Advertisement generation method and system, medium and device

Abstract

Provided are an advertisement generation method and system, a medium and a device. The method comprises: acquiring standard-product information and environment information of at least one standard product (S201); generating a standard-product image corresponding to the standard-product information on the basis of artificial intelligence generated content (AIGC), measuring each dimensionality of the standard-product image to obtain label information of each dimensionality of the standard-product image, and performing semantic combination on the label information of each dimensionality and the environment information to generate an advertisement copywriting (S202); and matching an advertisement template from an advertisement template library according to the label information of each dimensionality, and combining the standard-product image and the advertisement copywriting into the advertisement template to generate a standard-product advertisement (S203). The solution has the technical advantages of simplicity and high efficiency.

Description

Advertisement generation method, system, medium and device Technical Field

[0001] The present application relates to the field of Internet technology, and in particular to an advertisement generation method, system, medium, and device. Background Art

[0002] Advertising has always been an important method for product promotion. With the rapid development of the internet, it's becoming increasingly necessary to advertise online products on other platforms to attract users. Therefore, how to simply and efficiently generate advertisements is a technical challenge facing those skilled in the art.

[0003] Summary of the Invention

[0004] In view of this, the present application provides an advertisement generation method, system, medium and device, the main purpose of which is to provide a simple and efficient advertisement generation solution.

[0005] According to one aspect of the present application, a method for generating an advertisement is provided, comprising:

[0006] Obtaining product information and scene information of at least one standard product;

[0007] Based on artificial intelligence content generation AIGC, a standard product image corresponding to the standard product information is generated, each dimension of the standard product image is measured to obtain label information of each dimension of the standard product image, and the label information of each dimension is semantically combined with the scene information to generate advertising copy;

[0008] According to the label information of each dimension, an advertisement template is matched from an advertisement template library, and the standard product image and the advertisement copy are synthesized into the advertisement template to generate a standard product advertisement.

[0009] In one implementation,

[0010] The standard product information includes standard product text description information or standard product category information;

[0011] The artificial intelligence-based content production AIGC generates a standard product image corresponding to the standard product information, including: inputting the standard product text description information or standard product category information into the AIGC model, and outputting the standard product image corresponding to the standard product text description information according to the interactive prompt image generation algorithm or text image generation algorithm of the AIGC model.

[0012] In one implementation,

[0013] The standard product information includes standard product text description information or standard product category information and standard product sub-image, or the standard product information includes standard product text description information or standard product category information and standard product outline image;

[0014] The artificial intelligence-based content production AIGC generates a standard product image corresponding to the standard product information, including: inputting the standard product text description information or the standard product category information and the standard product sub-image, or the standard product text description information or the standard product category information and the standard product outline image, into the AIGC model; and outputting the standard product image corresponding to the standard product text description information or the standard product category information and the standard product sub-image, or outputting the standard product image corresponding to the standard product text description information or the standard product category information and the standard product outline image according to the image-to-image algorithm of the AIGC model.

[0015] In one implementation, the method further includes: training to obtain an AIGC model, further including:

[0016] On the basis of the open source stable diffusion model, the historical platform standard product data of the platform to which the standard product belongs is used for secondary training to obtain a base model composed of the local cultural graph model and the local map graph model;

[0017] Perform low-level adaptation LORA of the large language model for the base model, and generate a corresponding LORA model for each category of target items;

[0018] The base model and multiple LORA models constitute the AIGC model.

[0019] In one implementation, measuring each label dimension of the standard product image to obtain label information of each dimension of the standard product image includes:

[0020] Converting the visual information of the standard product image into semantic information based on the image single classification algorithm and / or the image multi-label classification algorithm of the image label model;

[0021] Based on the semantic information, label information of each dimension is measured and obtained, and the label information includes any one or more of image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information.

[0022] In one implementation, matching an advertisement template from an advertisement template library according to the tag information of each dimension includes:

[0023] According to any one or more of the image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information, the various templates in the advertising template library are matched, and an advertising template is matched according to the matching effect.

[0024] In one implementation, the semantically combining the label information of each dimension and the scene information to generate an advertisement copy includes:

[0025] Inputting the label information of each dimension and the scene information into a large language model, obtaining the syntax and semantics corresponding to the information, and obtaining label semantic information and scene semantic information;

[0026] The tag semantic information and the scene semantic information are combined to obtain the advertising copy.

[0027] In one implementation, synthesizing the standard product image and the advertising copy into the advertising template includes:

[0028] The standard product image is placed in the main body position of the advertising template, and the advertising copy is placed in the non-main body position of the advertising template. Alternatively, the standard product image is placed in the main body position of the advertising template, and short numbers or text are extracted based on the advertising copy and highlighted in the advertising template.

[0029] In one implementation, after generating the standard product image corresponding to the standard product information, the method further includes:

[0030] Performing quality inspection on the standard product image in at least one dimension to determine quality inspection results for each dimension; the at least one dimension includes image color dimension, picture naturalness dimension, image appearance dimension, image optical dimension, image clarity dimension, product body integrity dimension, font dimension, special character dimension, and watermark dimension;

[0031] During the advertisement synthesis process, the quality of each dimension is compared according to the detection threshold of each dimension to determine whether the image quality detection result passes. Advertisement synthesis is only performed on images that pass the quality detection.

[0032] In one implementation, the method further includes:

[0033] For standard product images that fail the image quality test, image correction is performed to obtain standard product images that meet the image quality test, or the standard product images are regenerated based on the AIGC model.

[0034] According to one aspect of the present application, there is provided an advertisement generation system, comprising:

[0035] An AIGC image generation engine for artificial intelligence content generation is used to obtain product information and scene information of at least one standard product and generate a standard product image corresponding to the standard product information based on the AIGC model;

[0036] An image labeling system engine, configured to measure each dimension of the standard product image based on an image labeling model to obtain label information of each dimension of the standard product image;

[0037] An intelligent copy generation engine, configured to semantically combine the label information of each dimension and the scenario information based on a large language model to generate advertising copy;

[0038] The material intelligent synthesis engine is used to match the advertising template from the advertising template library according to the label information of each dimension, and synthesize the standard product image and the advertising copy into the advertising template to generate a standard product advertisement.

[0039] In one implementation, the standard product information includes standard product text description information or standard product category information;

[0040] The AIGC image generation engine is specifically used to: input the standard product text description information or standard product category information into the AIGC model, and output the standard product image corresponding to the standard product text description information according to the interactive prompt image generation algorithm or text image generation algorithm of the AIGC model.

[0041] In one implementation,

[0042] The standard product information includes standard product text description information or standard product category information and standard product sub-image, or the standard product information includes standard product text description information or standard product category information and standard product outline image;

[0043] The AIGC image generation engine is specifically used to: input the standard product text description information or the standard product category information and the standard product sub-image, or the standard product text description information or the standard product category information and the standard product outline image, into the AIGC model, and output the standard product image corresponding to the standard product text description information or the standard product category information and the standard product sub-image, or output the standard product image corresponding to the standard product text description information or the standard product category information and the standard product outline image according to the image generation algorithm of the AIGC model.

[0044] In one implementation, the AIGC image generation engine is also used to: perform secondary training on the basis of an open source stable diffusion model using the historical platform standard product data of the platform to which the standard product belongs, and obtain a base model composed of a local cultural image model and a local map image model; perform low-order adaptation LORA of the large language model on the base model to generate a corresponding LORA model for each category of target products; and use the base model and multiple LORA models to form the AIGC model.

[0045] In one implementation, the image labeling system engine is specifically used to: convert the visual information of the standard product image into semantic information based on the image single classification algorithm and / or image multi-label classification algorithm of the image labeling model; based on the semantic information, measure and obtain label information of each dimension, and the label information includes any one or more of image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information.

[0046] In one implementation, the material intelligent synthesis engine is specifically used to: match the various templates in the advertising template library according to any one or more of the image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information, and match the advertising template according to the matching effect.

[0047] In one implementation, the intelligent copy generation engine is specifically used to input the label information of each dimension and the scene information into the large language model, obtain the syntax and semantics corresponding to the information, and obtain label semantic information and scene semantic information; combine the label semantic information and the scene semantic information to obtain the advertising copy.

[0048] In one implementation, the material intelligent synthesis engine is specifically used to: place the standard product image in the main position of the advertising template, place the advertising copy in the non-main position of the advertising template, or place the standard product image in the main position of the advertising template, extract short numbers or text based on the advertising copy and highlight them in the advertising template.

[0049] In one implementation, the method further includes:

[0050] An image quality audit engine is configured to perform quality inspection on the standard product image in at least one dimension and determine the quality inspection results of each dimension; the at least one dimension includes image color dimension, picture naturalness dimension, image appearance dimension, image optical dimension, image clarity dimension, product body integrity dimension, font dimension, special character dimension, and watermark dimension;

[0051] The material intelligent synthesis engine is also used to: during the advertisement synthesis process, compare the quality of each dimension according to the detection threshold of each dimension, determine whether the image quality detection result passes, and only synthesize advertisements for images that pass the quality detection.

[0052] In one implementation,

[0053] The material intelligent synthesis engine is also used to: perform image correction on standard product images that fail the image quality detection results to obtain standard product images that meet the image quality detection results, or regenerate standard product images based on the AIGC model.

[0054] According to one aspect of the present application, a storage medium is provided, in which a computer program is stored. The computer program is configured to execute the above-mentioned advertisement generation method when running.

[0055] According to one aspect of the present application, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the above-mentioned advertisement generation method.

[0056] By means of the above technical solution, the present application provides an advertisement generation method, system, medium and device, which obtains a standard product image through AIGC text-generated images or image-generated images, establishes multi-dimensional labels for the standard product images, and generates advertising copy based on label information and scene information, and finally synthesizes the standard product image and advertising copy into the advertising template selected according to the multi-dimensional label to obtain the final personalized standard product advertisement. The entire advertisement generation scheme only needs to provide information related to the standard product, and there is no need to manually select pictures or templates, etc., which is simple and efficient. In addition, since the AIGC model can provide a large number of high-quality images, it can enrich the content and form of advertisements. In addition, through image label analysis, the advertisement identity information can be determined from multiple angles, thereby matching the most suitable advertisement template, and by adding scene information, the LLM model can be used to generate personalized advertising copy that meets the standard product application scenario, which helps to improve the conversion rate of advertisements. In addition, in one method, in order to control the image generated by the AIGC model, the quality of the standard product image can be reviewed, and whether to perform image synthesis can be determined based on the quality review results. Therefore, by adding an image quality detection operation, the image quality and the advertisement quality can be improved.

[0057] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0059] FIG1 is a schematic diagram showing an implementation scenario of an advertisement generation method provided in an embodiment of the present application;

[0060] FIG2 shows a flow chart of an advertisement generation method provided by an embodiment of the present application;

[0061] FIG3 shows a schematic diagram of the structure of an advertisement generation system provided in an embodiment of the present application;

[0062] FIG4 is a schematic diagram showing the working principle of the AIGC image generation engine provided in an embodiment of the present application;

[0063] FIG5 shows a schematic diagram of an example of an AIGC image generation engine provided in an embodiment of the present application;

[0064] FIG6 shows a logic diagram of a base model in an AIGC model provided in an embodiment of the present application;

[0065] FIG7 shows a schematic diagram of the working principle of the image quality review engine provided in an embodiment of the present application;

[0066] FIG8 is a schematic diagram showing the working principle of the image tagging system engine provided in an embodiment of the present application;

[0067] FIG9 is a schematic diagram showing the working principle of the intelligent copywriting generation engine provided in an embodiment of the present application;

[0068] FIG10 shows a schematic diagram of an example of an intelligent copywriting generation engine provided in an embodiment of the present application;

[0069] FIG11 is a schematic diagram showing the working principle of the intelligent material synthesis engine provided in an embodiment of the present application;

[0070] FIG12 shows a schematic diagram of an example of a material intelligent synthesis engine provided in an embodiment of the present application. DETAILED DESCRIPTION

[0071] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only embodiments of a part of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of this application. It should be noted that, in the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0072] In Internet scenarios (such as food delivery apps), placing advertising materials on other external Internet channels is one of the important means of attracting new customers, and plays an important role in user growth and enhancing product influence. Image materials are one of the main types of external materials, and their original images mainly come from the local image library of the APP. These materials have problems such as poor image quality and many infringing elements such as merchant logos or text. In addition, it is currently impossible to automatically generate fixed categories of external materials according to scene requirements, which will result in the inability to timely guarantee advertising materials for certain short-term popular products when facing special seasons, events, and scenes. In addition, in the process of synthesizing advertising materials, only a few factors such as image color, subject position, and main color are relied upon to match the inherent template, and the synthesized materials lack concise and effective slogans, resulting in a large amount of waste in the delivery cost.

[0073] To this end, an embodiment of the present application provides an advertisement generation method. Referring to Figure 1, it is a schematic diagram of an application scenario of an advertisement generation method provided by an embodiment of the present application. In an embodiment of the present application, it is only necessary to provide the standard product information (such as standard dishes) and scene information (such as time of year, season, event or other scene) of the standard product to be generated, and then the advertisement generation system can automatically generate multiple high-definition, high-value-score, diverse and rich externally projected material results, saving the time cost and labor cost of selecting gallery images. In addition, this solution will also intelligently generate advertising slogans based on the time of year, season, event or other scenes and the dishes themselves, and match them with a variety of templates to improve the quality of generated materials. Finally, the generated advertisements are externally projected through the advertisement page.

[0074] 2 , which shows a flow chart of an advertisement generation method provided by an embodiment of the present application. The advertisement generation method includes the following steps S201 - S203 .

[0075] S201: Obtain standard product information and scene information of at least one standard product.

[0076] Standard products are products displayed or sold on websites, apps, or mini-programs, such as the individual dishes displayed on food delivery apps. Each standard product corresponds to a product category. For example, Coke belongs to the beverage category, while Kung Pao Chicken belongs to the dish category. Understandably, these broad categories can be further broken down into multiple subcategories.

[0077] In an embodiment of the present application, a product list including multiple products can be submitted to the advertisement generation system at one time. In the list, each product is distinguished by product information such as category number or product name. At the same time, in order to add a meaning that meets the scene requirements in the generated advertisement later, the product list can also be accompanied by scene information for each product. The scene information includes but is not limited to information about the time of year, season, event, or other scene. It is understood that multiple products belonging to the same category can be set in one product list. Since products belonging to the same category are generally suitable for the same or similar scenes, the scene information can be marked only once in the list.

[0078] In an embodiment of the present application, the standard product information can have multiple forms. One form is that the standard product information only includes standard product text description information or standard product category information. For example, the standard product text description information can be the standard product name or standard product descriptive text, and the standard product category information refers to the category identifier or category name to which the standard product belongs; another form is that in addition to the text description information or standard product category information, the standard product information also includes a standard product sub-image or a standard product outline image, wherein the standard product sub-image refers to an image corresponding to the standard product. For example, assuming that the standard product is "Kung Pao Chicken", the standard product sub-image is the dish image of "Kung Pao Chicken", and the standard product outline image refers to an image that only determines the image outline elements (image size and shape, etc.).

[0079] S202: Based on artificial intelligence content generation (AIGC), a standard product image corresponding to the standard product information is generated, each dimension of the standard product image is measured to obtain label information of each dimension of the standard product image, and the label information and scene information of each dimension are semantically combined to generate advertising copy.

[0080] In step S202 , a standard product image may be first generated based on the AIGC model, and then a multi-dimensional label may be established for the standard product image. Finally, an advertising copy may be generated based on the label information and the scene information.

[0081] In the implementation of generating standard product images based on the AIGC model, it can be based on text-generated images or on a "text + image" combination. As previously mentioned, in step S201, the input standard product information can include a single text form of text / category or a "text + image" form of "text / category + image". Therefore, for different forms of standard product information, standard product images are generated based on different algorithms of the AIGC model.

[0082] AIGC can be used to create images from text. By using text descriptions, text can be converted into images and displayed. AIGC-based text-to-image can provide the following functions: (1) Text-to-image conversion: Convert the input text into an image effect to make the text more vivid; (2) Image customization: Users can choose different colors, fonts, backgrounds, painting styles, etc. to customize their favorite images.

[0083] Therefore, when the standard product information includes standard product text description information or standard product category information, the process of generating a standard product image corresponding to the standard product information based on the AIGC model may include: inputting the standard product text description information or standard product category information into the AIGC model, and outputting the standard product image corresponding to the standard product text description information according to the interactive prompt image generation algorithm or text image generation algorithm of the AIGC model. The interactive prompt image generation algorithm may be, for example, a prompt image generation algorithm, and the text image generation algorithm may be, for example, a text image generation algorithm of Latent Diffusion Models (LDMs).

[0084] AIGC can also generate images from images, generating new images by modifying the parameters of a given image. AIGC-based image generation can also provide image customization: users can customize their favorite images based on the input image, with different colors, fonts, backgrounds, painting styles, and more.

[0085] Therefore, in the case where the standard product information includes standard product text description information or standard product category information and a standard product sub-image, or includes standard product text description information or standard product category information and a standard product outline image, the process of generating a standard product image corresponding to the standard product information based on the AIGC model may include: inputting the standard product text description information or standard product category information and the standard product sub-image, or the standard product text description information or standard product category information and the standard product outline image, into the AIGC model, and outputting the standard product image corresponding to the standard product text description information or standard product category information and the standard product sub-image, or outputting the standard product image corresponding to the standard product text description information or standard product category information and the standard product outline image according to the image generation algorithm of the AIGC model. The image generation algorithm may be, for example, an image generation algorithm based on a stable diffusion model, which supports generating new images by modifying parameters.

[0086] It is understood that before generating standard product images through the AIGC model, a training step can also be included to obtain the AIGC model. In one implementation, based on the open source Stable diffusion model, secondary training can be performed using historical platform standard product data from the platform to which the standard product belongs, to obtain a base model composed of a local cultural image model and a local map image model; a low-order adaptation (LORA) of the large language model is performed on the base model to generate a corresponding LORA model for each category of target products; and the base model and multiple LORA models are used to form the AIGC model.

[0087] In the process of generating label information for each dimension of an image based on an image labeling model, by assigning correct and appropriate labels to the image, the image's visual information is converted into semantic information, which facilitates better understanding and analysis of the image. Image labeling includes single image classification and multi-label image classification. Single image classification involves finding a classification label from a fixed set of classification labels and assigning it to the input image. In the real world, an image often contains rich semantic information, such as multiple objects, scenes, and behaviors. Multi-label image classification aims to assign multiple labels to an image to fully express the specific content contained in the image.

[0088] Therefore, based on the image label model, measuring each dimension of the standard product image to obtain label information for each dimension of the standard product image can include: using an image single classification algorithm and / or an image multi-label classification algorithm based on the image label model to convert the visual information of the standard product image into semantic information; based on the semantic information, measuring and obtaining label information for each dimension, wherein the label information includes any one or more of image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information. The image single classification algorithm or the image multi-label classification algorithm can be implemented based on an image classification algorithm using a convolutional neural network.

[0089] In the process of generating advertising copy by combining the label information and scene information of each dimension based on the large language model, the label information and scene information of each dimension are input into the large language model, the syntax and semantics corresponding to the information are obtained, and the label semantic information and scene semantic information are obtained; the label semantic information and scene semantic information are combined to obtain the advertising copy. Among them, the large language model (LLM) refers to a deep learning model trained with a large amount of text data, which can generate natural language text or understand the meaning of language text. In an embodiment of the present application, the "identity" information of each dimension of the advertisement can be fully analyzed through multi-dimensional labels, and combined with the scene information, the scene corresponding to the standard product can be further limited. Therefore, based on the advertisement identity information and scene information input into the LLM model, a highly summarized advertising copy can be output, such as a one-sentence slogan.

[0090] S203: Match an advertisement template from an advertisement template library according to the tag information of each dimension, and synthesize the standard product image and the advertisement copy into the advertisement template to generate a standard product advertisement.

[0091] After generating the standard product image and advertising copy, you can understand the label information of each dimension, match the most suitable advertising template in the advertising template library, and then embed the standard product image and advertising copy into the advertising template to generate the final advertisement.

[0092] As previously described, the label information of each dimension can be understood as the identity information of each dimension of the advertisement, including but not limited to image color distribution information, image subject information, image category information, image name information, standard product name information, standard product associated scene information, etc. Therefore, according to any one or more of the image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information, the templates in the advertisement template library can be matched to the advertisement template based on the matching effect. The matching effect can be determined by automatic machine scoring or user selection.

[0093] After selecting an advertising template, the standard product image and the advertising copy are synthesized into the advertising template. Specifically, the standard product image can be placed in the main position of the advertising template, and the advertising copy can be placed in the non-main position of the advertising template. Alternatively, the standard product image can be placed in the main position of the advertising template, and short numbers or text can be extracted based on the advertising copy and highlighted in the advertising template.

[0094] Thus, the standard product image is obtained through AIGC text-generated images or image-generated images, multi-dimensional labels are established for the standard product image, and advertising copy is generated based on the label information and scene information. Finally, the standard product image and advertising copy are synthesized into the advertising template selected according to the multi-dimensional label to obtain the final personalized standard product advertisement. The entire advertising generation solution only requires the provision of relevant information about the standard product, without the need for manual operations such as selecting pictures or templates, making it simple and efficient. In addition, since the AIGC model can provide a large number of high-quality images, it can enrich the content and form of advertisements. In addition, through image label analysis, the identity information of the advertisement can be determined from multiple angles, thereby matching the most suitable advertising template. Moreover, by adding scene information, the LLM model can be used to generate personalized advertising copy that conforms to the standard product application scenario, which helps to improve the conversion rate of advertisements.

[0095] In one implementation, to control the images generated by the AIGC model, a quality review of the standard product images can be performed, and the decision to proceed with image synthesis can be made based on the review results. By adding image quality testing, image quality can be improved, thereby enhancing the quality of advertisements.

[0096] Therefore, after generating the standard product image corresponding to the standard product information, the following steps may also be included: performing quality inspection on the standard product image in at least one dimension to determine the quality inspection results of each dimension; at least one dimension includes image color dimension, picture natural dimension, image appearance dimension, image optical dimension, image clarity dimension, product body integrity dimension, font dimension, special character dimension, and watermark dimension; during the advertising synthesis process, the quality of each dimension is compared according to the detection threshold of each dimension to determine whether the image quality inspection result passes, and only advertisements are synthesized for images that pass the quality inspection.

[0097] In addition, for standard product images that fail the image quality test, image correction can be performed until a standard product image that meets the image quality test is obtained, or a new standard product image can be generated based on the AIGC model.

[0098] The following uses dishes as an example of standard products to illustrate the embodiments of the present application from the perspective of the advertisement generation system.

[0099] Referring to FIG3 , there is shown a schematic diagram of the structure of an advertisement generation system provided in an embodiment of the present application.

[0100] In the advertisement generation system provided in the embodiment of the present application, first, the user (advertisement generator) publishes a list of standard dishes for key categories to be generated based on scene information such as seasons and festivals, and then enters the automated material generation process.

[0101] The process of generating automated external content based on AIGC in the scenario shown in Figure 3 includes the following parts:

[0102] 1) AIGC image generation engine: Generates dish images for each standard dish;

[0103] 2) Image quality review engine: This engine evaluates the quality of each dimension of the AIGC-generated results.

[0104] 3) Image labeling system engine: Calculates the attribute labels of each image based on the AIGC generated results;

[0105] 4) Intelligent copywriting generation engine: Generates advertising slogans using scene information, standard dish information, and related image information;

[0106] 5) Intelligent material synthesis engine: Utilize the output results of steps 1) to 4) above, combined with local external advertising material templates, to output external advertising materials.

[0107] The following is an illustrative introduction to the principles and examples of the AIGC image generation engine, image quality review engine, image labeling system engine, intelligent copy generation engine, and intelligent material synthesis engine.

[0108] The AIGC image generation engine uses intelligent prompt generation, AIGC text-to-image, and image-to-image algorithms to generate a greater quantity and variety of original external materials based on popular standard dishes and their accompanying images (if available). Compared to the current method of selecting from image libraries, this method offers greater timeliness and diversity in style.

[0109] Referring to Figure 4, a schematic diagram of the working principle of the AIGC image generation engine provided by an embodiment of the present application is shown. It can support various input forms: plain text input, text + original seed image, and text + outline image. In addition, input conditions can be added (such as customized generation ratio, customized background, customized perspective, customized style, etc.). In the AIGC image generation engine, the AIGC generation result of the original image is obtained through functions such as shape control, category fine-tuning and improvement (LORA), perspective control, style control, background replacement, package circle generation, and multi-size generation (smart cropping).

[0110] See Figure 5, which shows a schematic diagram of an example AIGC image generation engine provided by an embodiment of the present application. In this example, the AIGC image generation engine generates images of standard dishes. This example shows three input forms (plain text input, text + original seed image, and text + outline image), and the corresponding AIGC generation results.

[0111] The AIGC image generation engine consists of a local version of the Stable Diffusion base model + a variety of auxiliary generation capabilities. The local version of the Stable Diffusion base model is obtained by secondary training using local platform data based on the open source Stable Diffusion, which can greatly improve the localization effect.

[0112] See Figure 6, which shows a logical diagram of the base model in the AIGC model provided by an embodiment of this application. According to Figure 6, the stable diffusion model, when generating an image, primarily includes two inputs: a prompt and a seed (which generates a noise map). A fixed seed and a fixed resolution generate a fixed noise map, and the model uses this as the basis for image generation. The noise map is not a single image, but rather a representation in the latent space.

[0113] Various auxiliary capabilities enable further post-processing and tuning of large models to suit various application scenarios. Image LoRa tuning utilizes a minimally parameterized model structure to store fine-tuned parameter changes for a specific style or scene. Adding LoRa improves performance in certain scenarios. A LoRa model is trained for each category, achieving a structure consisting of one Stable Diffusion base and N LoRas.

[0114] ControlNet can be used to control the shape of the generated image. It supports various shape information inputs, such as Canny edge detection, Hed contour detection, and Depth detection, and applies the results to the generated image. Furthermore, it supports image background replacement and editing. Using inpainting, the background outside the subject is masked out, and a new background image is generated using diffusion. Existing background information can also be spliced ​​together to achieve a unified style. Furthermore, subject positioning can be used to determine the subject's relative position within the original image, allowing for intelligent cropping to produce generated images at various scales.

[0115] Because AIGC image generation technology belongs to a generative model, the results generated by such models are often divergent. To prevent illegal, infringing, and generated images from being released, a comprehensive image quality review engine is required. In an embodiment of the present application, the image quality review engine includes quality reviews across multiple dimensions, such as image color, naturalness, appearance, optics, clarity, product integrity, font detection, pseudo-logo recognition, special character pattern detection, raw food detection, ugly image detection, watermark detection, and image psoriasis detection. See Figure 7 for a schematic diagram of the operating principle of the image quality review engine provided in an embodiment of the present application. Because the results of generative models often have uncertainty, it is necessary to monitor the generated results from multiple dimensions to prevent the release of ugly, infringing, or value-incompatible images. The image quality review engine performs quality assessments on all dimensions of the image, typically using N discriminant models to obtain corresponding results. Quality assessments for each dimension can be controlled by setting different thresholds during the image synthesis process to control the generated results.

[0116] The image labeling system engine refers to the process of labeling various dimensions of an image through various image algorithms, which is used for subsequent intelligent matching of advertising templates, intelligent generation of advertising copy, generation of meal plans, etc., and can also support the systematic construction of local gallery labels so that each picture has clear information. The labeling system that can be constructed by the image labeling system engine mainly includes: image color distribution, image subject ratio, subject position, subject perspective, subject integrity, subject number, picture category, picture name, picture standard dish, dish common name, dish associated seasonal scene and other information. See Figure 8, which shows a schematic diagram of the working principle of the image labeling system engine provided by an embodiment of the present application. The role of the image quality labeling system engine is to measure the various dimensions of the generated result image and obtain the "identity" information of all dimensions of the generated result. This information will provide information input in multiple dimensions such as subsequent template synthesis, template color selection, cropping position, cropping size, text position, etc.

[0117] Normally, the external image advertising material includes not only the image itself, but also interest points such as red envelopes, advertising slogans or recommendations, etc. In the embodiment of the present application, the copywriting intelligent generation engine is used to generate recommendations or advertising slogans, which can refer to scene information such as time of year and season; food information such as dishes, categories, and flavors; visual information such as image appearance and number of subjects, and combine relevant information through technologies such as the LLM large language model to give advertising copy. See Figure 9, which shows a schematic diagram of the working principle of the copywriting intelligent generation engine provided in the embodiment of the present application. The copywriting intelligent generation engine mainly generates recommended advertising slogans related to the delivery scene and delivery materials. Based on the label information of each dimension of the image, the standard dish name information, and the delivery seasonal scene information, a one-sentence advertising recommendation result is generated based on the large language model. See Figure 10, which shows a schematic diagram of an example of the copywriting intelligent generation engine provided in the embodiment of the present application.

[0118] The material intelligent generation engine analyzes the label system of the images that have passed the quality review, and performs color matching, aesthetic matching, intelligent cropping, edge extension and other technologies on the original materials that have passed the quality review and the advertising template, and adds the advertising promotion copy of the image to synthesize the final external advertising material. Referring to Figure 11, a schematic diagram of the working principle of the material intelligent synthesis engine provided by the embodiment of the present application is shown. The material intelligent synthesis engine is a process that inputs the results of steps 1 to 4 above (original image, image quality results, image label analysis results, advertising copy) and matches the template from the template library. It mainly includes different analysis angles such as subject position analysis, subject integrity analysis, subject perspective analysis, extension direction analysis, subject color analysis, and generates the final result according to the analysis results and template matching. In addition, intelligent recommendations can also be obtained. As shown in Figure 12, an example of "beef pizza with hot peppers" is shown, which is a schematic diagram of material intelligent synthesis.

[0119] So far, the advertisement generation scheme provided by the embodiment of the present application has technical advantages over the existing scheme. The current synthesis of external advertising materials is mainly "material-driven", that is, the generation range of materials mainly refers to historical indicators such as the click-through rate and conversion rate of past materials. When faced with scenes such as festivals (Dragon Boat Festival), seasons (Winter Solstice), events (football matches), etc., it is difficult to synthesize the advertising materials that users are most interested in in a timely and effective manner; in addition, the original images of current image advertising materials come from local image libraries, and most of the images have low clarity, poor lighting, mismatched images and texts, the presence of merchant logos, excessive text, watermarks and other problems, which will greatly reduce the aesthetics of the synthesized materials; and the current materials are only simple combinations of original images + templates + interest points, lacking attractive introduction slogans and guide words, resulting in a waste of delivery costs. To this end, the advertising generation system provided by the embodiment of the present application can effectively solve the defects of the above-mentioned existing solutions. Among them, it proposes to generate external image advertising materials based on the AIGC model, which improves efficiency and saves delivery costs compared with the existing process; the AIGC technology is used to generate new image materials, which solves problems such as low clarity, ugly pictures, and infringing text and logo elements compared to screening images from the gallery, and increases the diversity and operability of the generation process; multiple algorithms are used to review image quality factors, and an image labeling system is established to perform quality analysis and labeling system construction on the generated image results, avoiding excessive manual review costs and improving efficiency; the LLM large language model is used, combined with multiple dimensions such as external projection scenarios, standard dish names, image labels, and benefit prices, to generate recommended advertising slogans of different lengths, which are more in line with scenario requirements and more targeted.

[0120] An embodiment of the present application further provides a storage medium storing a computer program, wherein the computer program is configured to execute the steps of any of the above method embodiments when run.

[0121] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:

[0122] Obtaining product information and scene information of at least one standard product;

[0123] Based on AIGC, a standard product image corresponding to the standard product information is generated, each dimension of the standard product image is measured to obtain label information of each dimension of the standard product image, and the label information of each dimension is semantically combined with the scene information to generate advertising copy;

[0124] According to the label information of each dimension, an advertisement template is matched from an advertisement template library, and the standard product image and the advertisement copy are synthesized into the advertisement template to generate a standard product advertisement.

[0125] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.

[0126] An embodiment of the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0127] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0128] Optionally, in this embodiment, the processor may be configured to execute the following steps through a computer program:

[0129] Obtaining product information and scene information of at least one standard product;

[0130] Based on AIGC, a standard product image corresponding to the standard product information is generated, each dimension of the standard product image is measured to obtain label information of each dimension of the standard product image, and the label information of each dimension is semantically combined with the scene information to generate advertising copy;

[0131] According to the label information of each dimension, an advertisement template is matched from an advertisement template library, and the standard product image and the advertisement copy are synthesized into the advertisement template to generate a standard product advertisement.

[0132] Optionally, specific examples in this embodiment may refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be described in detail here.

[0133] The serial numbers of the above embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0134] In the above embodiments of the present application, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0135] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0136] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0137] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0138] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0139] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. An advertisement generating method, characterized in that: include: Obtaining product information and scene information of at least one product; Based on artificial intelligence content production AIGC, a standard product image corresponding to the standard product information is generated, each dimension of the standard product image is measured to obtain label information of each dimension of the standard product image, and the label information of each dimension and the scene information are semantically combined to generate advertising copy; According to the label information of each dimension, an advertisement template is matched from an advertisement template library, and the standard product image and the advertisement copy are synthesized into the advertisement template to generate a standard product advertisement.

2. The method according to claim 1, characterized in that The standard product information includes standard product text description information or standard product category information; The method of generating a standard product image corresponding to the standard product information based on AIGC includes: inputting the standard product text description information or the standard product category information into the AIGC model, and outputting the standard product image corresponding to the standard product text description information according to the interactive prompt image generation algorithm or the text image generation algorithm of the AIGC model.

3. The method according to claim 1, characterized in that The standard product information includes standard product text description information or standard product category information and standard product sub-image, or the standard product information includes standard product text description information or standard product category information and standard product outline image; The method of generating a standard product image corresponding to the standard product information based on AIGC includes: inputting the standard product text description information or the standard product category information and the standard product sub-image, or the standard product text description information or the standard product category information and the standard product outline image, into the AIGC model, and outputting the standard product image corresponding to the standard product text description information or the standard product category information and the standard product sub-image, or outputting the standard product image corresponding to the standard product text description information or the standard product category information and the standard product outline image according to the image-to-image algorithm of the AIGC model.

4. The method according to claim 1, characterized in that: Also includes: The AIGC model is trained, further including: On the basis of the open source stable diffusion model, the historical platform standard product data of the platform to which the standard product belongs is used for secondary training to obtain a base model composed of a local cultural graph model and a local map graph model; A low-order adaptation LORA of the large language model is performed on the base model to generate a corresponding LORA model for each category of target products; The AIGC model is constructed by the base model and multiple LORA models.

5. The method according to claim 1, characterized in that The measuring each label dimension of the standard product image to obtain label information of each dimension of the standard product image includes: Based on the image single classification algorithm and / or the image multi-label classification algorithm of the image label model, converting the visual information of the standard product image into semantic information; Based on the semantic information, label information of each dimension is measured and obtained, and the label information includes any one or more of image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information.

6. The method according to claim 5, characterized in that The step of matching an advertisement template from an advertisement template library according to the tag information of each dimension includes: According to any one or more of the image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information, the various templates in the advertising template library are matched, and the advertising template is matched according to the matching effect.

7. The method according to claim 1, characterized in that The semantically combining the label information of each dimension and the scene information to generate an advertisement copy includes: Inputting the label information of each dimension and the scene information into a large language model, acquiring the syntax and semantics corresponding to the information, and obtaining label semantic information and scene semantic information; The tag semantic information and the scene semantic information are combined to obtain the advertisement copy.

8. The method according to claim 1, characterized in that: The step of synthesizing the standard product image and the advertisement text into the advertisement template includes: The standard product image is placed in the main body position of the advertising template, and the advertising copy is placed in the non-main body position of the advertising template. Alternatively, the standard product image is placed in the main body position of the advertising template, and short numbers or words are extracted based on the advertising copy and highlighted in the advertising template.

9. The method according to any one of claims 1 to 8, characterized in that: After generating the standard product image corresponding to the standard product information, the method further includes: Performing quality inspection on the standard product image in at least one dimension to determine the quality inspection results of each dimension; the at least one dimension includes image color dimension, picture natural dimension, image appearance dimension, image optical dimension, image clarity dimension, commodity body integrity dimension, font dimension, special character dimension, and watermark dimension; During the advertisement synthesis process, the quality of each dimension is compared according to the detection threshold of each dimension to determine whether the image quality detection result passes, and advertisement synthesis is performed only for images that pass the quality detection.

10. The method according to claim 9, characterized in that Also includes: For standard product images that fail the image quality test, image correction is performed to obtain standard product images that meet the image quality test, or the standard product images are regenerated according to the AIGC model.

11. An advertisement generation system, characterized in that: include: Artificial intelligence production content AIGC image generation engine, used to obtain standard product information and scene information of at least one standard product, and generate a standard product image corresponding to the standard product information based on AIGC; An image labeling system engine is used to measure each dimension of the standard product image to obtain label information of each dimension of the standard product image; A copywriting intelligent generation engine, used to semantically combine the label information of each dimension and the scene information to generate advertising copywriting; The material intelligent synthesis engine is used to match the advertising template from the advertising template library according to the label information of each dimension, and synthesize the standard product image and the advertising copy into the advertising template to generate a standard product advertisement.

12. The system according to claim 11, characterized in that The standard product information includes standard product text description information or standard product category information; The AIGC image generation engine is specifically used to: input the standard product text description information or standard product category information into To the AIGC model, according to the interactive prompt image generation algorithm or text image generation algorithm of the AIGC model, output the standard product image corresponding to the standard product text description information.

13. The system according to claim 11, characterized in that The standard product information includes standard product text description information or standard product category information and standard product sub-image, or the standard product information includes standard product text description information or standard product category information and standard product outline image; The AIGC image generation engine is specifically used to: input the standard product text description information or the standard product category information and the standard product sub-image, or the standard product text description information or the standard product category information and the standard product outline image, into the AIGC model, and output the standard product image corresponding to the standard product text description information or the standard product category information and the standard product sub-image, or output the standard product image corresponding to the standard product text description information or the standard product category information and the standard product outline image according to the image generation algorithm of the AIGC model.

14. The system according to claim 11, characterized in that The AIGC image generation engine is also used to: based on the open source stable diffusion model, use the historical platform standard product data of the platform to which the standard product belongs to perform secondary training to obtain a base model composed of a local cultural image model and a local map image model; perform low-order adaptation LORA of the large language model on the base model to generate a corresponding LORA model for each category of target products; and use the base model and multiple LORA models to form the AIGC model.

15. The system according to claim 11, characterized in that The image labeling system engine is specifically used to: convert the visual information of the standard product image into semantic information based on the image single classification algorithm and / or image multi-label classification algorithm of the image labeling model; based on the semantic information, measure and obtain label information of each dimension, and the label information includes any one or more of image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information.

16. The system according to claim 15, characterized in that The material intelligent synthesis engine is specifically used to: match the various templates in the advertising template library according to any one or more of the image color distribution information, image subject information, image category information, image name information, standard product name information, and standard product associated scene information, and match the advertising template according to the matching effect.

17. The system according to claim 11, characterized in that The intelligent copy generation engine is specifically used to input the label information of each dimension and the scene information into a large language model, obtain the syntax and semantics corresponding to the information, and obtain label semantic information and scene semantic information; combine the label semantic information and the scene semantic information to obtain the advertising copy.

18. The system according to claim 11, characterized in that The material intelligent synthesis engine is specifically used to: place the standard product image in the main position of the advertising template, place the advertising copy in the non-main position of the advertising template, or place the standard product image in the main position of the advertising template, extract short numbers or text based on the advertising copy and highlight them in the advertising template.

19. The system according to any one of claims 11 to 18, characterized in that: Also includes: The image quality audit engine is used to perform quality inspection on the standard product image in at least one dimension to determine the quality inspection results of each dimension; the at least one dimension includes image color dimension, picture natural dimension, image appearance dimension, image optical dimension, image clarity dimension, product body integrity dimension, font dimension, special character dimension, Watermark dimension; The material intelligent synthesis engine is also used to: during the advertisement synthesis process, compare the quality of each dimension according to the detection threshold of each dimension, determine whether the image quality detection result passes, and only synthesize advertisements for images that pass the quality detection.

20. The system according to claim 19, characterized in that The material intelligent synthesis engine is also used to: perform image correction on standard product images that fail to pass the image quality detection results to obtain standard product images that meet the image quality detection, or regenerate standard product images based on the AIGC model.

21. A storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 10 when executed.

22. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 10.

Images