sequential, which is slow, resulting in long user waiting times, low efficiency during large-scale production, and high operating costs.

[0053] 4. Background processing relies on manual operation: Sticker background processing usually requires manual operation, resulting in low efficiency and poor flexibility in generating transparent background stickers.

[0054] To address the aforementioned technical problems, this embodiment provides a method, apparatus, device, and storage medium for generating image sticker sets.

[0055] Please see Figure 1 The diagram illustrates an application environment for a method of generating sticker sets according to an exemplary embodiment. The application environment may include a server 01 and a client 02.

[0056] Specifically, in the embodiments of this specification, server 01 may include a standalone server, a distributed server, or a server cluster composed of multiple servers. It may also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Server 01 may include a network communication unit, a processor, and a memory, etc. Specifically, server 01 can be used to construct structured image description information based on the sticker description text input by the target object, then generate the set of stickers based on the obtained structured image description information, and send the set of stickers to client 02.

[0057] Specifically, in this embodiment of the specification, the client 02 may include physical devices such as smartphones, desktop computers, tablets, laptops, digital assistants, smart wearable devices, and in-vehicle terminals, and may also include software running on the physical device, such as web pages provided to users by some service providers, or applications provided to users by these service providers. Specifically, the client 02 can be used to display photo sets of stickers.

[0058] Figure 2 This is a flowchart illustrating a method for generating a set of stickers according to an exemplary embodiment, such as... Figure 2 As shown, this method can be applied to Figure 1 The server 01 shown includes the following steps.

[0059] In step S201, the sticker description text input by the target object is obtained; the sticker description text is the description information of the sticker set to be generated.

[0060] In this embodiment, the target object may include, but is not limited to, sticker development users, sticker creation users, or other objects that require stickers. The method of this embodiment can be applied to a sticker generation application, which can provide a text input interface. In response to the target object's text input operation on this interface, sticker description text can be obtained. The text input method may include, but is not limited to, voice input, character input, etc. The sticker description text is the description information of the sticker set to be generated. For example, the sticker description text may be text such as "Help me generate a set of stickers in the XX style, including xxx content, xx background, and xx decoration."

[0061] In step S203, the sticker description text is parsed to extract multiple key semantic information; and the multiple key semantic information is divided into at least two levels of hierarchical semantic information according to their visual function.

[0062] In this embodiment, natural language processing technology can be used to parse the sticker description text, extract multiple key semantic information, and divide these key semantic information into at least two levels of hierarchical semantic information according to their visual function. These different levels of hierarchical semantic information can be used to create multi-layered visual effects in the sticker; for example, multiple key semantic information can be layered into three levels according to their visual function: main content, secondary content, and decorative elements. The number of levels can also be adjusted according to actual needs; no specific limit is placed on the number of levels here.

[0063] In step S205, the hierarchical semantic information of each level is converted into descriptive information for generating the image, and structured image descriptive information is obtained based on the descriptive information corresponding to each of the at least two levels.

[0064] In this embodiment of the disclosure, the hierarchical semantic information of each level can be transformed into descriptive information for generating images. For example, a unified preset framework can be used to transform the semantic information of each level into structured image descriptive information. This framework uses a fixed sentence structure to ensure consistency in style, language, and semantics among the stickers. This allows the descriptive information corresponding to at least two levels to be aggregated, resulting in structured image descriptive information for the sticker set.

[0065] In step S207, the image sticker set is generated based on the structured image description information.

[0066] After obtaining the structured image description information, this embodiment can generate a set of stickers with rich visual layers that cover multiple levels of semantic information.

[0067] This embodiment obtains the sticker description text input by the target object; the sticker description text is the description information of the sticker set to be generated; the sticker description text is parsed to extract multiple key semantic information; the multiple key semantic information is divided into at least two levels of hierarchical semantic information according to visual function; then the hierarchical semantic information of each level is converted into description information for generating the image, and based on the description information corresponding to the at least two levels, structured image description information is obtained; the sticker set is generated according to the structured image description information. By extracting multi-level semantic information and then converting it into structured image description information to construct the sticker set, rich visual layers can be achieved in the constructed stickers, avoiding the monotony of the sticker's visual expression and realizing the diversification of sticker sets.

[0068] In some embodiments, generating the image set stickers based on the structured image description information may include: The structured image description information is converted into resource description information according to a preset resource description format; Based on the resource description information, a set of stickers generation task is constructed, and the set of stickers is generated according to the set of stickers generation task.

[0069] In this embodiment, structured image description information can be converted into resource description information according to a preset resource description format. This facilitates the aggregation of structured image description information into a structured list of resource description information, supporting efficient sticker generation scheduling and resource management, thereby ensuring the smooth progress of the sticker generation process. A generation task for each sticker in a sticker set can be constructed based on the resource description information, allowing the sticker set to be generated according to the sticker set generation task. This achieves efficient batch generation of high-quality sticker images, significantly improving generation speed and system reliability, and is suitable for large-scale production scenarios.

[0070] In this embodiment of the disclosure, the hierarchical semantic information of each level is converted into descriptive information for generating images, and then into resource description information to construct a sticker generation task. This can be used to automatically convert the sticker description text input by the user into a sticker generation task and automatically generate a set of hierarchical stickers, thereby improving the generation efficiency of the set of stickers.

[0071] In some embodiments, such as Figure 3 As shown, dividing the multiple key semantic information into at least two levels of hierarchical semantic information according to their visual function includes: S301: Based on the visual function corresponding to each key semantic information, extract the main content representing the core theme and key details of the image from the multiple key semantic information to obtain the first level semantic information; S303: Extract secondary content representing the image background and / or supplementary image information from the multiple key semantic information to obtain second-level semantic information; S305: Extract the decorative content representing the decorative elements of the image from the multiple key semantic information to obtain the third-level semantic information.

[0072] In this embodiment, multiple key semantic information may include the core theme of the image, key details, image background and / or supplementary image information, and decorative elements of the image. These multiple key semantic information can be hierarchically divided into main content, secondary content, and decorative elements. The main content representing the core theme and key details of the image can be used as the first-level semantic information; the secondary content representing the image background and / or supplementary image information can be used as the second-level semantic information; and the decorative content representing decorative elements of the image can be used as the third-level semantic information. This results in three levels of semantic information. During the construction of the sticker set, each sticker can be constructed based on these three levels of hierarchical semantic information, thereby enriching the visual hierarchy of the stickers, increasing the diversity of the sticker content hierarchy, and enhancing the visual effect of the stickers.

[0073] In some embodiments, the step of converting the hierarchical semantic information of each level into descriptive information for generating an image, and obtaining structured image descriptive information based on the descriptive information corresponding to each of the at least two levels, includes: The first-level semantic information is transformed into first-structured descriptive information according to a preset sentence framework; The second-level semantic information is transformed into second-structured descriptive information according to the preset sentence framework; The third-level semantic information is transformed into third-structured descriptive information according to the preset sentence framework; The structured image description information is determined based on the first structured description information, the second structured description information, and the third structured description information.

[0074] In this embodiment, a preset sentence framework can be pre-constructed, and then the first-level semantic information, the second-level semantic information, and the third-level semantic information can all be converted into structured description information according to the preset sentence framework; then the obtained first structured description information, second structured description information, and third structured description information are determined as structured image description information; thereby ensuring the consistency of stickers at each level in style, language, and semantics.

[0075] In some embodiments, the resource description information includes description information corresponding to multiple levels of semantic information, and the method further includes: Generate a resource identifier corresponding to each description information according to the hierarchical semantic information corresponding to each description information in the resource description information; The resource list stores the correspondence between the resource descriptors and each descriptive information in the resource description information.

[0076] In this embodiment of the disclosure, in order to distinguish the description information of different levels, a resource identifier corresponding to each description information can be generated according to the hierarchical semantic information corresponding to each description information in the resource description information; then, a resource list can be constructed according to the correspondence between the resource descriptor and each description information in the resource description information, so as to facilitate the construction of sticker generation tasks of each level according to the resource list, and then the stickers of each level are merged to generate multi-level stickers.

[0077] In some embodiments, such as Figure 4 As shown, the step of converting the structured image description information into resource description information according to a preset resource description format includes: S401: The structured image description information is verified according to a preset verification strategy to obtain the verification result; S403: Based on the verification result, the structured image description information is completed according to the standardized fields to obtain the completed description information; S405: Convert the completed description information into resource description information according to a preset resource description format.

[0078] In this embodiment of the disclosure, to ensure the semantic accuracy of the resource description information, the structured image description information can be validated first, and then semantic completion, logical correction, and other processing can be performed based on the validation results. Finally, it can be converted into resource description information according to a preset resource description format. A preset validation strategy can be pre-built; for example, the preset validation strategy can be a semantic integrity validation strategy. Validation is then performed according to this strategy to obtain the validation result. If the validation result indicates that the semantics of the structured image description information is incomplete, semantic completion processing can be performed on the structured image description information according to preset standardized fields to obtain completed description information.

[0079] In some embodiments, the preset verification strategy may also include a logical verification strategy, a syntax verification strategy, etc., so as to discover and identify errors, inconsistencies and anomalies in the data, and correct logical and syntax errors in the structured image description information, so as to ensure that the corrected semantic data is grammatically and logically correct and without contradictions.

[0080] In this embodiment, the preset resource description format can be a format that includes information such as theme, color, shape, category identifier, and resource identifier. The complete description information at each level can be converted into a list of resource description information according to the preset resource description format. Different resource identifiers correspond to different levels of complete description information; the resource identifiers are used to distinguish between different levels of description information.

[0081] In some embodiments, the step of constructing a sticker set generation task based on the resource description information and generating the sticker set according to the sticker set generation task includes: The resource description information is parsed to obtain the first resource description information corresponding to the first level semantic information, the second resource description information corresponding to the second level semantic information, and the third resource description information corresponding to the third level semantic information; A first sticker generation task is generated based on the first resource description information, and a first-level sticker is generated based on the first sticker generation task. A second sticker generation task is generated based on the second resource description information, and a second-level sticker is generated based on the second sticker generation task. A third sticker generation task is generated based on the third resource description information, and a third-level sticker is generated based on the third sticker generation task. The first layer sticker, the second layer sticker, and the third layer sticker are merged to generate the set of stickers.

[0082] In this embodiment of the disclosure, the resource description information may include the first resource description information corresponding to the first level semantic information, the second resource description information corresponding to the second level semantic information, and the third resource description information corresponding to the third level semantic information. Then, sticker generation tasks can be constructed according to the resource description information of each level in sequence, and stickers of each level can be generated according to each task. Finally, the first level sticker, the second level sticker, and the third level sticker are merged to generate at least one merged sticker, which can then be further aggregated to obtain a set of stickers. In some embodiments, the step of constructing a sticker set generation task based on the resource description information and generating the sticker set according to the sticker set generation task includes: Based on the resource description information, a generation task for each sticker image in the sticker set is constructed, and the sticker set generation task is constructed according to the generation task for each sticker image. Based on the generation task of each sticker image in the sticker set generation task and the text image model, generate the sticker image corresponding to each generation task. The set of stickers is constructed based on the sticker image corresponding to each generated task.

[0083] In this embodiment of the disclosure, a text-based image model can be pre-trained, and then sticker images can be generated using this model. For example, the training method for the text-based image model may include: Obtain sample resource description information, which is labeled with sample sticker image tags; The sample resource description information is input into a preset model for image generation processing to obtain a sample prediction image; Based on the difference between the predicted image of the sample and the label of the sample sticker image, the preset model is trained to obtain the text-based image model.

[0084] In this embodiment of the disclosure, the preset model may include, but is not limited to, a large model or OpenAI (which technically adopts an autoregressive structure and is optimized through reinforcement learning and human feedback). During the model training process, the target loss data can be determined based on the difference between the sample predicted image and the sample sticker image label. Then, the parameters of the preset model are adjusted based on the target loss data until the training termination condition is met, and the preset model at the end of training is used as the text-to-image model.

[0085] The core of the Wensheng graph model is cross-modal understanding and generation, which mainly consists of two steps: Text-image association: The model must first understand your textual description. This is typically accomplished by models such as CLIP (Contrastive Language–Image Pre-training). CLIP learns from massive text-image pairs, mapping text and images to the same semantic space, thereby understanding what visual features should correspond to the text like "cat".

[0086] Image generation: After understanding the text, the model begins to "create". The current mainstream technique is the diffusion model. It first adds noise to the training image step by step, and then learns how to reconstruct a clear image from pure noise. During generation, the model starts with random noise, removes noise step by step according to the text prompts, and finally forms an image that matches the description.

[0087] After model training is complete, sticker generation tasks for each level can be constructed based on resource description information. These tasks can include resource description information for each level of sticker and control text for sticker generation. The resource description information and control text are then input into the text-to-image model, which generates sticker images for each task. Finally, the sticker images from each level are fused together to quickly obtain a set of stickers.

[0088] In some embodiments, constructing the sticker set based on the sticker image corresponding to each generated task includes: Each sticker image is input into at least two stylization models for stylization processing, resulting in multiple stylized images corresponding to each sticker image; The multiple stylized images corresponding to each sticker image are combined to obtain the set of stickers.

[0089] In this embodiment, the sticker image corresponding to each generation task can be stylized, and then the stylized images of at least two levels of the same style can be fused to obtain multi-level stickers of each style. Since there can be multiple sticker images of the same level for each style, there can also be multiple multi-level stickers obtained after image fusion. Thus, multiple multi-level stickers of the same style can be combined to form an initial set of images; and initial sets of images of various styles can also be combined to generate sticker sets of multiple styles. For example, the stylization model can include, but is not limited to, a flat style model and a cartoon style model. For example, the stylization model can be a LoRA model. LoRa (Low-Rank Adaptation) is a lightweight model fine-tuning technique, particularly suitable for rapid adaptation and customization based on large pre-trained models. At least two stylization models can be included, but are not limited to, LoRA1, LoRA2, and LoRA3. Each model represents a style, and each stylization model can be used to output multiple stickers of the same style corresponding to semantic information at the same level. For example, stickers corresponding to semantic information at multiple levels output by the same stylization model can be fused to obtain a set of stickers of the same style. Specifically, multiple initial stickers corresponding to semantic information at the same level output by any stylization model can be obtained; then, one initial sticker corresponding to each level of semantic information can be obtained, and the initial stickers corresponding to each level of semantic information can be fused to obtain multi-level fused stickers; thus, multiple fused stickers of the same style can be obtained, and then the multiple fused stickers can be spliced ​​together to obtain a set of stickers of the same style. In some embodiments, multiple stylization images corresponding to each sticker image can also be combined to obtain a set of stickers with multiple styles, thereby improving the diversity of the set of stickers.

[0090] In some embodiments, combining multiple stylized images corresponding to each sticker image to obtain the sticker set includes: Combine multiple stylized images corresponding to each sticker image to obtain the initial set of images; Perform background segmentation on each image in the initial set of images, remove the original background from each image, and generate sticker materials with transparent backgrounds for each image; Multiple sticker materials are subjected to sticker enhancement processing to obtain multiple enhanced stickers; the sticker enhancement processing includes adding at least one of stroke, shadow, and highlight; The set of stickers is constructed based on the multiple enhanced stickers.

[0091] In this embodiment of the disclosure, after obtaining the initial set of images, background segmentation processing can be performed on each image in the initial set of images to remove the original background in each image and generate sticker materials with transparent backgrounds corresponding to each image; thereby improving the efficiency of sticker background removal and improving the clarity of the main content of the sticker materials; then, sticker enhancement processing can be performed on multiple sticker materials to obtain multiple enhanced stickers; the sticker enhancement processing includes adding at least one of stroke, shadow, and highlight; for example, stroke, shadow, and highlight processing can be added to each sticker material at the same time, thereby improving the adaptability of enhanced stickers in multiple scenarios and making them more beautiful and easy to use.

[0092] In some embodiments, after constructing the set of stickers based on the plurality of enhanced stickers, the method further includes: Upload the set of stickers to the resource storage system and generate the access path for the set of stickers; Based on the access path, generation time, and version number of the image set stickers, the resource description information is updated to obtain updated resource description information.

[0093] In this embodiment, the processed sticker sets can be uploaded to a resource storage system and an access path can be generated. The system will write back the path, generation time, version number, and other information to the resource description, enabling the resources to be managed, accessed, and traced. This structured storage method facilitates users' access to and retrieval of stickers at any time, improving resource management efficiency and ease of use.

[0094] In some embodiments, after constructing the set of stickers based on the plurality of enhanced stickers, the method further includes: Obtain the storage path, generation parameters, and stylization information of the sticker set; Based on the storage path, generation parameters, and stylization information of the image set stickers, generate image set description information; The correspondence between the sticker set and the description information of the sticker set is stored in a preset database.

[0095] In this embodiment of the disclosure, image set description information can be generated based on the storage path, generation parameters, and stylization information of the image set stickers. The system can output image set description information containing structured information such as sticker image storage path, generation parameters, and style settings, ensuring that the stickers are consistent in style, content hierarchy, and visual rhythm, supporting multi-scenario use as a set, facilitating user display, management, and integration, and ensuring the best effect in various applications.

[0096] In some embodiments, the method further includes: Obtain the image of the object to be processed; At least one sticker is selected from the set of stickers, and the at least one sticker is fused with the object image to generate a stylized object image.

[0097] In this embodiment of the disclosure, after generating a set of stickers, the user can quickly generate a stylized object image by inputting an image of the object to be processed. For example, in response to an image input command, the image of the object to be processed can be acquired; then, at least one sticker can be selected from the set of stickers, and the at least one sticker can be fused with the image of the object to be processed to generate a stylized object image; for example, the image of the object to be processed can also be fused with each sticker in the set of stickers separately, thereby generating multiple stylized object images in batches; this improves the generation efficiency of stylized object images.

[0098] In one exemplary embodiment, a method for generating image set stickers is provided, the main steps of which include: 1. Identify and hierarchically organize key semantic elements from the sticker description text input by the target object to form a structured semantic representation; 2. Transform semantic information into a structured description for image generation according to standardized rules to ensure consistent style and controllable generation; 3. Generate resource descriptions in a standard format and assign unique identifiers to support task tracking and large-scale management; 4. Image generation tasks are created based on resource descriptions, and a parallel scheduling strategy is adopted to improve concurrent processing capabilities; 5. Call the text image model that supports flat style control to generate sticker images with consistent style and stable quality based on the description; 6. Perform background removal and sticker-specific enhancement processing on the generated images to improve adaptability and visual effects in multiple scenes; 7. Unify the storage of generated results and output them as an integrable set of image resources, delivering sticker assets that are ready to use and consistent in style.

[0099] For example, a specific implementation of a method for generating image set stickers may include: 1. Element extraction and hierarchical processing The system uses natural language processing (NLP) to parse the sticker description text input by the user, extract key semantic information, and categorize it into primary content, secondary content, and decorative elements based on their visual function. This hierarchical structure serves as input to the generative model, ensuring that the image maintains consistent and controllable visual emphasis, layout, and style, thereby accurately generating stickers that meet expectations.

[0100] 2. Description Generation and Structured Representation After acquiring the hierarchical semantics, the system converts the content of each layer into a structured image description according to a pre-defined unified framework. This framework uses fixed sentence structures to ensure consistency in style, language, and semantics among the stickers: the main content contains the core theme and key details, with high information density; secondary content provides background or supplementary explanations; and decorative content serves as decorative elements to enhance the visual effect. The final description is organized into a standardized data structure for subsequent processes, ensuring a clear and orderly visual hierarchy for the stickers.

[0101] like Figure 5 As shown, Figure 5 This is a schematic diagram of a method for converting sticker description text into structured image description information. The method includes: semantic parsing of the sticker description text input by the user; specifically, Natural Language Processing (NLP) technology can be used for semantic parsing to extract multiple key semantic information; then, these key semantic information are divided into three levels according to their visual function: a first-level semantic information layer (main element layer L1), a second-level semantic information layer (minor element layer L2), and a third-level semantic information layer (embellishment element layer L3). The main element layer can include main element 1, main element 2, and main element 3; the minor element layer can include minor element 1, minor element 2, and minor element 3; and the embellishment element layer can include embellishment element 1, embellishment element 2, and embellishment element 3. Finally, a template layer with a fixed sentence structure is used to convert the elements at each level, thereby obtaining structured image description information. 3. Resource Description Construction like Figure 6 As shown, Figure 6 This diagram illustrates a method for constructing resource description information. It converts structured image description information into a unified resource description format, including core information such as theme, color, shape, category identifier, and resource identifier. Specifically, a merging and normalization module verifies the structured image description information, and then performs a completion operation using standardized fields based on the verification results to obtain complete description information. Next, according to the hierarchical semantic information corresponding to each description piece in the resource description information, a resource identifier is generated for each description piece. Based on the correspondence between resource descriptors and each description piece, resource entries are generated; and the correspondence between the resource descriptors and each description piece in the resource description information, i.e., resource entries, is stored in a resource list. All resource items are aggregated into a structured list according to rules, supporting efficient generation scheduling and resource management, ensuring the smooth progress of the generation process.

[0102] 4. Sticker generation task scheduling and concurrent processing The system creates a generation task for each sticker based on the resource description, employing a concurrent processing mechanism to achieve synchronous or asynchronous generation of multiple resource items while ensuring stability. By optimizing resource allocation and task scheduling, it can efficiently generate high-quality sticker images in batches, significantly improving generation speed and system reliability, making it suitable for large-scale production scenarios.

[0103] 5. Image generation service invocation and style control like Figure 7 As shown, Figure 7 This is a schematic diagram illustrating a method for generating sticker images. This embodiment utilizes a text-based image model (which can be a basic text-based image model) to automatically create sticker content, and then combines it with various LoRa (Low-Rank Adaptation) models (including LoRa1, LoRa2, and LoRa3) to achieve flat and cartoonish style control. Parameters such as color, lines, and shadows can be dynamically adjusted based on the resource description, generating diverse stickers while maintaining stylistic consistency, balancing consistency and application adaptability, and improving user experience.

[0104] For example, such as Figure 8 As shown, Figure 8 This is a schematic diagram of a method for generating image set stickers based on multiple stylization models. The stylization models can include two large models with different styles (large models 1 and 2). First, style routing judgment can be performed on the sticker image, and then the large model with the corresponding style can be selected to stylize the sticker image. For the two example styles, for style 1: first, the user input is analyzed for intent, then multiple key semantic information corresponding to the main effect is extracted, and then control text is generated. Based on the control text, the multiple key semantic information is parsed to obtain the first-level semantic information corresponding to Layout1 (main element layer), the second-level semantic information corresponding to Layout2 (secondary element layer), and the third-level semantic information corresponding to Layout3 (embellishment element layer). Then, the structured description information corresponding to Layout1, Layout2, and Layout3 is generated respectively, and then further converted into resource description information to generate the image set stickers.

[0105] For Style 2: First, perform intent analysis on the user input, then extract multiple key semantic information corresponding to the main body of the special effect, and then parse the multiple key semantic information to obtain the first-level semantic information corresponding to Layout1 (main element layer), the second-level semantic information corresponding to Layout2 (secondary element layer), and the third-level semantic information corresponding to Layout3 (embellishment element layer); then generate the main elements corresponding to Layout1, Layout2, and Layout3 in sequence, and further generate structured description information, which is then converted into resource description information to generate the set of stickers.

[0106] 6. Background processing and sticker enhancement To adapt to different sticker usage scenarios, the generated image undergoes background processing to remove the original background while retaining the main content, creating sticker materials with transparent backgrounds. This process employs image segmentation and background removal technology to ensure the subject is clear and allows for sticker enhancements as needed, such as adding outlines, shadows, or highlights, improving adaptability to various scenarios and making the stickers more aesthetically pleasing and user-friendly.

[0107] 7. Resource storage and information write-back Once processed, the image will be uploaded to the resource storage system and an access path will be generated. The system will then write back the path, generation time, version number, and other information to the resource description, enabling the resource to be managed, accessed, and traced. This structured storage method allows users to access and call the stickers at any time, improving resource management efficiency and ease of use.

[0108] 8. Output of image set stickers The system outputs a description of the sticker set resource containing structured information such as sticker image storage path, generation parameters, and style settings. This ensures that the stickers are consistent in style, content hierarchy, and visual rhythm, supports multi-scenario use as a set, facilitates user display, management, and integration, and guarantees the best results in various applications.

[0109] This disclosure provides an automated production line for creating cute and adorable sticker sets. This production line covers all stages from design and production to finished product output. Simultaneously, it optimizes the connections between each process, ensuring a more efficient, stable, and controllable production process. This allows for the provision of higher-quality cute and adorable sticker sets with special effects. The sticker sets in this embodiment can be used to overlay multiple layers of visual content—including "face enhancement + cute style effects + themed stickers + ambient lighting effects"—on an existing object image (e.g., a portrait); presenting a cute, warm, and fun visual effect with a unified style.

[0110] Figure 9 This is a block diagram illustrating an apparatus for generating sticker sets according to an exemplary embodiment. (Refer to...) Figure 9 The device includes: The description text acquisition module 910 is configured to acquire the sticker description text input by the target object; the sticker description text is the description information of the set of stickers to be generated; The semantic extraction module 920 is configured to parse the sticker description text, extract multiple key semantic information, and divide the multiple key semantic information into at least two levels of hierarchical semantic information according to their visual function. The structured information acquisition module 930 is configured to convert the hierarchical semantic information of each level into descriptive information for generating an image, and obtain structured image description information based on the descriptive information corresponding to each of the at least two levels. The sticker set generation module 940 is configured to generate the sticker set based on the structured image description information.

[0111] In one exemplary embodiment, the sticker set generation module includes: The resource information conversion unit is configured to convert the structured image description information into resource description information according to a preset resource description format; The sticker generation unit is configured to perform a task to build a set of stickers based on the resource description information, and to generate the set of stickers according to the task.

[0112] In one exemplary embodiment, the semantic extraction module includes: The first semantic extraction unit is configured to perform visual functions corresponding to each key semantic information to extract the main content representing the core theme and key details of the image from the multiple key semantic information to obtain the first level semantic information. The second semantic extraction unit is configured to extract secondary content representing image background and / or image supplementary information from the plurality of key semantic information to obtain second-level semantic information. The third semantic extraction unit is configured to extract decorative content representing decorative elements of the image from the multiple key semantic information to obtain third-level semantic information.

[0113] In one exemplary embodiment, the structured information acquisition module is further configured to perform the following steps: converting the first-level semantic information into first structured description information according to a preset sentence framework; converting the second-level semantic information into second structured description information according to the preset sentence framework; converting the third-level semantic information into third structured description information according to the preset sentence framework; and determining the structured image description information based on the first structured description information, the second structured description information, and the third structured description information.

[0114] In one exemplary embodiment, the resource information conversion unit includes: The verification unit is configured to perform verification on the structured image description information according to a preset verification strategy and obtain a verification result. The completion unit is configured to perform completion processing on the structured image description information according to the standardized fields based on the verification result, so as to obtain completed description information; The resource information conversion unit is configured to convert the completed description information into resource description information according to a preset resource description format.

[0115] In one exemplary embodiment, the resource description information includes description information corresponding to multiple levels of semantic information, and the device further includes: The resource identifier generation module is configured to generate a resource identifier corresponding to each description information according to the hierarchical semantic information corresponding to each description information in the resource description information. The relational storage module is configured to store the correspondence between the resource descriptors and each descriptor in the resource description information in the resource list.

[0116] In one exemplary embodiment, the sticker generation unit includes: The parsing unit is configured to parse the resource description information to obtain first resource description information corresponding to the first level semantic information, second resource description information corresponding to the second level semantic information, and third resource description information corresponding to the third level semantic information. The first sticker generation unit is configured to perform a first sticker generation task based on the first resource description information, and to generate a first-level sticker based on the first sticker generation task. The second sticker generation unit is configured to perform a second sticker generation task based on the second resource description information, and to generate a second-level sticker based on the second sticker generation task. The third sticker generation unit is configured to perform a third sticker generation task based on the third resource description information, and to generate a third-level sticker based on the third sticker generation task. The fusion unit is configured to perform fusion of the first-layer sticker, the second-layer sticker, and the third-layer sticker to generate the set of stickers.

[0117] In one exemplary embodiment, the sticker generation unit includes: The task construction unit is configured to execute a task to generate each sticker image in the sticker set based on the resource description information, and to construct the sticker set generation task according to the generation task of each sticker image. The sticker image generation unit is configured to perform a generation task based on each sticker image in the sticker set generation task and the text image model, and generate a sticker image corresponding to each generation task. The sticker set construction unit is configured to construct the sticker set based on the sticker image corresponding to each generation task.

[0118] In one exemplary embodiment, the sticker set construction unit includes: The stylization processing subunit is configured to perform stylization processing on each sticker image by inputting at least two stylization models to obtain multiple stylized images corresponding to each sticker image; The sticker combination subunit is configured to combine multiple stylized images corresponding to each sticker image to obtain the set of stickers.

[0119] In one exemplary embodiment, the sticker combination subunit is further configured to perform the following: combine multiple stylized images corresponding to each sticker image to obtain an initial set of images; perform background segmentation processing on each image in the initial set of images to remove the original background in each image and generate sticker material with a transparent background corresponding to each image; perform sticker enhancement processing on the multiple sticker materials respectively to obtain multiple enhanced stickers; the sticker enhancement processing includes adding at least one of stroke, shadow, and highlight; and construct the set of stickers based on the multiple enhanced stickers.

[0120] In one exemplary embodiment, the apparatus further includes: The path generation module is configured to upload the set of stickers to the resource storage system and generate the access path for the set of stickers; The information update module is configured to update the resource description information based on the access path, generation time, and version number of the image sticker set, thereby obtaining updated resource description information.

[0121] In one exemplary embodiment, the apparatus further includes: The information acquisition module is configured to acquire the storage path, generation parameters, and stylization information of the image set stickers; The description information generation module is configured to generate description information for the set of stickers based on the storage path, generation parameters, and stylization information of the sticker set. The relationship storage module is configured to store the correspondence between the sticker set and the description information of the sticker set in a preset database.

[0122] In one exemplary embodiment, the apparatus further includes: The image acquisition module is configured to acquire images of the objects to be processed. The image fusion module is configured to perform image fusion processing on the at least one sticker selected from the set of stickers and the at least one sticker and the object image to be processed, thereby generating a stylized object image.

[0123] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0124] In one exemplary embodiment, an electronic device is also provided, including a processor; a memory for storing processor-executable instructions; wherein, when the processor is configured to execute the instructions stored in the memory, it implements the method for generating overlay stickers provided in any of the above embodiments.

[0125] The electronic device can be a terminal, a server, or a similar computing device. Taking a server as an example... Figure 10 This is a block diagram of a server according to an exemplary embodiment, such as... Figure 10 As shown, the server 1000 can vary significantly due to different configurations or performance. It may include one or more central processing units (CPUs) 1010 (CPUs 1010 may include, but are not limited to, microprocessors (MCUs) or programmable logic devices (FPGAs), a memory 1030 for storing data, and one or more storage media 1020 (e.g., one or more mass storage devices) for storing application programs 1023 or data 1022. The memory 1030 and storage media 1020 may be temporary or persistent storage. The program stored in the storage media 1020 may include one or more modules, each module may include a series of instruction operations on the server. Furthermore, the CPU 1010 may be configured to communicate with the storage media 1020 and execute the series of instruction operations in the storage media 1020 on the server 1000. Server 1000 may also include one or more power supplies 1060, one or more wired or wireless network interfaces 1050, one or more input / output interfaces 1040, and / or one or more operating systems 1021, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, etc.

[0126] The input / output interface 1040 can be used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of server 1000. In one example, the input / output interface 1040 includes a network interface controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the input / output interface 1040 may be a radio frequency (RF) module used for wireless communication with the Internet.

[0127] Those skilled in the art will understand that Figure 10 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, server 1000 may also include... Figure 10 The more or fewer components shown, or having the same Figure 10 The different configurations shown.

[0128] In one exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 1030 including instructions, which can be executed by a processor 1010 of server 1000 to perform the above-described method. Optionally, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0129] Figure 11 This is a block diagram illustrating an electronic device for generating image stickers according to an exemplary embodiment. The electronic device may be a terminal, and its internal structure diagram may be as follows: Figure 11 As shown, the electronic device includes a processor, memory, network interface, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for generating overlay stickers. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the device's casing, or an external keyboard, touchpad, or mouse. Those skilled in the art will understand that Figure 11The structure shown is merely a block diagram of a portion of the structure related to the present disclosure and does not constitute a limitation on the electronic device to which the present disclosure is applied. A specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0130] In an exemplary embodiment, an electronic device is also provided, comprising: A processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the above-described method for generating image stickers.

[0131] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory including instructions, which can be executed by a processor of an electronic device to complete the above-described method for generating the overlay stickers. Optionally, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0132] In an exemplary embodiment, a computer program product is also provided, including a computer program that, when executed by a processor, implements the above-described method for generating image stickers.

[0133] This disclosure obtains sticker description text input by a target object; the sticker description text is the description information of the sticker set to be generated; the sticker description text is parsed to extract multiple key semantic information; the multiple key semantic information is divided into at least two levels of hierarchical semantic information according to visual function; the hierarchical semantic information of each level is then converted into description information for generating images, and structured image description information is obtained based on the description information corresponding to the at least two levels; the sticker set is generated according to the structured image description information. This invention extracts multiple levels of semantic information and then converts it into structured image description information to construct sticker sets, thereby achieving rich visual layers in the constructed stickers, avoiding the monotony of sticker visual expression, and realizing the diversification of sticker sets.

[0134] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0135] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0136] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for generating a set of sticker images, characterized in that, include: Get the sticker description text input by the target object; The sticker description text is the description information of the sticker set to be generated; The sticker description text is parsed to extract multiple key semantic information; and the multiple key semantic information is divided into at least two levels of hierarchical semantic information according to their visual function. The hierarchical semantic information of each level is transformed into descriptive information for generating images, and structured image description information is obtained based on the descriptive information corresponding to each of the at least two levels. The set of stickers is generated based on the structured image description information.

2. The method according to claim 1, characterized in that, The step of generating the image set stickers based on the structured image description information includes: The structured image description information is converted into resource description information according to a preset resource description format; Based on the resource description information, a set of stickers generation task is constructed, and the set of stickers is generated according to the set of stickers generation task.

3. The method according to claim 2, characterized in that, The step of dividing the multiple key semantic information into at least two levels according to their visual function includes: Based on the visual function corresponding to each key semantic information, the main content representing the core theme and key details of the image is extracted from the multiple key semantic information to obtain the first level of semantic information; Secondary content representing the image background and / or supplementary image information is extracted from the multiple key semantic information to obtain second-level semantic information; The decorative elements representing the decorative elements of the image are extracted from the multiple key semantic information to obtain the third level of semantic information.

4. The method according to claim 3, characterized in that, The process of converting the hierarchical semantic information of each level into descriptive information for generating images, and obtaining structured image description information based on the descriptive information corresponding to each of the at least two levels, includes: The first-level semantic information is transformed into first-structured descriptive information according to a preset sentence framework; The second-level semantic information is transformed into second-structured descriptive information according to the preset sentence framework; The third-level semantic information is transformed into third-structured descriptive information according to the preset sentence framework; The structured image description information is determined based on the first structured description information, the second structured description information, and the third structured description information.

5. The method according to claim 2, characterized in that, The step of converting the structured image description information into resource description information according to a preset resource description format includes: The structured image description information is verified according to a preset verification strategy to obtain the verification result; Based on the verification result, the structured image description information is completed according to the standardized fields to obtain the completed description information; The completed description information is converted into resource description information according to a preset resource description format.

6. The method according to claim 5, characterized in that, The resource description information includes description information corresponding to multiple levels of semantic information, and the method further includes: Generate a resource identifier corresponding to each description information according to the hierarchical semantic information corresponding to each description information in the resource description information; The resource list stores the correspondence between the resource descriptors and each descriptive information in the resource description information.

7. The method according to claim 4, characterized in that, The step of constructing a sticker set generation task based on the resource description information, and generating the sticker set according to the sticker set generation task, includes: The resource description information is parsed to obtain the first resource description information corresponding to the first level semantic information, the second resource description information corresponding to the second level semantic information, and the third resource description information corresponding to the third level semantic information; A first sticker generation task is generated based on the first resource description information, and a first-level sticker is generated based on the first sticker generation task. A second sticker generation task is generated based on the second resource description information, and a second-level sticker is generated based on the second sticker generation task. A third sticker generation task is generated based on the third resource description information, and a third-level sticker is generated based on the third sticker generation task. The first layer sticker, the second layer sticker, and the third layer sticker are merged to generate the set of stickers.

8. The method according to claim 2, characterized in that, The step of constructing a sticker set generation task based on the resource description information, and generating the sticker set according to the sticker set generation task, includes: Based on the resource description information, a generation task for each sticker image in the sticker set is constructed, and the sticker set generation task is constructed according to the generation task for each sticker image. Based on the generation task of each sticker image in the sticker set generation task and the text image model, generate the sticker image corresponding to each generation task. The set of stickers is constructed based on the sticker image corresponding to each generated task.

9. The method according to claim 8, characterized in that, The step of constructing the set of stickers based on the sticker image corresponding to each generated task includes: Each sticker image is input into at least two stylization models for stylization processing, resulting in multiple stylized images corresponding to each sticker image; The multiple stylized images corresponding to each sticker image are combined to obtain the set of stickers.

10. The method according to claim 9, characterized in that, The step of combining multiple stylized images corresponding to each sticker image to obtain the sticker set includes: Combine multiple stylized images corresponding to each sticker image to obtain the initial set of images; Perform background segmentation on each image in the initial set of images, remove the original background from each image, and generate sticker materials with transparent backgrounds for each image; Multiple sticker materials are subjected to sticker enhancement processing to obtain multiple enhanced stickers; the sticker enhancement processing includes adding at least one of stroke, shadow, and highlight; The set of stickers is constructed based on the multiple enhanced stickers.

11. The method according to claim 10, characterized in that, After constructing the set of stickers based on the plurality of enhanced stickers, the method further includes: Upload the set of stickers to the resource storage system and generate the access path for the set of stickers; Based on the access path, generation time, and version number of the image set stickers, the resource description information is updated to obtain updated resource description information.

12. The method according to claim 10, characterized in that, After constructing the set of stickers based on the plurality of enhanced stickers, the method further includes: Obtain the storage path, generation parameters, and stylization information of the sticker set; Based on the storage path, generation parameters, and stylization information of the image set stickers, generate image set description information; The correspondence between the sticker set and the description information of the sticker set is stored in a preset database.

13. The method according to claim 10, characterized in that, The method further includes: Obtain the image of the object to be processed; At least one sticker is selected from the set of stickers, and the at least one sticker is fused with the object image to generate a stylized object image.

14. A device for generating photo sticker sets, characterized in that, include: The description text acquisition module is configured to acquire the sticker description text input from the target object. The sticker description text is the description information of the sticker set to be generated; The semantic extraction module is configured to parse the sticker description text, extract multiple key semantic information, and divide the multiple key semantic information into at least two levels of hierarchical semantic information according to their visual function. The structured information acquisition module is configured to convert the hierarchical semantic information of each level into descriptive information for generating images, and to obtain structured image description information based on the descriptive information corresponding to each of the at least two levels. The sticker set generation module is configured to generate the sticker set based on the structured image description information.

15. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the method for generating sticker sets as described in any one of claims 1-13.

16. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by an electronic device processor, the electronic device is able to perform the method for generating sticker sets as described in any one of claims 1-13.

17. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the method for generating the sticker set according to any one of claims 1-13.

Images