METHOD AND APPARATUS, DEVICE, MEANS AND PRODUCT OF A PROGRAM FOR GENERATING AN IMAGE

Description

1 / 41 METHOD AND APPARATUS, DEVICE, MEANS AND PRODUCT OF A PROGRAM FOR GENERATING AN IMAGE FIELD

[001] The modalities of this disclosure generally relate to the field of image processing and, more specifically, to the method and apparatus, device, means and program product for generating an image. BACKGROUND

[002] Currently, machine learning is becoming increasingly important in people's daily lives and work, gradually becoming an indispensable tool. People are increasingly processing tasks using machine learning models. For example, text processing tasks, image processing tasks, video processing tasks, and the like are processed using machine learning. In particular, machine learning models with multimodal data processing capabilities offer a significant advantage in processing multimodal data.

[003] With the rapid development of machine learning technology, the procedure for processing diverse multimodal data has become faster and more accurate. For example, in the case of text and image processing, the multimodal machine learning model can be used to assist users in processing operations related to text and images. Furthermore, to meet the development needs of text and image processing technology, the machine learning model has been more broadly applied in image generation. SUMMARY

[004] Modalities of the present disclosure provide a method and apparatus, a device, a means and a program product for generating a Petition 870250090513, dated 03 / 10 / 2025, page 19 / 84 2 / 41 image.

[005] According to a first aspect of the present disclosure, a method is provided for generating an image. The method comprises generating, based on user-generated content, a descriptive text of the content for the user-generated content. The method also comprises generating, based on the descriptive text of the content, a set of template elements of a template for the user-generated content. The method further comprises generating, based on the set of template elements and the user-generated content, a target composite image.

[006] According to a second aspect of the present disclosure, an apparatus is provided for generating an image. The apparatus comprises a content description text generation module configured to generate, based on user-generated content, a content description text for the user-generated content; a model element set generation module configured to generate, based on the content description text, a model element set for the user-generated content; and a target composite image generation module configured to generate, based on the model element set and the user-generated content, a target composite image.

[007] According to a third aspect of the present disclosure, an electronic device is provided, comprising at least one processor; and a memory for storing at least one program, the at least one program, when executed by at least one processor, causes the at least one processor to perform the method according to the first aspect of the present disclosure.

[008] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided for therein. Petition 870250090513, dated 03 / 10 / 2025, page 20 / 84 3 / 41 with computer programs, the computer programs, when executed by a processor, cause the processor to perform the method according to the first aspect of this disclosure.

[009] According to a fifth aspect of this disclosure, a computer program product is provided. The computer program product includes computer programs, computer programs which, when executed by a processor, cause the processor to perform the method according to the first aspect of this disclosure.

[010] It should be noted that the content described in this Summary is not intended to identify key or essential attributes of the features of this disclosure, nor to limit the scope of this disclosure. Other attributes of this disclosure will be more easily understood through the description below. BRIEF DESCRIPTION OF THE DRAWINGS

[011] By means of the detailed description below of the exemplary embodiments of this disclosure, with reference to the accompanying drawings, the above and other objectives, attributes and advantages of this disclosure will become more evident. In the exemplary embodiments of this disclosure, the same reference sign generally indicates the same component.

[012] Figure 1 illustrates a schematic diagram of an exemplary environment in which the device and / or method according to some embodiments of the present disclosure can be implemented; Figure 2 illustrates a schematic diagram of an exemplary method for generating an image according to some embodiments of the present disclosure; Figure 3 illustrates a schematic flowchart of a process for generating an image according to some of the modalities of this disclosure; Petition 870250090513, dated 03 / 10 / 2025, p. 21 / 84 4 / 41 Figure 4 illustrates a schematic diagram of an example for generating an image according to some modalities of the present disclosure; Figure 5 illustrates a schematic diagram of another example to generate an image in accordance with some modalities of the present disclosure; Figure 6 illustrates a schematic diagram of an additional example for generating an image in accordance with some modalities of the present disclosure; Figure 7 illustrates a schematic diagram of an additional example for generating an image in accordance with some embodiments of the present disclosure; Figure 8 illustrates a schematic diagram of a composite image including identification information according to some embodiments of the present disclosure; Figure 9 illustrates a schematic diagram of exemplary positions of user-generated content and the set of model materials for generating an image in accordance with the modalities of this disclosure; Figure 10 illustrates a schematic diagram of a specific modality for generating an image in accordance with some modalities of this disclosure; Figure 11 illustrates a schematic diagram of an example for generating summarized descriptive text according to some modalities of this disclosure; Figure 12 illustrates a schematic diagram of an example of a training procedure for an image generation model to generate an image according to some modalities of the present disclosure; Figure 13 illustrates a schematic diagram of an apparatus for generating an image according to some embodiments of the present disclosure; Figure 14 illustrates a schematic block diagram of an exemplary device adapted to implement a plurality of modalities of Petition 870250090513, dated 03 / 10 / 2025, p. 22 / 84 5 / 41 present disclosure.

[013] In each drawing, the same reference mark or a corresponding reference mark indicates the same component or a corresponding component. DETAILED DESCRIPTION OF THE MODALITIES

[014] It should be understood that the data involved in the technical solutions of this disclosure, including, but not limited to, the data itself, and the acquisition or use of the data, must comply with the requirements of the relevant laws, regulations and rules.

[015] It should be noted that, prior to the use of the technical solutions disclosed through various methods of this disclosure, the type, scope of use and application scenario of the personal information involved in this disclosure are disclosed to users through appropriate means, in accordance with relevant laws and regulations, to obtain the user's authorization.

[016] For example, in response to receiving an active request from users, a prompt information is sent to users to clearly inform them that the requested operation needs to obtain and use their personal information. Consequently, users can voluntarily select, according to the prompt information, whether they wish to provide their personal information to the software or hardware that performs the operations of the technical solution, such as an electronic device, application program, server, or storage medium.

[017] As an optional and non-restrictive implementation, in response to receiving an active request from users, a prompt is sent to users, where the prompt may be in the form of a pop-up window, for example, and the prompt may be displayed as text in the pop-up window. Furthermore, the pop-up window may also Petition 870250090513, dated 03 / 10 / 2025, page 23 / 84 6 / 41 be provided with a selection control through which users can choose to “agree” or “disagree” to the provision of personal information to the electronic device.

[018] It should be noted that the above procedure for informing users and obtaining user authorization is merely illustrative and does not restrict implementations of this disclosure. Other methods may also be applied to implementations of this disclosure, provided they comply with relevant regulations and laws.

[019] The embodiments of this disclosure will be described below in more detail, with reference to the drawings. Although the drawings illustrate some embodiments of the present disclosure, it should be appreciated that the present disclosure can be implemented in various ways and should not be limited to the embodiments explained in the present invention. Rather, the embodiments are provided for a more complete and thorough understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are provided for illustrative purposes only, rather than restricting the scope of protection of the present disclosure.

[020] In describing the modalities of this disclosure, the term “includes” and its variants should be read as open terms meaning “includes, but is not limited to”. The term “based on” should be read as “based at least in part on”. The term “a modality” or “this modality” should be read as “at least one exemplary modality”. The terms “first”, “second”, and so forth may refer to the same object or to different objects. The following text may also include other explicit and implicit definitions.

[021] There are still many problems to be addressed during image generation. For example, usually when users want to generate Petition 870250090513, dated 03 / 10 / 2025, p. 24 / 84 7 / 41 Individualized content based on existing information (such as user-generated content, which is multi-mode and usually video or image), they need to collect related information and process it on their own. Sometimes, they may even need to write related scripts themselves. However, the above process, which is very time-consuming and resource-intensive, is relatively inefficient and impacts the user experience.

[022] For example, in traditional solutions, users can mine different types of user-generated content (UGC), such as video creation content, image creation content, audio creation content, and text creation content, etc. Users then further process the materials into the various types of user-generated content and then combine the processed materials into a new material that can be placed on the market. In this solution, it is usually necessary to manually mine user content, extract the mined user-generated content, find suitable and available materials, and position them in a predefined fixed model.

[023] In general, the fixed template includes background content, product identifier, and script text, etc. An “empty slot” is left for materials to be replaced in a fixed position, and the slot is often fixed in the fixed template. The fixed template is made manually and in limited quantities. In this case, it is impossible to satisfy the need to release a large quantity of materials. The template, therefore, is monotonous and unattractive for other uses. Additionally, the generated fixed template is generally irrelevant to the mined materials. Frequently, the same fixed template includes materials of different styles and types. Therefore, a sense of distance is observed between the fixed template and the materials in the “fixed slot.” It is impossible to provide users with an individualized match, and the user experience is quite Petition 870250090513, dated 03 / 10 / 2025, page 25 / 84 8 / 41 impacted.

[024] To at least solve the above and other potential problems, the embodiments of the present disclosure provide a method for generating an image. In this method, a content description text for the user-generated content can be initially generated on the computing device where the user-generated content is created by the user. For example, the user-generated content is video or image or a combination thereof. The content description text is obtained by applying the user-generated content to a machine learning model. Then, the computing device also processes the content description text to generate a set of model elements for the user-generated content. Finally, the computing device generates a target composite image with the generated set of model elements and the user-generated content.In this method, a set of template elements associated with user-generated content is created based on the content generated from the user's creation. The associated user-generated content and the set of template elements are further combined, so that the generated target composite image is more accurate and relevant. Therefore, the efficiency of image generation is improved and the user experience is enhanced.

[025] Embodiments of the present disclosure are described in detail below with reference to the drawings, wherein Figure 1 illustrates an exemplary environment in which the device and / or method of the embodiments of the present disclosure can be implemented. In environment 100, computing device 102 first generates a descriptive text of the content 106 for user-generated content 104, wherein user-generated content is created by the user. For example, user-generated content 104 is a Petition 870250090513, dated 03 / 10 / 2025, page 26 / 84 9 / 41 video, or an image, or a combination thereof. The descriptive text of the content 106 is obtained by processing the user-generated content 104, such as applying the user-generated content 104 to a machine learning model. Then, the computing device 102 generates a set of model elements 108 from a model for the user-generated content 104 with the descriptive text of the content 106. Finally, after determining the set of model elements 108, the computing device 102 additionally combines it with the user-generated content 104 to generate a target composite image 110.

[026] Examples of computing devices 102 include, but are not limited to, personal computers, server computers, portable devices or laptops, mobile devices (such as cell phones, Personal Digital Assistants (PDAs), media players and the like), multiprocessor systems, consumer electronics, minicomputers, mainframe computers and a distributed computing environment, including any of the above systems or devices, etc.

[027] As shown in Figure 1, the computing device 102 can first generate the descriptive content text 106 for the user-generated content 104, wherein the user-generated content 104 is created by the user. For example, the user-generated content 104 is a video, an image, or a combination thereof. In one example, the user-generated content 104 might be a video segment, for example, a 10-second video created by the user. In another example, the user-generated content 104 might be an image, for example, an image created by the user through gathering, cropping, and merging, among other operations. In yet another example, the user-generated content 104 might be a combination of video and image. For example, the user creates the user-generated content 104 by merging and combining the video and image using editing software. Petition 870250090513, dated 03 / 10 / 2025, page 27 / 84 10 / 41 Alternatively, user-generated content can also be audio or a combination of audio and video.

[028] In some modalities, the descriptive text of the content 106 includes information extracted from the user-generated content 104, and the information may include one or more of the following: what is primarily described by the user-generated content 104, what is the style of the user-generated content 104, and what is the theme color of the user-generated content 104, and so on. In addition, the theme of the user-generated content 104 may be extracted and determined according to the descriptive text of the content 106. For example, the theme of the user-generated content 104 may be determined as live music streaming, football match and game entertainment, etc.

[029] In some embodiments, the descriptive text of content 106 is obtained by applying user-generated content 104 to the machine learning model. For example, the descriptive text of content 106 is obtained by applying user-generated content 104 to a visual model. Additionally, the descriptive text of content 106 is obtained by applying user-generated content 104 to a large-scale visual model. It should be understood that the examples are provided merely to describe the present disclosure, rather than restricting it.

[030] Next, the computing device 102 generates a set of template elements 108 from the template for the user-generated content 104 with descriptive text of the content 106. In some embodiments, a set of template elements 108 includes only background images. In some other embodiments, a set of template elements 108 includes background images and stickers. In some additional embodiments, a set of template elements 108 includes background images and summarized descriptive text. Petition 870250090513, dated 03 / 10 / 2025, page 28 / 84 11 / 41 In some additional embodiments, a set of 108 template elements includes background images, stickers, and brief descriptive text. Additionally, a set of 108 template elements may also include a product identifier, company identifier, or username watermark, etc. For ease of description and understanding, the brief descriptive text is also referred to as a script.

[031] In some embodiments, the computing device can process the descriptive text of the content 106 with an image generation model, so as to generate background images and stickers in the model element set 108. In some other embodiments, the computing device can process the descriptive text of the content 106 with a large language model to generate a script for the user-generated content 104.

[032] After determining the set of template elements 108, the computing device 102 additionally combines it with the user-generated content 104 to generate the target composite image 110. In some embodiments, the computing device generates more than one sticker and more than one script and combines a plurality of stickers with a plurality of scripts according to a predetermined position or a predefined rule to generate a set of template elements 108.

[033] Through this method, a set of template elements associated with user-generated content is generated based on the content generated from the user's creation. The associated user-generated content and the set of template elements are further combined, so that the generated target composite image is more accurate and relevant. Therefore, the efficiency of image generation is improved and the user experience is enhanced. Petition 870250090513, dated 03 / 10 / 2025, p. 29 / 84 12 / 41

[034] The schematic diagram of an exemplary environment in which the device and / or method according to some embodiments of the present disclosure can be implemented was described above with reference to Figure 1. Next, a schematic diagram of an exemplary method for generating an image according to some embodiments of the present disclosure is depicted below with reference to Figure 2. The method in Figure 2 can be executed by the computing device 102 in Figure 1 and / or any suitable computing devices.

[035] As shown in Figure 2, in exemplary method 200, computing device 102 generates descriptive text of content 106 for user-generated content 104 based on user-generated content 104, wherein user-generated content 104 is created by the user. User-generated content 104 can be video or image or a combination thereof.

[036] In some embodiments, the content description text 106 is obtained by applying the user-generated content 104 to the machine learning model. For example, the content description text 106 is obtained by applying the user-generated content 104 to a visual model. Additionally, the content description text 106 is obtained by applying the user-generated content 104 to a large-scale visual model. In some embodiments, there is a predetermined mapping relationship between the user-generated content and the content description text. After obtaining the user-generated content 104, the content description text 106 corresponding to the user-generated content can be obtained from the predetermined mapping relationship. The above examples are provided only to describe the present disclosure, rather than restricting it.

[037] Next, in block 204, computing device 102 generates a Petition 870250090513, dated 03 / 10 / 2025, page 30 / 84 13 / 41 set of model elements 108 of the model for the user-generated content 104 based on the descriptive text of the content 106. In order to best display the user-generated content, it is necessary to determine the model combined with the user-generated content according to the descriptive text of the content 106 determined based on the user-generated content. The model consists of a set of model elements.

[038] In some embodiments, the computing device can process the content description text 106 with the image generation model, to generate background images and stickers in the model element set 108. In addition, the computing device processes the content description text 106 with the machine learning model to generate a prompt information for the background image and a prompt information for the sticker. For example, the machine learning model is a large language model. Then, the prompt information for the background image and the prompt information for the sticker are fed into the image generation model to generate the background image and the sticker. Alternatively, the computing device 102 can also obtain a mapping relationship between the description text and the background image and a mapping relationship between the description text and the sticker in advance.After obtaining the descriptive text of content 106, the computing device 102 obtains the background image corresponding to the descriptive text of content 106 according to the mapping relationship between the descriptive text and the background image. The computing device 102 can also obtain the sticker corresponding to the descriptive text of content 106 according to the mapping relationship between the descriptive text and the sticker. The above examples are provided only to describe the present disclosure, rather than restrict it.

[039] In some forms, the computing device may Petition 870250090513, dated 03 / 10 / 2025, page 31 / 84 14 / 41 process the descriptive text of the content 106 with the large language model, to generate the script for the user-generated content 104. In some embodiments, the computing device 102 may obtain a mapping relationship between the descriptive text and the script. After obtaining the descriptive text of the content 106, the computing device 102 may search for the script corresponding to the descriptive text using this mapping relationship. The above examples are provided only to describe the present disclosure, rather than restrict it.

[040] Finally, in block 206, computing device 102 generates, based on the model element set 108 and the user-generated content 104, the target composite image 110. The model element set 108, after being obtained, can be appropriately combined with the user-generated content 104 to generate the target composite image.

[041] In some embodiments, the target composite image 110 only includes the background image and the user-generated content 104, and the user-generated content 104 is in front of the background image. In this case, the computing device can adjust the size of the user-generated content 104 according to a predefined ratio. For example, the computing device adjusts the user-generated content 104 to occupy 40% of the size of the background image. Furthermore, at this point, the computing device can position the user-generated content in any suitable position on the background image.

[042] In some embodiments, the target composite image 110 includes a background image, a sticker, and user-generated content 104, and the user-generated content 104 and the sticker are also in front of the background image. In this case, the computing device can position the background image and the sticker according to the position information. Petition 870250090513, dated 03 / 10 / 2025, page 32 / 84 15 / 41 predefined rules from the template element set 108 are used to generate the composite image with the user-generated content 104. In one example, the sticker and the user-generated content 104 do not touch in the background image. In another example, the sticker and the user-generated content 104 partially touch each other in the background image. For example, part of the sticker covers the user-generated content 104 and is displayed in front of the user-generated content 104. Additionally, the computing device can also position the background image and the sticker according to the predefined rule of the template element set 108. For instance, the user can set the sticker to cover the user-generated content 104. The threshold ratio of the user-generated content 104 covered by the sticker can be set to 20%.

[043] In some embodiments, the target composite image 110 includes the background image, the summary descriptive text, and the user-generated content 104, and the user-generated content 104 and the summary descriptive text are also in front of the background image. In this case, the computing device can position the background image and the summary descriptive text according to the predefined positioning information of the model element set 108 and generate the composite image with the user-generated content 104. It should be noted that there may be multiple summary descriptive texts scattered in front of the background image. In one example, the summary descriptive text does not contact the user-generated content. In another example, the summary descriptive text and the user-generated content partially contact each other.In addition, the computing device can also position the background image and the summary descriptive text according to the predefined rule of the 108 model element set. For instance, the user can define the reason for the generated content. Petition 870250090513, dated 03 / 10 / 2025, page 33 / 84 16 / 41 by user 104 covered by the summarized descriptive text. The threshold ratio of content generated by user 104 covered by the summarized descriptive text can be defined as 10%.

[044] In some embodiments, the target composite image 110 includes the background image, the sticker, the script, and the user-generated content, and the user-generated content 104, the sticker, and the script are also in front of the background image. In this case, the computing device can arrange the user-generated content 104, the sticker, and the script according to the predefined position information of the model element set 108 and / or the predefined rule, and the computing device can position the user-generated content 104, the sticker, and the script based on the predefined position information. The computing device can also position the user-generated content 104, the sticker, and the script according to the predefined rule. For example, the user-generated content 104, the sticker, and the script can be placed in contact with each other according to a predefined ratio.Additionally, the user can also define the display priority of the user-generated content 104, the sticker, and the script. For example, the script's display priority is set to the highest. In this case, the script will never come into contact with the user-generated content 104 and / or the sticker and will always be displayed in front of the composite image. The example above is an example of the present disclosure and not a restriction.

[045] In some embodiments, the set of model elements 108 also includes identification information, in addition to the user-generated content 104, the sticker and the script, and the identification information includes the product identifier, the company identifier or the user name watermark etc. In one example, the user can define the transparency of Petition 870250090513, dated 03 / 10 / 2025, page 34 / 84 17 / 41 Identification information. For example, the transparency of the identification information is set to 50%. In another example, the user can display the identification information in bold or highlight it. In a further example, the user can define the positioning information of the identification information. For example, the identification information is positioned in the four corners of the composite image. Additionally, the user can configure that any other elements, except the background image, are not allowed within a certain range centered on the identification information. For example, the presence of any elements other than the background image is not allowed within a range defined by diffusing the identification information outwards by 50%.

[046] Through this method, a set of model elements associated with user-generated content is generated based on the user-created content. The associated user-generated content and the set of model elements are further combined, so that the generated target composite image is more accurate and relevant. Therefore, the efficiency of image generation is improved and the user experience is enhanced.

[047] The schematic diagram of an exemplary method for generating an image according to some embodiments of the present disclosure was described above with reference to Figure 2. Next, a schematic flowchart of a process for generating an image according to some embodiments of the present disclosure should be depicted together with Figure 3. The example in Figure 3 can be executed by computing device 102 in Figure 1 or any suitable computing devices.

[048] In example 300 shown in Figure 3, computing device 102 first generates descriptive content text 304 for the Petition 870250090513, dated 03 / 10 / 2025, page 35 / 84 18 / 41 User-generated content 302, and user-generated content 302 is created by the user. User-generated content 302 is a video or an image, or a combination thereof. The descriptive text of content 304 is obtained by applying user-generated content 302 to the machine learning model. Additionally, the machine learning model is a large-scale visual model.

[049] In some embodiments, after applying the user-generated content 302 to the large-scale visual model, the computing device obtains various extracted information for the user-generated content 302. For example, in some embodiments, the descriptive text of the content 304 includes extracted information for the user-generated content 302, and the information includes at least one of the following: what is primarily described by the user-generated content 302, what is the style of the user-generated content 302, and what is the theme color of the user-generated content 302, and so on. In addition, the theme of the user-generated content 302 can be extracted and further determined according to the descriptive text of the content 304. For example, the theme of the user-generated content 302 can be determined as live music streaming, gaming entertainment, and football matches, etc.

[050] After obtaining the various information above, the computing device additionally generates the summarized descriptive text 306 based on the information contained in the content descriptive text 304. For ease of description, the summarized descriptive text 306 is also referred to as a script. The script is generated by extracting the text information for the user-generated content in the content descriptive text. The computing device can generate the text information using the large language model. Additionally, the large language model is part of the visual model in Petition 870250090513, dated 03 / 10 / 2025, page 36 / 84 19 / 41 large scale and the summarized descriptive text 306 is an element in the set of model elements.

[051] Meanwhile, the computing device also generates image prompt information 308 in accordance with the descriptive content text 304, wherein image prompt information 308 includes first image prompt information and second image prompt information.

[052] In some modalities, the first image prompt information includes: the text for the main description of the user-generated content 302, the text for the style of the user-generated content 302, and the text for the theme color of the user-generated content 302, among other text information. In addition, the text for the region or date of the user-generated content 302 may also be obtained. The second image prompt information includes text for the theme of the user-generated content 302. For example, the theme of the user-generated content 302 may be determined as live music streaming, football game, and game entertainment, etc.

[053] Next, the computing device applies the image prompt information 308 to the image generation model 310 to further generate the background image 312 and the sticker 318. Both the background image 312 and the sticker 318 are elements in the model element set. Additionally, the image generation model 310 is a diffusion model. Furthermore, the image generation model 310 is a Stable Diffusion model. It should be understood that this is only an example and does not restrict the present disclosure.

[054] In some modes, the background image 312 is generated by the computing device according to the information from the first prompt. Petition 870250090513, dated 03 / 10 / 2025, page 37 / 84 20 / 41 image. The background image 312 is always behind the set of template elements and the user-generated content 302, and is an indispensable part of the final composite image. In other words, the composite image consists of at least the background image and the user-generated content 302. Furthermore, the background image is strongly correlated with the user-generated content.

[055] In some embodiments, sticker 318 is generated by the computing device according to the information in the second image prompt. Specifically, sticker 318 is generated according to the text for the theme of the user-generated content 302 in the information in the second image prompt. In one example, when the theme of the user-generated content 302 is determined to be live music streaming, the sticker might be an image related to singing or music. In another example, when the theme of the user-generated content 302 is determined to be a football game, the sticker might be an image related to football or a match. In a further example, when the theme of the user-generated content 302 is determined to be gaming entertainment, the sticker might be an image related to games or esports.

[056] After the computing device determines the summary descriptive text 306, the background image 312, and the sticker 318, the three elements above are determined as a set of model elements. In addition, the computing device may also determine other model elements, considering the requirements, and add them to the set of model elements above. Subsequently, the computing device may apply the set of model elements above to the layout calculation 314 to determine a plurality of candidate composite images.

[057] In some modalities, the three elements above may be Petition 870250090513, dated 03 / 10 / 2025, page 38 / 84 21 / 41 are positioned according to predefined positioning information, such as a first plurality of positions. Additionally, user-generated content also has predefined positioning information, for example, a second plurality of positions. The computing device positions the three elements and the user-generated content according to their respective positioning information.

[058] In some modes, the three elements above can also be positioned according to the predefined rule. For example, the size of user-generated content 104 is adjusted according to a predefined ratio. For example, user-generated content 302 is adjusted to occupy 40% of the background image size. Furthermore, at this point, the computing device can position the user-generated content in any suitable position on the background image.

[059] In some embodiments, a portion of sticker 318 covers user-generated content 302 and is displayed in front of user-generated content 302. Additionally, the user can define the ratio of user-generated content 302 covered by the sticker. For instance, the user can define a threshold ratio of user-generated content 104 covered by the sticker as being 20%.

[060] In some embodiments, the computing device determines a set of candidate composite images, including user-generated content 302 and the three elements according to a plurality of predefined position information and a plurality of predefined rules, for example, the positioning positions and positioning rules described above, etc. The set of candidate composite images includes a plurality of candidate composite images, and the score of each set of candidate composite images can be determined by a model of Petition 870250090513, dated 03 / 10 / 2025, page 39 / 84 22 / 41 score.

[061] In some modes, the user can define a threshold score. In this case, after the layout calculation 314, the candidate composite image is determined as the target composite image 316 if its score reaches or exceeds the threshold score. Additionally, if a plurality of candidate composite images has a score that reaches or exceeds the threshold score, the candidate composite image with the highest score among the plurality of candidate composite images is determined as the target composite image 316.

[062] In this method, a set of template elements associated with user-generated content is generated with the user-created content. The associated user-generated content and the set of template elements are further combined, so that the generated target composite image is more accurate and relevant. Therefore, the efficiency of image generation is improved and the user experience is enhanced.

[063] The schematic flowchart of a process for generating an image according to some embodiments of the present disclosure was described above with reference to Figure 3. Next, a schematic diagram of an example for generating an image according to some embodiments of the present disclosure should be depicted with reference to Figure 4.

[064] In example 400, background image 402 and user-generated content 404 form the target composite image. In this example, the target composite image only includes background image 402 and user-generated content 404.

[065] In some modes, user-generated content 404 is created by the user. User-generated content 404 is a video or an image or a combination thereof. Additionally, the generated content Petition 870250090513, dated 03 / 10 / 2025, page 40 / 84 23 / 41 by user 404 may be a video generated from the recording of the live stream.

[066] In some embodiments, the computing device can adjust the size of the user-generated content 404 according to a predefined ratio. For example, the computing device adjusts the user-generated content 404 to occupy 40% of the background image size. Furthermore, at this point, the computing device can position the user-generated content in any suitable position on the background image.

[067] The schematic diagram of an example for generating an image according to some embodiments of the present disclosure was described above with reference to Figure 4. Next, a schematic diagram of another example for generating an image according to some embodiments of the present disclosure is depicted below with reference to Figure 5.

[068] In example 500, based on the previous example 400, in addition to the background image 502 and the user-generated content 508, sticker 504 and sticker 506 are added. The size and shape of sticker 504 and sticker 506 are defined voluntarily by the user and both are positioned in front of the background image.

[069] In some modes, there is at least one sticker, and the sticker can be positioned according to a predefined position. Additionally, the sticker can also be positioned according to a predefined rule. In one example, the sticker and the user-generated content do not touch in the background image. In another example, the sticker and the user-generated content partially touch each other in the background image. For example, part of the sticker covers the user-generated content and is displayed in front of the user-generated content. The user can also define the reason for the Petition 870250090513, dated 03 / 10 / 2025, page 41 / 84 24 / 41 user-generated content 104 covered by the sticker. For instance, the user can set the threshold ratio of user-generated content 104 covered by the sticker to 20%.

[070] The schematic diagram of another example for generating an image according to some embodiments of the present disclosure was described above with reference to Figure 5. Next, a schematic diagram of an additional example for generating an image according to some embodiments of the present disclosure is depicted below with reference to Figure 6.

[071] In example 600, based on the previous example 400, in addition to the background image 602 and the user-generated content 606, the summary descriptive text 604 and the summary descriptive text 606 are added. For ease of description, the summary descriptive text 604 is referred to as script 1 and the summary descriptive text 606 is referred to as script 2.

[072] In some modes, neither script 1 nor script 2 comes into contact with the user-generated content, and script 1 and script 2 are positioned on the background image according to the predefined position.

[073] In some embodiments, script 1 and / or script 2 come into contact with user-generated content and partially cover it. The computing device can define the ratio of user-generated content covered by the script, considering the predefined rule. For example, the computing device can define the threshold ratio of user-generated content covered by the script as 10%.

[074] The schematic diagram of a further example for generating an image according to some embodiments of the present disclosure was described above with reference to Figure 6. Next, a schematic diagram of a further example for generating an image according to some Petition 870250090513, dated 03 / 10 / 2025, page 42 / 84 25 / 41 of the modalities of this disclosure should be depicted below with reference to Figure 7.

[075] In the example, based on previous examples 400, 500 and 600, in addition to background image 702 and user-generated content 710, the summary descriptive text 704, the sticker 706, the sticker 708 and the summary descriptive text 712 are positioned in front of background image 702. To facilitate description, the summary descriptive text 704 is referred to as script 1 and the summary descriptive text 712 is referred to as script 2.

[076] In some embodiments, the computing device may position the user-generated content 710, the sticker 706, the sticker 708, the script 1, and the script 2 according to predetermined positioning information. The computing device may also position the user-generated content, the sticker, and the script according to a predetermined rule. For example, the user-generated content, the sticker, and the script are positioned according to the predefined rule. The sticker's border is displayed in bold, and the script is highlighted, for instance. Additionally, the user may also define the display priority of the user-generated content, the sticker, and the script. For example, the script's display priority is set to the highest. In this case, the script will never come into contact with the user-generated content and / or the sticker and will always be displayed in front of the composite image. The above example is an example of the present disclosure and not a restriction.

[077] The schematic diagram of a further example for generating an image according to some embodiments of the present disclosure was described above with reference to Figure 7. Next, a schematic diagram of a composite image including identification information according to some embodiments of the present disclosure is depicted below with reference to Figure 8. Petition 870250090513, dated 03 / 10 / 2025, page 43 / 84 26 / 41

[078] Based on the previous example 700, an identifier 804 is also included in example 800, in addition to the background image 802, the short descriptive text 806, the sticker 808, the sticker 810, the user-generated content 812, and the short descriptive text 814. The identifier 804 includes a product identifier, a company identifier, or a username watermark, etc.

[079] In some modes, the user can set the transparency of the identification information, for example, to 50%. In another example, the user can display the identification information in bold or highlight it. In a further example, the user can set the positioning of the identification information. For example, the identification information is positioned in the four corners of the composite image. In addition, the user can configure that any other elements, except the background image, are not allowed within a certain range centered on the identification information. For example, the presence of any elements other than the background image is not allowed within a range defined by diffusing the identification information outwards by 50%.

[080] In some embodiments, the computing device sets the position of identifier 804 to have the highest priority. That is, any other elements in the composite image would not cover identifier 804, and identifier 804 may appear in front of the composite image in any case.

[081] The schematic diagram of a composite image including identification information according to some embodiments of the present disclosure was described above in Figure 8. Next, a schematic diagram of exemplary positions of the user-generated content and the Petition 870250090513, dated 03 / 10 / 2025, page 44 / 84 27 / 41 The set of model materials for generating an image in accordance with some modalities of the present disclosure should be depicted below with reference to Figure 9.

[082] Example 900 consists of background image 902, identifier 904, sticker 906, short descriptive text 908, sticker 910, user-generated content 912, sticker 914 and short descriptive text 916.

[083] In some embodiments, the position of identifier 904 is fixed and identifier 904 is positioned according to a predefined range of positions. In one example, the computing device positions identifier 904 within 5% of the upper left corner of background image 902. In another example, identifier 904, after being positioned, does not make contact with any edge of the background image.

[084] In some embodiments, it is positioned according to the predefined position and displayed in front of the background image. In some embodiments, the summary descriptive text 908 and the sticker 910 are positioned according to the predefined rule. For example, part of the summary descriptive text 908 covers the sticker 910, and the summary descriptive text 908 is always displayed in front of the sticker 910. Additionally, the ratio of sticker 910 covered by the summary descriptive text 908 does not exceed a predetermined threshold, such as 30% of the display area of ​​sticker 910.

[085] In some embodiments, the 914 sticker may also cover user-generated content 912. Similar to that described above, the proportion of user-generated content 912 covered by the 914 sticker must also not exceed a predetermined threshold, such as 5% of the display area of ​​the user-generated content.

[086] In some modalities, the textual content of the descriptive text Petition 870250090513, dated 03 / 10 / 2025, page 45 / 84 The summarized descriptive text 908 and the summarized descriptive text 916 may be the same and displayed according to different display styles. In some versions, the summarized descriptive text 908 is displayed in bold and highlighted, while the summarized descriptive text 916 is displayed in an artistic font.

[087] In some forms, the textual content of the summarized descriptive text 908 and the summarized descriptive text 916 may be different. Additionally, the content of the summarized descriptive text 908 and the summarized descriptive text 916 must be strongly correlated. For example, the content summary must address the same topic or include the same keyword.

[088] The schematic diagram of exemplary positions of user-generated content and the set of model materials for generating an image according to some embodiments of the present disclosure was described above with reference to Figure 9. Next, a schematic diagram of a specific embodiment for generating an image according to some embodiments of the present disclosure is depicted below with reference to Figure 10.

[089] Example 1000 includes background image 1102 and user-generated content 1014, wherein user-generated content 1014 is a content segment in which an anchor explains specific details of a music festival and mentions that there will be mystery guests at the festival. The music festival is sponsored and held by sponsors and organizers. Additionally, background image 1002 is generated according to the prompt information from the first image for user-generated content 1014.

[090] In some modalities, the identifier 1004 is the identifier of the sponsor and organizer. In some other modalities, the identifier 1004 is the watermark of the anchor's username. Petition 870250090513, dated 03 / 10 / 2025, p. 46 / 84 29 / 41

[091] In some modes, the transparency of identifier 1004 can be defined. For example, the transparency of identifier 1004 is set to 50%. Identifier 1004 can also be displayed in bold or highlighted. The positioning information of identifier 1004 can also be configured. For example, identifier 1004 is positioned in the four corners of the composite image. Additionally, the user can configure that any other elements, except the background image, are not allowed within a certain centered range of the identification information. For example, the presence of any elements other than the background image is not allowed within a range defined by diffusing the identification information by 50%.

[092] In some modalities, the summary descriptive information 1008 states that “the mystery guest will make a surprise appearance, and please stay tuned”. This summary descriptive information is generated from the summary of a summary description of user-generated content using the large language model. Similarly, the summary descriptive information 1018, which says “cheerful music festival”, is also generated based on the same large language model. Furthermore, the summary descriptive information 1008 and the summary descriptive information 1018 are correlated and can be obtained directly or indirectly from user-generated content.

[093] In some modalities, sticker 1016 is a CD; sticker 1010 is a speaker; sticker 1012 is a random playback identifier; and sticker 1016 is a note. The stickers above are all generated according to the prompt information in the second image for user-generated content 1014. In addition, they are generated based on the theme information in the prompt information in the second image, and the theme information. Petition 870250090513, dated 03 / 10 / 2025, page 47 / 84 30 / 41 include “music festival” and “music”, etc.

[094] The schematic diagram of a specific embodiment for generating an image in accordance with some embodiments of the present disclosure was described above with reference to Figure 10. A schematic diagram of an example of generating corresponding summary descriptive text in accordance with user-generated content in accordance with some embodiments of the present disclosure is depicted below with reference to Figure 11.

[095] In example 1100, a scene from user-generated content 1102 shows a pet dog celebrating its birthday smiling at the camera.

[096] In some embodiments, the computing device applies the user-generated content 1102 to the large-scale visual model and then interprets the user-generated content 1102 with the large-scale visual model to obtain the descriptive text of the user-generated content 1102. The text for the main description of the user-generated content 1102, the text for the style of the user-generated content 1102, and the text for the theme color of the user-generated content 1102, among other text information, are additionally obtained from the descriptive text of the generated content. In addition, the text for the region or date of the user-generated content 1102 can also be obtained.

[097] Then, the computing device extracts keywords from the various types of textual information obtained and provides the appropriate summarized descriptive information for the current scene and relevant to the target object that appears in the current scene, in view of the scene.

[098] For example, script 1104 which says “Every pet deserves a big stage” is generated based on user-generated content 1102. Petition 870250090513, dated 03 / 10 / 2025, page 48 / 84 31 / 41

[099] In some embodiments, the computing device can also generate a plurality of candidate scripts with the large-scale visual model and classify the plurality of candidate scripts via the scoring model. The script with a score that reaches or exceeds the threshold score among the plurality of candidate scripts is determined as the target script 1104. Additionally, the semantic and keyword integrity in the generated candidate scripts can serve as benchmarks for classifying the candidate scripts.

[100] The schematic diagram of an example of generating a corresponding summary descriptive text according to user-generated content according to some embodiments of the present disclosure was described above with reference to Figure 11. Next, a schematic diagram of an example of a training procedure for an image generation model to generate an image according to some embodiments of the present disclosure is depicted below with reference to Figure 12.

[101] In example 1200, the image generation model is trained. The computing device can first obtain sample image prompt information 1202. Sample image prompt information 1202 can be directed to the background sample image. Sample image prompt information is text information and can be adjusted by the user according to requirements.

[102] The computing device then applies the sample image prompt information 1202 to the image generation model 1206 to generate a predicted image 1208. The predicted image 1208 may be the generated background image. After generating the predicted image 1208, the computing device further compares the predicted image 1208 with the sample image 1204 to determine the differences between them. The computing device Petition 870250090513, dated 03 / 10 / 2025, page 49 / 84 32 / 41 additionally adjusts the image generation model parameters to complete the image generation model training. Additionally, when it is necessary to generate the sticker using the image generation model, the above approach can also be applied to the sticker to train the image generation model. In some embodiments, the user can train the image generation model based on different sample images and sample image prompt information for the sample images, in order to enhance the image generation model's ability to handle different image types.

[103] The schematic diagram of an example of a training procedure for an image generation model to generate an image according to some modalities of the present disclosure was described above with reference to Figure 12. Next, a schematic block diagram of an apparatus 1300 to generate an image according to some modalities of the present disclosure is depicted below with reference to Figure 13.

[104] As shown in Figure 13, the device 1300 comprises a content description text generation module 1302 configured to generate, based on user-generated content, a content description text for the user-generated content; a model element set generation module 1304 configured to generate, based on the content description text, a model element set for the user-generated content; and a target composite image generation module 1306 configured to generate, based on the model element set and user-generated content, a target composite image.

[105] In some modes, where the model element set includes a background image, where module 1304 is generated Petition 870250090513, dated 03 / 10 / 2025, page 50 / 84 33 / 41 model elements include: a module to generate initial image prompt information configured to generate, based on the descriptive text of the content, initial image prompt information for the user-generated content; and a module to generate background image configured to generate, based on the initial image prompt information, the background image for the model.

[106] In some embodiments, where the set of model elements also includes at least one of: a summary descriptive text or a sticker, where module 1304 for generating model elements also includes at least one of: a module for generating summary descriptive text configured to generate, based on the descriptive text of the content, the summary descriptive text for the model; or a module for generating stickers configured to generate, based on the descriptive text of the content, the sticker for the model.

[107] In some embodiments, the sticker generation module includes: a module to generate second image prompt information configured to generate, based on the descriptive text of the content, second image prompt information for the user-generated content; and the sticker generation module is configured to generate, based on the second image prompt information, the sticker for the template.

[108] In some embodiments, where the first image prompt information includes at least one of: background image content or background image color.

[109] In some embodiments, the module for generating the background image includes: a module for applying the image generation model configured to generate the background image by applying the first image prompt information in an image generation model, the image generation model Petition 870250090513, dated 03 / 10 / 2025, page 51 / 84 34 / 41 being a diffusion model.

[110] In some embodiments, where a module for training the image generation model includes: a module to obtain sample image prompt information and sample image configured to obtain sample image prompt information and a sample image; a module to obtain predicted image configured to obtain a predicted image by applying the sample image prompt information to the image generation model; and a module to adjust parameters configured to adjust image generation model parameters based on the sample image and the predicted image.

[111] In some embodiments, the target composite image generation module 1306 includes: a module to generate a set of candidate composite images configured to generate a set of candidate composite images based on the set of model elements and user-generated content; and a module to select the target composite image configured to select the target composite image from the set of candidate composite images.

[112] In some embodiments, the module for generating a set of candidate composite images includes: a module for determining a first plurality of positions and a second plurality of positions configured to determine a first plurality of positions available for positioning model elements in the set of model elements and a second plurality of positions available for positioning user-generated content; and a module for generating a set of candidate composite images configured to generate the set of candidate composite images by positioning the model elements respectively in the first plurality of positions and positioning the user-generated content respectively in the Petition 870250090513, dated 03 / 10 / 2025, page 52 / 84 35 / 41 second plurality of positions.

[113] In some embodiments, a candidate composite image set includes: a module to determine a plurality of predefined rules configured to determine a plurality of predetermined rules for positioning the model element set and user-generated content; and a module to generate a candidate composite image set configured to generate the candidate composite image set based on the plurality of predetermined rules.

[114] In some embodiments, the target composite image selection module includes: a module to determine a score set configured to determine a score set for the candidate composite image set; and the target composite image selection module configured to select the target composite image from the candidate composite image set based on the score set, where a score of the target composite image exceeds a threshold score.

[115] In some embodiments, module 1302 for generating descriptive text of the content includes: a module for applying a machine learning model configured to obtain descriptive text of the content for the user-generated content by applying the user-generated content to a machine learning model.

[116] In some modes, the user-generated content is an image or a video, and the machine learning model is a visual model.

[117] Figure 14 illustrates a schematic block diagram of an example device 1400 for implementing embodiments of the present disclosure. The computing device 102 in Figure 1 can be implemented by the device 1400. As shown in Figure 14, the device 1400 Petition 870250090513, dated 03 / 10 / 2025, pp. 53 / 84 36 / 41 comprises a central processing unit (CPU) 1401, which can perform various actions and appropriate processing based on computer program instructions stored in read-only memory (ROM) 1402 or computer program instructions loaded into random access memory (RAM) 1403 from storage unit 1408. RAM 1403 can also store all types of programs and data necessary for the operation of device 1400. CPU 1401, ROM 1402, and RAM 1403 are connected to each other via a bus 1404. The input / output (I / O) interface 1405 is also connected to bus 1404.

[118] A plurality of components in the 1400 device are connected to the I / O interface 1405, including: an input unit 1406, such as a keyboard, mouse, and the like; an output unit 1407, for example, various types of displays and speakers, etc.; a storage unit 1408, such as a disk and optical disc, etc.; and a communication unit 1409, such as a network card, modem, wireless transceiver, and the like. The communication unit 1409 allows the 1400 device to exchange information / data with other devices via the computer network, such as the internet, and / or various telecommunications networks.

[119] The procedure and processing described above, such as method 200 and examples 300 and 1200, can be executed by processing unit 1401. For example, in some embodiments, method 200 and examples 300 and 1200 can be implemented as a computer software program tangibly included in the machine-readable medium, for example, storage unit 1408. In some embodiments, the computer program can be partially or fully loaded and / or mounted on the device 1400 via ROM 1402 and / or communication unit 1409. When the computer program is loaded into RAM 1403 and executed by CPU 1401, one or more actions of method 200 and examples 300 and 1200 Petition 870250090513, dated 03 / 10 / 2025, pp. 54 / 84 37 / 41 described above can be implemented.

[120] The present disclosure may be a method, apparatus, system and / or computer program product. The computer program product may include a computer-readable storage medium on which the computer-readable program instructions for performing various aspects of the present disclosure are loaded.

[121] The computer-readable storage medium may be a tangible device that holds and stores instructions used by instruction-executing devices. The computer-readable storage medium may be, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any appropriate combination thereof.More concrete examples of computer-readable storage media (non-exhaustive list) include: laptop disk, hard disk, random access memory (RAM), read-only memory (ROM), programmable and erasable read-only memory (EPROM or flash), static random access memory (SRAM), portable compact disc (CD-ROM) read-only memory, digital versatile disc (DVD), pen drive, floppy disk, mechanical encoding devices, punched card stored with instructions on it, or a projection in a slot, and any appropriate combinations thereof. The computer-readable storage media used in the present invention is not interpreted as transient signals per se, such as radio waves or freely propagating electromagnetic waves, electromagnetic waves propagated via waveguides or other transmission media (such as optical pulses via fiber optic cables), or electrical signals propagated via electrical wires. Petition 870250090513, dated 03 / 10 / 2025, pp. 55 / 84 38 / 41

[122] The described computer-readable program instructions can be downloaded from the computer-readable storage medium for each computing / processing device, or to an external computer or external storage via the Internet, local area network, wide area network, and / or wireless network. The network may comprise copper cable, fiber optic transmission, wireless transmission, router, firewall, switch, network port computer, and / or edge server. The network adapter card or network interface in each computing / processing device receives computer-readable program instructions from the network and forwards them for storage on the computer-readable storage medium of each computing / processing device.

[123] The computer program instructions for performing operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state configuration data, or source code or target code written in any combination of one or more programming languages, wherein programming languages ​​comprise object-oriented programming languages, for example, Smalltalk, C++, and so forth, and traditional procedural programming languages, such as the C language or similar programming languages. The computer-readable program instructions may be implemented wholly on the user's computer, partially on the user's computer as a stand-alone software package, partially on the user's computer and partially on the remote computer, or completely on the remote computer or server.In the case of a remote computer, it can be connected to the user's computer via any type of network, including local area network (LAN) and wide area network. Petition 870250090513, dated 03 / 10 / 2025, pp. 56 / 84 39 / 41 distance (WAN), or to an external computer (e.g., connected via the internet using the internet service provider). In some embodiments, the state information of computer-readable program instructions is used to customize an electronic circuit, for example, a programmable logic circuit, a field-programmable gate array (FPGA), or a programmable logic array (PLA). The electronic circuit may execute computer-readable program instructions to implement various aspects of this disclosure.

[124] Several aspects of the present disclosure are described in the present invention with reference to the flowchart and / or block diagram of the method, apparatus (system) and computer program products, according to the embodiments of the present disclosure. It should be understood that each block of the flowchart and / or block diagram and the combination of several blocks in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[125] Computer-readable program instructions may be provided to the processing unit of a general-purpose computer, dedicated computer, or other programmable data processing devices to manufacture a machine, so that the instructions, when executed by the computer's processing unit or other programmable data processing devices, generate a device to implement functions / actions stipulated in one or more blocks in the flowchart and / or block diagram. Computer-readable program instructions may also be stored on computer-readable storage media and cause the computer, programmable data processing device, and / or other devices to function in a specific manner, such that the computer-readable media stored with Petition 870250090513, dated 03 / 10 / 2025, page 57 / 84 40 / 41 instructions encompass a manufacturing article, including instructions for implementing various aspects of the functions / actions stipulated in one or more blocks of the flowchart and / or block diagram.

[126] Computer-readable program instructions can also be loaded into the computer, other programmable data processing devices, or other devices in order to execute a series of operational steps on the computer, other programmable data processing devices, or other devices to generate a computer-implemented procedure. Therefore, instructions executed on the computer, other programmable data processing devices, or other devices implement functions / actions stipulated in one or more blocks of the flowchart and / or block diagram.

[127] The flowchart and block diagram in the drawings illustrate the architecture, functions, and operations of the system that can be implemented by the computer program system, method, and product according to multiple implementations of the present disclosure. In this sense, each block in the flowchart or block diagram can represent a module, a part of a program segment or code, wherein the module and part of a program segment or code include one or more executable instructions to perform stipulated logical functions. In some alternative implementations, it should be noted that the functions indicated in the block may also occur in a different order than that indicated in the drawings. For example, two successive blocks may be executed in parallel or sometimes in reverse order, depending on the functions involved.It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented by a system based on unique hardware. Petition 870250090513, dated 03 / 10 / 2025, pp. 58 / 84 41 / 41 the execution of stipulated functions or actions, or by a combination of dedicated hardware and computer instructions.

[128] Several embodiments of the present disclosure have been described above, the description above being merely illustrative, not exhaustive, and not limited to the embodiments of the present disclosure. Many modifications and alterations, without departing from the scope and spirit of the various embodiments, are obvious to those skilled in the art. The selection of terms in the text aims to best explain the principles and actual applications of each embodiment and the technical improvements introduced into the market by each embodiment, or to allow those skilled in the art to understand the embodiments of the present disclosure. Petition 870250090513, dated 03 / 10 / 2025, p. 59 / 84

Claims

1 / 5 CLAIMS 1. A method for generating an image, characterized in that it comprises: generating, based on user-generated content, a descriptive text of the content for the user-generated content; generating, based on the descriptive text of the content, a set of model elements of a model for the user-generated content; and generating, based on the set of model elements and the user-generated content, a target composite image.

2. A method according to claim 1, characterized in that the set of template elements includes a background image, and wherein generating a set of template elements from a template to user-generated content comprises: generating, based on the descriptive text of the content, initial image prompt information for the user-generated content; and generating, based on the initial image prompt information, the background image for the template.

3. Method, according to claim 1, characterized in that the set of template elements further comprises at least one of: a summary descriptive text or a sticker, wherein generating a set of template elements from a template for user-generated content further comprises at least one of: generating, based on the descriptive text of the content, the summary descriptive text for the template; or generating, based on the descriptive text of the content, the sticker for the template.

4. Method, according to claim 3, characterized in that generating, based on the descriptive text of the content, the sticker for the Petition 870250090513 model, dated 03 / 10 / 2025, page 80 / 84 2 / 5, comprises: generating, based on the descriptive text of the content, secondary image prompt information for the user-generated content; and generating, based on the secondary image prompt information, the sticker for the model.

5. Method according to claim 2, characterized in that the initial image prompt information comprises at least one of: background image content or a background image color.

6. A method according to claim 2, characterized in that generating, based on the first image prompt information, the background image for the model comprises: generating the background image by applying the first image prompt information to an image generation model, wherein the image generation model is a diffusion model.

7. A method according to claim 6, characterized in that the training of the image generation model comprises: obtaining a sample image prompt information and a sample image; obtaining a predicted image by applying the sample image prompt information to the image generation model; and adjusting parameters of the image generation model based on the sample image and the predicted image.

8. Method, according to claim 1, characterized in that generating, based on the set of model elements and user-generated content, a target composite image comprises: generating a set of candidate composite images based on the set of model elements and user-generated content; and selecting the target composite image from the set of candidate composite images.

9. A method according to claim 8, characterized in that generating a set of candidate composite images based on the set of model elements and user-generated content comprises: determining a first plurality of available positions for placing a model element in the set of model elements and a second plurality of available positions for placing user-generated content; and generating the set of candidate composite images by placing the model element respectively in the first plurality of positions and placing the user-generated content respectively in the second plurality of positions.

10. A method according to claim 8, characterized in that generating a set of candidate composite images based on the set of model elements and user-generated content comprises: determining a plurality of predetermined rules for positioning the set of model elements and user-generated content; and generating the set of candidate composite images based on the plurality of predetermined rules.

11. Method, according to claim 8, characterized in that selecting the target composite image from the set of candidate composite images comprises: determining a set of scores for the set of candidate composite images; and selecting the target composite image from the set of candidate composite images based on the set of scores, wherein a Petition 870250090513, dated 10 / 03 / 2025, page 82 / 84 4 / 5 score of the target composite image exceeds a threshold score.

12. Method, according to claim 1, characterized in that generating a descriptive text of the content for the user-generated content comprises: obtaining a descriptive text of the content for the user-generated content by applying the user-generated content to a machine learning model.

13. A method according to claim 11, characterized in that the user-generated content is an image or a video, and the machine learning model is a visual model.

14. Device for generating an image, characterized in that it comprises: a content description text generation module configured to generate, based on user-generated content, a content description text for the user-generated content; a model element set generation module configured to generate, based on the content description text, a model element set for the user-generated content; and a target composite image generation module configured to generate, based on the model element set and user-generated content, a target composite image.

15. Electronic device, characterized in that it comprises: at least one processor; and a memory for storing instructions which, when executed by at least one processor, cause the at least one processor to perform the method defined in any one of claims 1 to 13.

16. Computer-readable storage medium, characterized by the fact that it has instructions stored on it which, when executed by a processor, cause the processor to perform the method defined in any of claims 1 to 13.

17. Product, process, system, kit, means or use, characterized in that it comprises one or more elements described in the descriptive report, claims, drawings, sequence listing, or summary of this application, when applicable. Petition 870250090513, dated 03 / 10 / 2025, pp. 84 / 84

Images