Information processing apparatus, information processing method, and program

Abstract

To improve usability for obtaining content to be arranged in a product in an information processing apparatus for generating data of the product such as a poster.SOLUTION: The first prompt determined by the determining means when the receiving means has received a first target impression is different from the second prompt determined by the determining means when the receiving means has received a second target impression that is different from the first target impression. The information processor is provided with a receiving means that receives a user's designation of a target impression, which is an impression that the product is ultimately required to hold, and a determining means that determines a prompt for the generation AI to generate content to be arranged in the product.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, an information processing method, and a program. [Background technology]

[0002] One method for creating poster design data using an information processing device such as a PC or smartphone is to use a template in which the shapes and arrangements of images, text, graphics, etc. to be placed on the poster are predetermined. Patent Document 1 also discloses an application program for automatically generating poster designs. With this application program, when a user specifies the impression of the poster they want to create (target impression) and the images and text (hereinafter, both will be collectively referred to as content), poster data with a design that matches the target impression with the content arranged is generated. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2024-004399 Summary of the Invention [Problem to be solved by the invention]

[0004] Patent Document 1 describes an example of a method for specifying an image to be used in a poster, in which the user selects a file saved in a storage device from a dialog screen. However, there are cases where the user does not have the content they want to use in the poster and it is difficult to specify an image.

[0005] The present disclosure aims to improve usability for obtaining content to be placed on a work, such as a poster, in an information processing device that generates data on the work. [Means for solving the problem]

[0006] The present disclosure provides an information processing device that generates data for a work, comprising: a reception means that receives a user's specification of a target impression, which is the impression that the work is ultimately required to retain; and a determination means that determines a prompt for a generation AI to generate content to be placed in the work, wherein a first prompt determined by the determination means when the reception means receives a first target impression is different from a second prompt determined by the determination means when the reception means receives a second target impression that is different from the first target impression. [Effects of the Invention]

[0007] According to the present disclosure, in an information processing device that generates data for a work such as a poster, usability for obtaining content to be placed on the work is improved. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 2 is a block diagram showing the hardware configuration of the poster generation device. [Figure 2] FIG. 1 is a software block diagram of a poster creation application. [Figure 3] FIG. 2 is a software block diagram of an image generating unit. [Figure 4] FIG. 1 is a diagram illustrating a skeleton. [Figure 5] FIG. 10 is a diagram illustrating a color scheme pattern. [Figure 6] 10A and 10B are diagrams showing a generation condition setting screen and a content setting screen provided by a poster creation application. [Figure 7] FIG. 10 is a diagram showing an image designation screen provided by a poster creation application. [Figure 8] FIG. 10 illustrates a prompt selection screen provided by a poster creation application. [Figure 9] FIG. 10 is a diagram showing an image selection screen provided by a poster creation application. [Figure 10] FIG. 10 is a diagram showing a preview screen provided by a poster creation application. [Figure 11] 10 is a flowchart showing a process for quantifying a poster impression. [Figure 12] FIG. 10 is a diagram illustrating a subjective evaluation of a poster. [Figure 13] 10 is a flowchart showing a content impression quantification process. [Figure 14] 10 is a flowchart showing a poster generation process. [Figure 15] 4 is a flowchart showing an image generation process in the first embodiment. [Figure 16] FIG. 10 illustrates a prompt impression table. [Figure 17] 10 is a flowchart illustrating a prompt change process. [Figure 18] FIG. 10 is a diagram illustrating a skeleton selection method. [Figure 19] 10A and 10B are diagrams illustrating a method for selecting a color scheme pattern and a font. [Figure 20] FIG. 2 is a software block diagram illustrating the layout unit in detail. [Figure 21] 10 is a flowchart showing a layout process. [Figure 22] FIG. 10 is a diagram for explaining input to a layout section. [Figure 23] FIG. 10 is a diagram illustrating the operation of a layout unit. [Figure 24] 10 is a flowchart showing a poster generation process according to a modified example of the first embodiment. [Figure 25] FIG. 10 is a software block diagram of a poster creation application according to a second embodiment. [Figure 26] FIG. 10 is a software block diagram for explaining in detail an image conversion unit of the second embodiment. [Figure 27] 10A and 10B are diagrams showing a generation condition setting screen and a content setting screen in the second embodiment. [Figure 28] FIG. 10 is a diagram showing an image designation screen in the second embodiment. [Figure 29] FIG. 10 is a diagram showing an image selection screen in the second embodiment. [Figure 30] 10 is a flowchart showing an image generation process executed in the second embodiment. [Figure 31] 10 is a flowchart showing an image conversion process executed in the second embodiment. [Figure 32] 10 shows a prompt input screen and a content setting screen in the second embodiment. [Figure 33] 10 is a flowchart of an image conversion process in a modified example of the second embodiment. [Figure 34] FIG. 10 is a software block diagram of a poster creation application according to a third embodiment. [Figure 35] 10 is a setting screen provided by a poster creation application according to a third embodiment. [Figure 36] FIG. 11 is a diagram showing an image designation screen in the third embodiment. [Figure 37] FIG. 11 is a diagram showing a poster preview screen according to the third embodiment. [Figure 38] 11 is a flowchart showing a poster generation process in the third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the claims, and not all of the combinations of features described in the embodiments are necessarily essential to the present disclosure. Note that the same components are given the same reference numerals and descriptions thereof will be omitted.

[0010] <<First Embodiment>> In the following embodiments, a method for automatically generating designed poster data by running a poster creation application (hereinafter also referred to as an "app") on an information processing device will be described as an example. The poster creation application of this embodiment acquires content to be used for the poster using a generation AI (Artificial Intelligence). For example, when using an image generated by an image generation AI, the user needs to specify a text prompt (hereinafter referred to as a prompt) to input to the image generation AI. Furthermore, the user needs to visually confirm whether the intended image has been generated and provide manual feedback as appropriate. Therefore, it is necessary to repeatedly specify a prompt and check the generated image. Therefore, the poster creation application of the first embodiment determines a prompt for the generation AI to generate content to be placed in the work based on a target impression specified by the user. This makes it easier to obtain the intended content and reduces the number of attempts to specify a prompt. In the following embodiments, a case will be described in which the content generated by the generation AI is an image. It should be noted that the content is not limited to images, and the processing described in the following embodiments may also be applied to cases in which the generation AI generates text information to be placed on a poster.

[0011] In the following description, unless otherwise specified, the term "image" includes still images and frame images extracted from videos. In the following embodiments, posters are described as examples of works, but works are not limited to posters. The term "image" can be used for any work that includes at least one of image content and text content, such as flyers, menus, banners, calendars, photo collages, certificates, business cards, shop cards, postcards, invitations, membership cards, and other works. These works can be used not only in print but also as electronic content on websites, social networking sites, virtual spaces, and the like.

[0012] <System configuration> 1 is a block diagram showing the hardware configuration of a poster generation device. The poster generation device 100 is an information processing device, such as a personal computer (hereinafter referred to as a PC) or a smartphone. In this embodiment, the poster generation device is described as a PC. The poster generation device 100 includes a CPU 101, a ROM 102, a RAM 103, a HDD 104, a display 105, a keyboard 106, a pointing device 107, a data communication unit 108, and a GPU 109.

[0013] A CPU (Central Processing Unit / Processor) 101 comprehensively controls the poster generation device 100, and realizes the operation of this embodiment by, for example, reading a program stored in a ROM 102 into a RAM 103 and executing it. Although FIG. 1 shows one CPU, the device may be configured with multiple CPUs.

[0014] The ROM 102 is a general-purpose ROM and stores, for example, a program executed by the CPU 101. The RAM 103 is a general-purpose RAM and is used, for example, as a working memory for temporarily storing various pieces of information when the CPU 101 executes a program.

[0015] The HDD (hard disk) 104 is a storage medium (storage unit) for storing image files, a database for holding processing results such as image analysis, and skeletons and the like used by the poster creation application.

[0016] The display 105 is a display unit that displays the user interface (UI) of this embodiment and an electronic poster as a layout result of image data (hereinafter also referred to as "image") to the user. The keyboard 106 and pointing device 107 accept instructions and operations from the user. The display 105 may also have a touch sensor function.

[0017] The keyboard 106 is used, for example, when the user inputs text information such as the title of the poster he or she wants to create on the UI displayed on the display 105, or a prompt to the generation AI.

[0018] The pointing device 107 is used, for example, when the user clicks a button on the UI displayed on the display 105 .

[0019] The data communication unit 108 communicates with external devices via a wired or wireless network, etc. For example, the data communication unit 108 transmits data laid out by the automatic layout function to a printer or server that can communicate with the poster generation device 100.

[0020] The GPU 109 is a processor that performs image processing in response to instructions from the CPU 101. The GPU 109 performs, for example, analysis of the image to be placed on the poster, estimation of the impression of the image or text, estimation of the impression of the poster, layout and color arrangement of the image, text, etc. for the skeleton, and generates a poster image.

[0021] The data bus 110 connects the blocks in FIG. 1 so that they can communicate with each other. Note that the configuration shown in FIG. 1 is merely an example and is not limited to this. For example, the poster generation device 100 may not have the display 105 and may display the UI on an external display.

[0022] The poster creation application in this embodiment is stored in HDD 104. The poster creation application is started by the user performing an operation such as clicking or double-clicking the application icon displayed on display 105 with pointing device 107.

[0023] <Software block diagram> 2 is an example of a software block diagram of a poster creation application. The poster creation application includes a poster creation condition specification unit 201, a text specification unit 202, an image specification unit 203, a target impression specification unit 204, a poster display unit 205, an image generation unit 220, and a poster creation unit 210. The poster creation unit 210 includes an image acquisition unit 211, an image analysis unit 212, a skeleton acquisition unit 213, a skeleton selection unit 214, a color scheme pattern selection unit 215, a font selection unit 216, a layout unit 217, a poster impression estimation unit 218, and a poster selection unit 219. FIG. 2 particularly shows a software block diagram related to the poster creation unit 210, which executes the automatic poster creation function.

[0024] When the poster creation application is installed in poster generation device 100, a startup icon is displayed on the top screen (desktop) of the OS (operating system) running on poster generation device 100. When the startup icon is operated (e.g., double-clicked) with pointing device 107, the program of the poster creation application stored in HDD 104 is loaded into RAM 103 and executed by CPU 101. This starts the poster creation application.

[0025] The poster creation application described above includes program modules corresponding to the components shown in Fig. 2. CPU 101 executes each program module, causing CPU 101 to function as each component shown in Fig. 2. Hereinafter, each component shown in Fig. 2 will be described as executing various processes.

[0026] The poster creation condition specification unit 201 specifies poster creation conditions to the poster generation unit 210 in response to UI operations using the pointing device 107. In this embodiment, the poster creation conditions include the poster size, number of posters to be created, and purpose category. The poster size may be specified as actual dimensions of width and height, or a paper size such as A1 or A2. The purpose category indicates the purpose for which the poster will be used, and examples include restaurants, school events, sales, and educational activities. The creation conditions specified by the poster creation condition specification unit 201 are input to the skeleton acquisition unit 213, skeleton selection unit 214, color scheme pattern selection unit 215, font selection unit 216, and poster selection unit 219.

[0027] The text designation unit 202 accepts designation of character information to be placed on the poster by the user through UI operation using the keyboard 106. The character information to be placed on the poster represents, for example, character strings representing a title, date and time, location, etc. The text designation unit 202 also links each piece of character information with information (tag or attribute information) indicating the type of character information, such as whether it is a title, or information indicating a date and time or location, and outputs the information to the skeleton acquisition unit 213 and the layout unit 217.

[0028] The image designation unit 203 accepts user designation of one or more image data to be arranged on a poster. The image data can be designated, for example, from image data stored in the HDD 104 based on the structure of a file system containing the image data, such as a device and a directory. The image data may also be designated based on associated information for identifying the image, such as the shooting date and time, or attribute information. The image designation unit 203 may also designate image data included in the poster creation application and provided as material (hereinafter also referred to as "application material image"). The image designation unit 203 may also designate image data held by an external image providing service linked to the poster creation application (hereinafter also referred to as "linked material image"). The image designation unit 203 may also designate image data generated by an image generation AI (hereinafter also referred to as "AI-generated image"). A generative AI is a machine learning model that generates new data based on learned data, and an image generation AI is a generative AI that generates an image. Specifically, an image generation AI is an AI that can generate an image from an image or text using a diffusion model, a GAN model, or the like. The image designation unit 203 outputs the file paths of the designated image and the generated image acquired from the image generation unit 220 to the image acquisition unit 211. The image designation unit 203 also accepts a user designation of a prompt to be input to the image generation AI, and outputs it to the image generation unit 220.

[0029] The target impression designation unit 204 accepts a user's designation of a target impression for a poster to be created. The target impression is an impression that the poster to be created is ultimately required to maintain, and is set so as to be given to people who view the created poster (product). In this embodiment, a UI operation using the pointing device 107 is used to designate the intensity of a word or combination of words that expresses an impression, indicating the degree to which that impression should be given. Information indicating the target impression designated by the target impression designation unit 204 is shared by the skeleton selection unit 214, color scheme selection unit 215, font selection unit 216, poster selection unit 219, and image generation unit 220. Details of impressions will be described later.

[0030] The image generation unit 220 acquires a prompt from the image designation unit 203 and acquires a target impression from the target impression designation unit 204. The image generation unit 220 generates an image to be used for a poster using the acquired prompt and image generation AI, and saves the generated image (hereinafter also referred to as a generated image) in the HDD 104. The image generation AI may be configured to be included in the poster creation application, or may be configured to use an external image generation AI service via the data communication unit 108 without being included in the poster creation application. When an external image generation AI service is used, the image generation unit 220 transmits the prompt acquired from the image designation unit 203 to the external image generation AI service and receives the generated image generated by the image generation AI service, thereby acquiring the image to be used for the poster. The image generation unit 220 outputs the file path of the generated image to the image designation unit 203.

[0031] The image generation unit 220 determines a prompt to be input to the image generation AI in order to have the image generation AI generate an image to be placed on a poster. Specifically, the image generation unit 220 determines the prompt according to the target impression specified by the user. That is, if the target impression is different, the determined prompt will also be different. A first prompt determined when a first target impression is received from the target impression designation unit 204 is different from a second prompt determined when a second target impression different from the first target impression is received from the target impression designation unit 204.

[0032] In the first embodiment, when a prompt specified by the user is acquired from image designation unit 203, image generation unit 220 determines the prompt to be input to the image generation AI by changing the prompt acquired from image designation unit 203. Image generation unit 220 changes the prompt acquired from image designation unit 203 so that the impression estimated from the changed prompt is closer to the target impression than before the change. When a first target impression is acquired, image generation unit 220 changes the prompt acquired from image designation unit 203 to a first prompt, and when a second target impression different from the first target impression is acquired, image generation unit 220 changes the prompt acquired from image designation unit 203 to a second prompt.

[0033] More specifically, the image generation unit 220 changes the acquired prompt to the first prompt or the second prompt based on the difference between the target impression acquired from the target impression designation unit 204 and the impression of the prompt acquired from the image designation unit 203.

[0034] Furthermore, the image generation unit 220 may determine a prompt to be input to the generation AI based on the impression estimated from the image and the target impression. The image may be an image generated by the image generation AI using the prompt acquired from the image designation unit 203 or a changed prompt, or may be an image designated by the user in the image designation unit 203. The image generation unit 220 determines a prompt that is close to the target impression based on the difference between the target impression and the impression estimated from the image. When determining a prompt based on the impression estimated from an image generated by the image generation AI and the target impression, the image generation unit 220 determines a prompt that is close to the target impression by changing the prompt used to generate the image.

[0035] Fig. 3 is a software block diagram of the image generation unit 220. As shown in Fig. 3, the image generation unit 220 has an acquisition unit 301, an impression estimation unit 302, an evaluation unit 303, a modification unit 304, and a generation unit 305. Fig. 3 shows software blocks relating to the function of modifying the prompt acquired by the image generation unit 220 based on the target impression.

[0036] The acquisition unit 301 acquires the prompt specified by the image designation unit 203, the prompt change permission information set by the image designation unit 203, and the target impression specified by the target impression designation unit 204. The acquisition unit 301 stores the acquired prompt in RAM 108. This is to use the acquired prompt as a base prompt, which will be described later. The prompt change permission information is information indicating a user instruction as to whether or not to permit a prompt change. The acquisition unit 301 switches its operation depending on the content of the acquired prompt change permission information. That is, when the acquisition unit 301 acquires prompt change permission information indicating permission to change the prompt, the change unit 304 changes the prompt as necessary, and the generation unit 305 generates an image using the changed prompt (hereinafter referred to as the changed prompt). When the acquisition unit 301 does not acquire prompt change permission information indicating permission to change the prompt, the change unit 304 does not change the prompt, and the generation unit 305 executes image generation processing using the prompt acquired by the acquisition unit 301.

[0037] The impression estimation unit 302 estimates the impression of the prompt acquired by the acquisition unit 301. The impression estimation of the prompt can be performed using a machine learning model for text impression estimation generated in the text impression quantification process described later.

[0038] Furthermore, the impression estimation unit 302 estimates the impression of the image (generated image) generated by the generation unit 305. The image impression estimation can be performed using a machine learning model for image impression estimation generated in the image impression quantification process described below. Note that when the generation unit 305 generates images multiple times, the impression is estimated for each of the images generated in the multiple image generation processes.

[0039] The evaluation unit 303 calculates the difference (hereinafter also referred to as impression difference) and distance between the target impression acquired by the acquisition unit 301 and the impression of the prompt estimated by the impression estimation unit 302. The calculated impression difference represents the amount of change required to change the impression of the prompt to one closer to the target impression. In this embodiment, Euclidean distance is used as the distance (hereinafter, when simply referring to distance, it will be referred to as Euclidean distance). The smaller the value indicated by the distance, the closer the target impression is to the impression estimated from the prompt. Note that the distance calculated by the evaluation unit 303 is not limited to Euclidean distance, and any distance between vectors, such as Manhattan distance or cosine similarity, may be calculated. The evaluation unit 303 determines whether the calculated distance is greater than a predetermined threshold. If the distance calculated by the evaluation unit 303 is greater than the predetermined threshold, the evaluation unit 303 instructs the change unit 304 to change the prompt. If the distance calculated by the evaluation unit 303 is not greater than the predetermined threshold, the evaluation unit 303 does not instruct the change of the prompt.

[0040] The evaluation unit 303 also calculates the difference and distance between the target impression acquired by the acquisition unit 301 and the impression of the generated image estimated by the impression estimation unit 302, and associates the difference with the corresponding image. The calculated impression difference represents the amount of change required to change the impression of the generated image to the target impression. The smaller the value indicating the distance, the closer the target impression and the impression of the generated image are. The evaluation unit 303 determines whether all the distances calculated for each generated image are greater than a predetermined threshold. If all the distances are greater than the predetermined threshold, that is, if an image suitable for the target impression cannot be generated even after changing the prompt multiple times, the evaluation unit 303 displays a warning screen on the display 105, for example, indicating that it is difficult to change the prompt suitable for the target impression. The image generation process may then be stopped. Alternatively, the evaluation unit 303 may retain and acquire the top N generated images with the smallest distances that have been generated up to that point.

[0041] The change unit 304 changes the prompt acquired by the acquisition unit 301 to a prompt suitable for the target impression based on the evaluation result by the evaluation unit 303. The change unit 304 determines the changed prompt so that a value indicating the distance between the target impression and an impression estimated from the prompt changed by the change unit 304 (changed prompt) is smaller than a predetermined threshold.

[0042] The modified prompt determined by the modification unit 304 includes, for example, a base prompt that is a prompt used as a base, and one or more additional prompts to be added to the base prompt. The modification unit 304 determines the additional prompt based on the difference between the target impression and the impression estimated from the base prompt. In this embodiment, the additional prompt can be acquired from a prompt impression table (impression information) stored in advance in the HDD 104. The prompt impression table is information in which a character string (additional prompt) is pre-associated with a value indicating the impression of the character string (additional prompt). The modification unit 304 acquires from the prompt impression table a character string (additional prompt) for which the distance between the difference between the target impression and the impression estimated from the base prompt and the impression of the additional prompt is smaller than a predetermined threshold.

[0043] When determining the changed prompt, the change unit 304 selects a prompt to be used for generating content from one or more changed prompts. At this time, a screen for selecting a prompt to be used for generating content from one or more changed prompts may be displayed and the user's selection may be accepted. The screen for selecting a prompt will be described later.

[0044] The generation unit 305 acquires the prompt acquired by the acquisition unit 301 or the changed prompt from the change unit 304, generates a random number as an initial value to be input to the image generation AI, and inputs the acquired prompt and the generated random number to the image generation AI to generate an image. The image generation AI can use known technology for generating images from prompts. In this embodiment, Stable Diffusion is used as the image generation AI. Other known image generation AIs, including Midjourney (https: / / www.midjourney.com / home / ), may also be used, or image generation AIs to be developed in the future may also be used. Any technology can be used as long as it can generate an image according to the content of the input prompt. If the generation unit 305 acquires multiple prompts, the generation unit 305 generates an image for each acquired prompt and acquires multiple generated images.

[0045] Furthermore, generation unit 305 selects an image to actually be used on the poster from the generated images. That is, an image is generated by the generation AI using one or more changed prompts determined (changed) by change unit 304, and a generated image to be placed on the poster is selected from the one or more generated images. At this time, generation unit 305 may display a screen for selecting a generated image to be placed on the poster from one or more generated images, and may accept a selection by the user. The screen for selecting an image will be described later.

[0046] Returning to the explanation of Fig. 2, the software configuration of the poster generator 210 will now be described in detail.

[0047] The image acquisition unit 211 acquires one or more image data items specified by the user in the image designation unit 203 from a specified acquisition source. The image acquisition unit 211 outputs the acquired image data to the image analysis unit 212. Image acquisition sources include the HDD 104, storage areas on a network, etc. Examples of images to be acquired include still images, frame images extracted from videos, stock images created in advance for this application or provided by an image provision service, and images generated by a generation AI. The still images and frame images are acquired from an imaging device such as a digital camera or a smart device. The imaging device may be included in the poster generation device 100 or an external device. If the imaging device is an external device, the images are acquired via the data communication unit 108. As another example, the still images may be illustration images created using image editing software or CG images created using CG production software. The still images and extracted images may be images acquired from a network or a server via the data communication unit 108. Examples of images acquired from a network or server include social networking service images (hereinafter referred to as "SNS images"), stock images, externally provided images, and images generated using generation AI. The program executed by CPU 101 analyzes data attached to each image to determine the source of the image. For example, SNS images may be acquired from an SNS via an application, and the acquisition source may be managed within the application. Images are not limited to those described above, and may be other types of images.

[0048] The image analysis unit 212 performs an analysis process on the image data acquired from the image acquisition unit 211, and acquires information indicating image feature amounts. Specifically, the image analysis unit 212 performs an object recognition process, which will be described later, and acquires feature amounts of the image data. The image analysis unit 212 also associates the information indicating the acquired feature amounts with the image data and outputs the information to the layout unit 217.

[0049] The skeleton acquisition unit 213 acquires, from the HDD 104, one or more skeletons that meet the conditions specified by the poster creation condition specification unit 201, the text specification unit 202, and the image acquisition unit 211. In this embodiment, a skeleton is information that indicates the arrangement of content (character strings and images) and figures, etc., to be arranged on a poster.

[0050] FIG. 4 is a diagram illustrating an example of a skeleton. Three graphic objects 402, 403, and 404, one image object 405, and four text objects 406, 407, 408, and 409, in which text is arranged, are arranged on a skeleton 401 in FIG. 4(a). Each object is recorded with metadata necessary for generating a poster, in addition to the position, size, and angle indicating the placement location. FIG. 4(b) is a diagram illustrating an example of metadata. For example, text objects 406 to 409 hold, as metadata attributes, the type of text information to be arranged. Here, text object 406 indicates that a title will be arranged, text object 407 indicates that a subtitle will be arranged, and text objects 408 and 409 indicate that the main text will be arranged. Furthermore, graphic objects 402 to 404 hold, as metadata attributes, color scheme numbers (color scheme IDs) indicating the shape and color scheme pattern of the graphic. Here, the attributes of graphic objects 402 and 403 indicate that they are rectangular, and the attribute of graphic object 404 indicates that they are elliptical. Also, assume that color scheme number 1 is assigned to graphic object 402, and color scheme number 2 is assigned to graphic objects 403 and 404. Here, the color scheme number is information referenced when assigning colors, which will be described later, and indicates that different colors are assigned to different color scheme numbers. Note that the types of objects and metadata are not limited to these. For example, there may be a map object for placing a map, or a barcode object for placing a QR code (registered trademark) or barcode. Furthermore, metadata for character objects may include metadata indicating the line spacing and character spacing. The metadata may include the purpose of the skeleton and be used to control whether or not to use the skeleton depending on the purpose.

[0051] The skeletons may be stored in HDD 104 in, for example, CSV format, or in a DB format such as SQL. Skeleton acquisition unit 213 outputs one or more skeletons acquired from HDD 104 to skeleton selection unit 214.

[0052] The skeleton selection unit 214 selects one or more skeletons that match the desired impression specified by the desired impression specification unit 204 from the skeletons acquired from the skeleton acquisition unit 213, and outputs them to the layout unit 217. Since the layout of the entire poster is determined by the skeleton, preparing various types of skeletons in advance can increase the variety of posters that can be generated.

[0053] Color scheme selection unit 215 obtains from HDD 104 one or more color schemes that match the desired impression specified by desired impression specification unit 204, and outputs them to layout unit 217. A color scheme is a combination of colors to be used in a poster.

[0054] FIG. 5 is a diagram showing an example of a color scheme table. In this embodiment, a color scheme pattern is expressed as a combination of four colors. The color scheme ID column in FIG. 5 is an ID for uniquely identifying a color scheme pattern. The color 1 to color 4 columns represent colors in RGB order, with each RGB color value ranging from 0 to 255 ((R, G, B) = (0 to 255, 0 to 255, 0 to 255)). Note that, although this embodiment uses a color scheme pattern consisting of a combination of four colors, other numbers of colors may be used, or multiple numbers of colors may be mixed.

[0055] Font selection unit 216 selects one or more font patterns that match the desired impression specified by desired impression specification unit 204, obtains them from HDD 104, and outputs them to layout unit 217. A font pattern is a combination of at least one of a title font, a subtitle font, and a body font.

[0056] The layout unit 217 generates one or more poster data items equal to or greater than the specified number of posters to be created by laying out various data items for each of the one or more skeletons obtained from the skeleton selection unit 214. The layout unit 217 arranges the text obtained from the text specification unit 202 and the image data obtained from the image analysis unit 212 for each skeleton. The layout unit 217 also applies the color pattern obtained from the color pattern selection unit 215 and the font pattern obtained from the font selection unit 216. The layout unit 217 outputs the generated one or more poster data items to the poster impression estimation unit 218.

[0057] The poster impression estimation unit 218 estimates an impression for each of the multiple poster data acquired from the layout unit 217, and links the estimated impression (estimated impression) to each poster data. The poster impression estimation unit 218 then outputs one or more poster data linked with the estimated impression to the poster selection unit 219.

[0058] The poster selection unit 219 compares the target impression specified by the target impression specification unit 204 with each estimated impression of the multiple poster data associated with an estimated impression obtained from the poster impression estimation unit 218, and selects poster data associated with an estimated impression that is close to the target impression. The poster selection unit 219 selects the number of posters to be created specified by the poster creation condition specification unit 201 or more. At this time, the poster selection unit 219 selects the number of posters to be created or more in ascending order of the value (distance) indicating the difference between the target impression and the estimated impression. The closeness between the target impression and the estimated impression is determined based on the distance calculated from the difference in impression value for each impression factor. The selection result is saved in the HDD 104. The poster selection unit 219 outputs the selected poster data to the poster display unit 205.

[0059] The poster display unit 205 displays a poster image based on the poster data acquired from the poster selection unit 219 on the display 105. The poster image is, for example, bitmap data. Note that the poster generation unit 210 generates poster data for the number of posters specified by the poster creation condition specification unit 201 or more, and therefore, a list of previews of the poster images is displayed on the display 105. When the user clicks on a poster image with the pointing device 107, the poster image is selected.

[0060] The poster creation application may be provided with a function (not shown) that allows the user to edit the layout, color, shape, etc. of images, text, and graphics through additional operations after the generated result is displayed on the poster display unit 205, and further change the design to suit the user's needs. Also, if the application is provided with a function that prints the poster data stored in the HDD 104 using a printer according to the conditions specified in the poster creation condition specification unit 201, the user will be able to obtain a printed copy of the poster they have created.

[0061] <Example of display screen> 6A and 6B are diagrams showing examples of a generation condition setting screen 622 and a content setting screen 601 provided by the poster creation application. The generation condition setting screen 622 shown in FIG. 6A and the content setting screen 601 shown in FIG. 6B are displayed on the display 105. The user specifies the text and images, which are the content to be placed on the poster, the target impression of the poster to be created, and the poster creation conditions (size, number to be created, and use category) via the generation condition setting screen 622 and the content setting screen 601. The poster creation condition specification unit 201, the target impression specification unit 204, the image specification unit 203, and the text specification unit 202 acquire the user's specifications via these UI screens.

[0062] The impression sliders 608 to 611 on the generation condition setting screen 622 are operation objects that allow the user to set values ​​indicating the degree of each factor (hereinafter referred to as impression factor) of the target impression of the poster to be created. For example, the impression slider 608 is a slider that sets a value indicating the degree of the target impression related to the impression factor "luxury," and the target impression is set so that the more to the right the slider is slid, the more luxurious the poster will be, and the more to the left the poster will be slid, the less luxurious (cheaper) it will be. Furthermore, by combining the target impression factors set with each slider, an overall target impression is set that reflects not only the impression factor set with one slider, but also the impression factors set with the other sliders.

[0063] For example, if the impression slider 608 corresponding to the impression factor "luxury" is set to the right of the center, and the impression slider 611 corresponding to the impression factor "profoundness" is set to the left of the center, a poster with a high sense of luxury but low sense of profoundness, and an elegant impression, is generated. Alternatively, if the impression slider 608 corresponding to the impression factor "luxury" is set to the right of the center, and the impression slider 611 corresponding to the impression factor "profoundness" is set to the right of the center, a poster with a high sense of luxury and profoundness, and an elegant impression, is generated. In this way, by combining target impression factors indicated by multiple impression sliders, it is possible to set target impressions with different directionalities, such as a "refined" target impression and a "gorgeous" target impression, even if the common target impression factor of "luxury" is set.

[0064] In other words, the target impression is composed of and determined by multiple factors that indicate the impression, but it may also be determined by a single factor that indicates the impression. In this embodiment, the state in which the slider is set to the far left is set to -2, and the state in which it is set to the far right is set to +2, and the value indicating the impression is corrected to a value between -2 and +2. These numerical values ​​indicate that the impression is low (-2), slightly low (-1), neutral (0), slightly high (+1), and high (+2). Note that the purpose of correcting to -2 to +2 is to align the scale with the estimated impression to facilitate distance calculation, which will be described later, but this is not limiting and normalization using values ​​between 0 and 1 may also be used.

[0065] Radio buttons 612 are buttons that can control whether the settings of each impression factor are enabled or disabled. The user can enable or disable each setting of each impression factor by pressing radio button 612 to set it to on / off. For example, by selecting off with radio button 612, that impression factor is excluded from impression control. For example, if the user wants to create a subdued poster with low dynamism but has no particular specifications for other impressions, the user can generate a poster that focuses on low dynamism by turning off the radio buttons 612 other than dynamism. Note that FIG. 6 shows a state in which luxury and familiarity are set to on, and dynamism and profoundness are set to off. This allows for highly flexible control, such as whether all impression factors or only some impression factors are used to generate a poster. Note that if the state in which each slider is set to the leftmost position is equivalent to a state in which the corresponding impression factor is not set, radio button 612 may be omitted. In this case, the user can disable the setting of each target impression by setting each slider to the leftmost position.

[0066] The size list box 613 is a list box for setting the size of the poster to be created. By clicking with the pointing device 107, the user can display a list of poster sizes that can be created and select one. The number to create box 614 can set the number of posters to be created. The category list box 615 can set the purpose category of the poster to be created.

[0067] The reset button 616 is a button for resetting each setting information on the generation condition setting screen 622. The next button 617 is a button for transitioning to the content setting screen 601 shown in FIG.

[0068] When the user presses the Next button 617, the displayed screen switches to the content setting screen 601. The poster creation condition specification unit 201 and the target impression specification unit 204 output the information set on the generation condition setting screen 622 to the poster creation unit 210. At that time, the poster creation condition specification unit 201 acquires the size of the poster to be created from the size list box 613, the number of posters to be created from the number to be created box 614, and the purpose category of the poster to be created from the category list box 615. The target impression specification unit 204 acquires the target impression of the poster to be created from the impression sliders 608 to 611 and the radio button 612. Note that the poster creation condition specification unit 201 and the target impression specification unit 204 may process the values ​​set on the generation condition setting screen 622. For example, the target impression specification unit 204 may correct the values ​​of the target impression specified by the impression sliders 608 to 611.

[0069] The title box 602, subtitle box 603, and main text box 604 on the content setting screen 601 accept user specification of text information to be placed on the poster. Note that in this embodiment, three types of text information are accepted, but this is not limiting. For example, additional text information such as location, date and time may also be accepted. Also, text information does not need to be entered in all boxes, and some boxes may be left blank.

[0070] The image designation area 605 is an area for designating and displaying an image to be placed on the poster. An image 606 represents a thumbnail of the designated image. An add image button 607 is a button for adding an image to be placed on the poster. When the user presses the add image button 607, the image designation unit 203 displays an image designation screen 701 for selecting an image file and accepts the image file selection by the user. A thumbnail of the selected image is then added to the image designation area 605.

[0071] Here, the image designation screen 701 will be described with reference to Fig. 7. The image designation screen 701 is displayed on the display 105. The user can use the image designation screen 701 to designate the acquisition destination of an image to be placed on a poster, or the acquisition destination of an image folder containing multiple images. The image designation unit 203 acquires the settings from the user through this UI screen.

[0072] Radio buttons 702 to 706 are buttons for setting the method for specifying candidate image data. The user can set the method for specifying image data by pressing radio buttons 702 to 706 to set on / off. Although multiple radio buttons are displayed, only one radio button can be set to on. In other words, when an off radio button is set to on, that radio button is set to on, but any radio button that was on before the setting is automatically set to off.

[0073] Radio button 702 is a button for setting a method for specifying one or more image data as a method for specifying an image. Specification box 708 accepts the specification of one or more image data. The user can specify candidate image data by specifying the file path of the image data in specification box 708. Browse button 709 is a button for specifying one or more image data. When the user presses Browse button 709, image specification unit 203 displays a dialog screen for selecting a file saved in HDD 104 and accepts the user's selection of an image file.

[0074] Radio button 703 is a button for setting a method for specifying an image by specifying a folder containing one or more image data as the acquisition destination for the image group. Specification box 710 accepts the specification of a folder containing one or more image data. By specifying a folder path in specification box 710, the user can specify all image data contained in the folder as the image group. Browse button 711 is a button for specifying the acquisition destination folder. When the user presses Browse button 711, image specification unit 203 displays a dialog screen for selecting a folder saved in HDD 104 and accepts the folder selection by the user.

[0075] Radio button 704 is a button for setting a method for specifying an application material image as a method for specifying an image. Designation box 712 displays the name of the application material image specified via browse button 713. Browse button 713 is a button for specifying one or more application material images. When the user presses browse button 713, image designation unit 203 displays a dialog screen for selecting an application material image and accepts the image selection by the user. Note that if tag information is assigned to the application material image, a configuration may be adopted in which all application material images with the tag can be selected at once by specifying the tag.

[0076] The radio button 705 is a button for setting a method for specifying a linked material image as a method for specifying an image. The specification box 714 displays the name of the linked material image specified through the browse button 715. The browse button 715 is a button for specifying one or more linked material images. When the browse button 715 is pressed by the user, the image specification unit 203 displays a dialog screen for selecting a linked material image and accepts the image selection by the user. Note that if tag information is assigned to the linked material image, it may be possible to select all linked material images with the tag by specifying the tag.

[0077] Radio button 706 is a button for setting a method for generating an image using an image generation AI as a method for specifying an image. Prompt box 716 accepts specification of a prompt to be used as input for the image generation AI. Then, image specification unit 203 generates an image using the specified prompt and image generation AI, and saves the generated image in HDD 104. Then, image specification unit 203 specifies the file path of the saved AI-generated image.

[0078] Check box 719 is a box for setting prompt change permission information that indicates whether or not automatic change of the prompt specified in prompt box 716 is permitted during image generation processing. If check box 719 is checked, image designation unit 203 designates prompt change permission information that permits change of the prompt. If check box 719 is not checked, image designation unit 203 designates prompt change permission information that does not permit change of the prompt.

[0079] The cancel button 717 is a button for canceling the designation of an image. When the cancel button 717 is pressed, each setting information on the image designation screen 701 is ignored, and the screen displayed on the display 105 transitions to the content setting screen 601. When the user presses the OK button 718, the screen displayed on the display 105 transitions to the content setting screen 601. At this time, thumbnails of one or more images designated on the image designation screen 701 are added to the image designation area 605 of the content setting screen 601. Note that if the OK button 718 is pressed while the radio button 706 indicating AI image generation on the image designation screen 701 is on, the image generation process shown in FIG. 15 is executed, and then the screen transitions to a prompt selection screen (FIG. 8) or an image selection screen (FIG. 9), which will be described later.

[0080] Returning to Fig. 6(b), the back button 623 is a button for canceling the designation on the content setting screen 601 and returning to the generation condition setting screen 622. The reset button 620 is a button for resetting each setting information on the content setting screen 601.

[0081] When the user presses OK button 621, text designation unit 202 and image designation unit 203 output the content (text information and image) set on content setting screen 601 to poster generation unit 210. At this time, image designation unit 203 acquires the image file path to be placed on the poster from image designation area 605. Text designation unit 202 acquires text information to be placed on the poster from title box 602, subtitle box 603, and body box 604. Note that text designation unit 202 and image designation unit 203 may process the values ​​set on content setting screen 601. For example, text designation unit 202 may remove unnecessary blank characters at the beginning or end of the input text information.

[0082] 8(a) is a diagram showing an example of a prompt selection screen displayed on the display 105 when the prompt has been changed by the image generation unit 220. When the OK button 718 is pressed with the radio button 706 and check box 719 enabled on the image specification screen 701 and the prompt is changed in the image generation process, the screen displayed on the display 105 transitions to a prompt selection screen 810.

[0083] The prompt selection screen 810 is a screen that is displayed when there is one or more types of prompts after the change. The user can specify one or more prompts for generating an image via the prompt selection screen 810.

[0084] The prompt selection screen 810 displays a plurality of prompts 812 that have been changed by the change unit 304 of the image generation unit 220 side by side. A check box 813 is also displayed for each prompt. The user can select (ON) the prompt by clicking the check box 813 corresponding to the prompt they want to use with the pointing device 107. Note that multiple check boxes 813 can be turned ON. The prompt 812 includes multiple prompts before and after the change. The cancel button 806 is a button for canceling the selection of a prompt. When the cancel button 806 is pressed, the processing by the image generation unit 220 is stopped, and the screen displayed on the display 105 transitions to the image specification screen 701. When the user presses the OK button 807, the processing by the image generation unit 220 is resumed using the specified prompt.

[0085] Furthermore, if there is only one type of prompt after the change, the image generation unit 220 may display a prompt selection screen 801 shown in FIG. 8(b). The user can specify one prompt to be specified for image generation via the prompt selection screen 801. A display box 802 is a box that displays the prompt before the change that the user input in the prompt box 716 in the image specification screen 701 of FIG. 7. A display box 803 is a box that displays the prompt after the change that has been changed by the image generation unit 220. A radio button 804 is a button that specifies the prompt displayed in the display box 802 as the prompt to be used for image generation. A radio button 805 is a button that specifies the prompt displayed in the display box 803 as the prompt to be used for image generation.

[0086] 9 is a diagram showing an example of an image selection screen on which an image generated by the image generation unit 220 is displayed on the display 105. When the OK button 807 is pressed on the prompt selection screen 801 or the prompt selection screen 810 shown in FIG. 8 and image generation is completed, the screen displayed on the display 105 transitions to the image selection screen 901.

[0087] One or more generated images 902 generated by the image generation unit 220 are displayed side by side on the image selection screen 901. Since the image generation unit 220 generates one or more images, the generated images 902 are also displayed as a list on the image selection screen 901. When the user designates one of the generated images 902 with the pointing device 107, the designated generated image 902 is selected and a check mark 903 is displayed. Note that multiple generated images 902 can be selected.

[0088] The information display area 904 is an area that displays information related to image generation. In this embodiment, the prompt and random number used to generate the corresponding generated image 902 are displayed as information related to image generation. The random number is a value input to the generation AI as an initial value for the generated image 902. The cancel button 905 is a button for canceling the selection of the generated image 902. When the cancel button 905 is pressed, the selection of the generated image 902 is canceled, and the screen displayed on the display 105 transitions to the prompt selection screen 810 or the prompt selection screen 801. Note that the transition destination screen is not limited to the prompt selection screen 810 or the prompt selection screen 801; the image generation process may be canceled and the transition to the image specification screen 701 may be performed. When the user presses the OK button 906, the image generation unit 220 saves the selected generated image 902 in the HDD 104 and transitions the screen displayed on the display 105 to the content setting screen 601.

[0089] 10 is a diagram showing an example of a poster preview screen 1001 on which a poster image 1002 generated by the poster display unit 205 is displayed on the display 105. When the OK button 621 on the content setting screen 601 is pressed and the poster generation is completed, the screen displayed on the display 105 transitions to the poster preview screen 1001.

[0090] Poster image 1002 is a poster image output by poster display unit 205. Poster generation unit 210 generates poster data in the number of pieces specified by poster creation condition specification unit 201 or more, and therefore displays a list of poster images 1002 in the number corresponding to the generated poster data. When the user clicks on one of the poster images 1002 with pointing device 107, the poster data corresponding to poster image 1002 is selected.

[0091] An edit button 1003 is a button for transitioning to a function for editing the selected poster data. With the edit function, the poster data can be edited through a UI (not shown).

[0092] A print button 1004 is a button for transitioning to a function for printing the selected poster data. With the print function, the poster data can be printed via a control UI of a printer (not shown).

[0093] <Quantifying the impression of posters> Here, a process for quantifying the impression of a poster (hereinafter referred to as poster impression quantification process) will be described. The poster impression quantification process is a pre-processing required for executing the poster impression estimation process (S1412 in FIG. 14) described later.

[0094] The poster impression quantification process is performed by a vendor or the like who develops the poster creation application during the development stage of the poster creation application. The poster impression quantification process may be executed by the poster generation device 100, or may be executed by an information processing device different from the poster generation device 100. When executed by an information processing device different from the poster generation device 100, the process is executed by the CPU of that information processing device.

[0095] The poster impression quantification process quantifies the impressions people have of various posters. At the same time, a correspondence between the poster image and the poster impression is derived. This makes it possible to estimate the impression of the poster from the generated poster image. If the impression can be estimated, it becomes possible to control the impression of the poster by modifying the poster image, or to search for a poster image that has a certain target impression. The poster impression quantification process is performed, for example, by running an impression learning application in a poster generation device to learn the impression of the poster image in advance of the poster generation process.

[0096] Fig. 11 is a flowchart showing the poster impression quantification process. The flowchart shown in Fig. 11 is realized, for example, by CPU 101 reading a program stored in HDD 104 into RAM 103 and executing it. The poster impression quantification process will be described with reference to Fig. 11. Note that the symbol "S" in the description of each process represents a step in the flowchart (the same applies hereinafter in this specification).

[0097] In S1101, CPU 101 acquires subjective evaluations of impressions of a poster. FIG. 12 illustrates an example of a method for subjectively evaluating impressions of a poster. CPU 101 presents a poster to a subject and acquires subjective evaluations of the impressions of the poster from the subject. Measurement methods such as the Semantic Differential (SD) method and the Likert scale can be used. FIG. 11 illustrates an example of a questionnaire using the SD method, in which adjective pairs expressing impressions are presented to multiple evaluators and scores are assigned to the adjective pairs evoked by the target poster. After acquiring subjective evaluation results from multiple subjects for multiple posters, CPU 101 calculates the average of the responses to each adjective pair and sets the average as a representative score for the corresponding adjective pair. Note that a subjective evaluation method other than the SD method may be used as long as words expressing impressions and corresponding scores are determined.

[0098] In S1102, CPU 101 performs a factor analysis of the subjective evaluation results acquired in S1101. If the subjective evaluation results are left as they are, the number of adjective pairs will increase the number of dimensions, making control complex. Therefore, it is desirable to reduce the number of dimensions to an efficient level using an analytical method such as principal component analysis or factor analysis. In this embodiment, the description will be given assuming that the dimensions have been reduced to four factors using factor analysis. Naturally, this number varies depending on the adjective pairs selected for the subjective evaluation and the factor analysis method. Furthermore, the output of the factor analysis is assumed to be standardized. That is, each factor is scaled so that the mean and variance of the poster used in the analysis are 0 and 1, respectively. This allows the impressions of -2, -1, 0, +1, and +2 specified by the target impression specification unit 204 to directly correspond to -2σ, -1σ, the mean value, +1σ, and +2σ for each impression, facilitating the calculation of the distance between the target impression and the estimated impression, as described below. In this embodiment, the four factors shown in Figure 6(a) are luxury, familiarity, dynamism, and solidity, but these are names given for convenience in order to convey impressions to the user through the user interface, and each factor is composed of the interaction of multiple adjective pairs. Also, CPU 101 stores in HDD 104 a conversion formula (hereinafter referred to as "impression conversion formula") that converts the subjective evaluation results of each adjective pair into the value of each impression, obtained by factor analysis.

[0099] In S1103, the CPU 101 associates poster images with impressions. While it is possible to quantify posters that have been subjectively evaluated using the method described above, it is also necessary to estimate impressions for posters to be created in the future without subjective evaluation. The association between poster images and impressions can be achieved by learning a model that estimates impressions from poster images. Specifically, for example, deep learning methods using Convolution Neural Network (CNN) or Visual Transformer (ViT), or machine learning methods using decision trees, can be used. In this embodiment, the CPU 101 inputs a poster image and outputs four factors to perform supervised deep learning using CNN. That is, a deep learning model is created by learning the subjectively evaluated poster images and the corresponding impressions as correct answers, and an impression is estimated by inputting an unknown poster image into the learning model.

[0100] In S1104, the CPU 101 stores in the HDD 104 the model configuration and learned parameters of the deep learning model for impression estimation created in S1103.

[0101] The poster impression estimation unit 218 loads the deep learning model stored in the HDD 104 into the RAM 103 and executes it. The poster impression estimation unit 218 converts the poster data acquired from the layout unit 217 into an image, and estimates the impression of the poster by running the deep learning model loaded in the RAM 103 on the CPU 101 or the GPU 109. Note that, although a deep learning method is used in this embodiment, this is not limiting. For example, when using a machine learning method such as a decision tree, it is possible to extract feature amounts such as the average brightness value and edge amount of the poster image by image analysis, and create a machine learning model that estimates the impression based on the feature amounts.

[0102] <Quantifying content impressions> Next, with reference to FIG. 13, a process for quantifying the impression of content (hereinafter referred to as content impression quantification process) will be described. The content impression quantification process is a pre-processing for executing the prompt impression estimation process (S1503 in FIG. 15) and the image impression estimation process (S1510 in FIG. 15). Hereinafter, both text and images will be collectively referred to as "content." The content impression quantification process is performed by a vendor or the like that develops a poster creation application during the development stage of the poster creation application. Note that the content impression quantification process may be executed by the poster generation device 100, or may be executed by an information processing device different from the poster generation device 100. Note that when executed by an information processing device different from the poster generation device 100, it is executed by the CPU of that information processing device.

[0103] The content impression quantification process derives a correspondence between the content itself and the impression of the content in a space where the impression of the poster is quantified. This makes it possible to find content that matches the impression of the poster you want to generate. The content impression quantification process is executed, for example, in a poster generation device by running an impression learning application for learning the impression of the content in advance of the poster generation process. Furthermore, the content impression quantification process must be executed after the poster impression quantification process because it uses the impression transformation formula acquired in the poster impression quantification process shown in Figure 11.

[0104] Fig. 13 is a flowchart showing the content impression quantification process. The flowchart shown in Fig. 13 is implemented, for example, by the CPU 101 reading a program stored in the HDD 104 into the RAM 103 and executing it. First, the image impression quantification process will be described with reference to Fig. 13(a).

[0105] In S1301, CPU 101 executes acquisition of subjective evaluations of impressions of images. The subjective evaluation may be performed using the same method as the subjective evaluation performed in quantifying impressions of posters. After acquiring subjective evaluation results from multiple subjects for multiple images, CPU 101 calculates the average value of the responses to each adjective pair and sets this average value as the representative score value for the corresponding adjective pair. Note that the subjective evaluation method for impressions may be other than the SD method, as long as words expressing impressions and their corresponding scores are determined.

[0106] In S1302, the CPU 101 acquires from the HDD 104 the impression conversion formula obtained by the factor analysis performed in the poster impression quantification process, and applies it to the subjective evaluation results acquired in S1301 to acquire an impression value for each image. By applying the impression conversion formula obtained in the poster impression quantification, the impression of the image can be quantified on a dimension that has the same meaning as the impression of the poster.

[0107] In S1303, the CPU 101 associates images with impressions. While the above-described method can quantify images that have been subjectively evaluated, the poster generation process of this embodiment requires that impressions be estimated for unknown images without subjective evaluation. The association between images and impressions can be achieved by learning a model that estimates impressions from images. Specifically, for example, deep learning methods using Convolution Neural Network (CNN) or Visual Transformer (ViT), or machine learning methods using decision trees, can be used. In this embodiment, the CPU 101 performs supervised deep learning using CNN, taking images as input and four factors as output. That is, a deep learning model is created by learning the subjectively evaluated images and the corresponding impressions as correct answers, and an impression is estimated by inputting an unknown image into the learning model.

[0108] In S1304, the CPU 101 stores in the HDD 104 the model configuration and learned parameters of the deep learning model for impression estimation created in S1303.

[0109] Next, the text impression quantification process will be described with reference to FIG.

[0110] In S1311, CPU 101 executes acquisition of subjective evaluations of impressions of the text. The subjective evaluation may be performed using the same method as the subjective evaluation performed in quantifying impressions of the poster. After acquiring subjective evaluation results from multiple subjects for multiple texts, CPU 101 calculates the average value of the responses to each adjective pair and sets this average value as the representative score value for the corresponding adjective pair. Note that the subjective evaluation method for impressions may be other than the SD method, as long as words expressing impressions and their corresponding scores are determined.

[0111] In S1312, CPU 101 obtains from HDD 104 the impression conversion formula obtained by the factor analysis performed in the poster impression quantification, and applies it to the subjective evaluation results obtained in S1311 to obtain an impression value for each piece of text. By applying the impression conversion formula obtained in the poster impression quantification, it is possible to quantify the impression of the text on a dimension that has the same meaning as the impression of the poster.

[0112] In S1313, CPU 101 associates text with impressions. While it is possible to quantify images that have been subjectively evaluated using the method described above, it is also necessary to estimate impressions for unknown text without subjective evaluation. Association between text and impressions can be achieved by training a model that estimates impressions from text, using, for example, a deep learning method using Transformer or a machine learning method using a decision tree. In this embodiment, CPU 101 inputs an image and performs supervised deep learning using Transformer, with four factors as output. That is, a deep learning model is created by learning the subjectively evaluated image and the corresponding impression as the correct answer, and an impression is estimated by inputting unknown text into the learning model.

[0113] In S1314, the CPU 101 stores in the HDD 104 the model configuration and learned parameters of the deep learning model for impression estimation created in S1313.

[0114] <Poster generation process> Fig. 14 is a flowchart showing the poster generation process executed by poster generation unit 210 of the poster creation application. The flowchart shown in Fig. 14 is realized, for example, by CPU 101 reading a program stored in HDD 104 into RAM 103 and executing it. In this embodiment, it is described that CPU 101 executes the poster creation application, causing each component shown in Fig. 2 to execute processing corresponding to each function and realize the function. As described above, the flowchart shown in Fig. 14 starts when the user sets various setting items on the poster creation application and presses the OK button.

[0115] In S1401, the poster creation application displays the generation condition setting screen 622 shown in Fig. 6(a) on the display 105. The user inputs each setting via the UI screen of the generation condition setting screen 622 using the keyboard 106 or the pointing device 107.

[0116] In S1402, the poster creation condition specification unit 201 and the target impression specification unit 204 acquire the corresponding settings from the creation condition setting screen 622. That is, the poster creation condition specification unit 201 acquires the poster size, number of posters to be created, and usage category designated by the user. The target impression specification unit 204 acquires the target impression designated by the user.

[0117] In S1403, the poster creation application displays the content setting screen 601 on the display 105. The text designation unit 202 and the image designation unit 203 accept the user's designation of text and image for each setting item displayed on the content setting screen 601. The user inputs setting values ​​for each setting item using the keyboard 106 or the pointing device 107. The image acquisition unit 211 acquires image data. Specifically, the image acquisition unit 211 reads an image file from the acquisition destination (e.g., HDD 104) designated by the image designation unit 203 into the RAM 103. The CPU 101 also acquires character information entered in the title box 602, subtitle box 603, and body box 604.

[0118] When the user wants to specify an image, he or she presses the add image button 607 on the content setting screen 601. When the add image button 607 is pressed, the image specification unit 203 displays the image specification screen 701 and accepts the user's selection of the image specification method. When the radio button 706 on the image specification screen 701 is set to on and the OK button 718 is pressed with a prompt for causing the image generation AI to generate an image entered in the prompt box 716, the image generation unit 220 starts the image generation process shown in FIG.

[0119] <Image generation processing> The image generation process will be described in detail using Fig. 15. Fig. 15 is a flowchart for explaining the image generation process in detail. The image generation process is executed by the image generation unit 220. Specifically, it is executed by the acquisition unit 301, impression estimation unit 302, evaluation unit 303, change unit 304, and generation unit 305 shown in Fig. 3.

[0120] In S1501, the acquisition unit 301 acquires the prompt specified by the user in the prompt box 716 on the image specification screen 701 and the prompt change permission information set in the check box 719. The acquisition unit 301 also stores the acquired prompt in the RAM 108.

[0121] In S1502, the acquiring unit 301 switches the subsequent processing depending on the contents of the prompt change permission information acquired in S1501. If the prompt change permission information acquired by the acquiring unit 301 indicates information that permits prompt change (S1502; YES), the process proceeds to S1503. If the information indicates information that does not permit prompt change (S1502; NO), the processes of S1503 to S1507 are skipped and the process proceeds to S1508.

[0122] In S1503, the impression estimation unit 302 estimates the impression of the prompt acquired in S1501. The impression of the prompt can be estimated using the stored learning model by the text impression quantification process shown in Fig. 13(b).

[0123] In S1504, the evaluation unit 303 calculates the difference (impression difference) and distance between the target impression acquired in S1402 and the impression of the prompt estimated in S1503. The calculated impression difference is used as the amount of change for changing the impression of the prompt to the target impression.

[0124] In S1505, the evaluation unit 303 determines whether the distance calculated in S1504 is greater than a predetermined threshold. If the distance calculated in S1504 is greater than the threshold, the evaluation unit 303 transitions to S1506, and if not, the evaluation unit 303 transitions to S1508.

[0125] In S1506, the change unit 304 changes the prompt acquired in S1501 to a prompt suitable for the target impression. Details of the prompt change process in S1506 will be described below with reference to Figures 16 and 17. In this embodiment, a method for determining a changed prompt is described in which a series of prompts is determined by adding one or more additional prompts that modify a base prompt.

[0126] FIG. 16 shows information associating character strings representing additional prompts with the impressions of those character strings. Hereinafter, this information will be referred to as a prompt impression table 1600. In the prompt impression table 1600 shown in FIG. 16, character strings are written in English because prompts used as input to the generation AI are typically written in English. Other languages, such as Japanese, may be used as long as the generation AI can support them. The columns for luxury, familiarity, dynamism, and profoundness contain values ​​indicating the degree to which each additional prompt affects each impression factor (luxury, familiarity, dynamism, profoundness). The numerical values ​​indicating the impressions corresponding to each additional prompt can be determined using a method similar to the text impression quantification described with reference to FIG. 13(b). Specifically, the impressions of the additional prompts can be derived by performing the processes of S1311 and S1312.

[0127] The impression of the additional prompt can also be derived by applying the impression estimation model saved in S1314 to the string representing the additional prompt. Furthermore, the impression corresponding to the additional prompt can also be derived by fixing the base prompt, synthesizing each additional prompt with the base prompt, generating an image using an image generation AI, and estimating the impression of the generated image. For example, first, an image is generated using an image generation AI with the base prompt "cafe" as input, and the impression of the generated image is estimated. Next, an image is generated using an image generation AI with the prompt "cute cafe" obtained by synthesizing the additional prompt "cute" with the base prompt, and the impression of the generated image is estimated. Then, by calculating the difference between the two impressions, the impression corresponding to the additional prompt in the base prompt can be obtained. The impression corresponding to the additional prompt can be calculated using various base prompts, and the impression corresponding to the additional prompt can be obtained by performing statistical processing such as averaging on the multiple impressions obtained. The additional prompt can be any word that influences the impression. For example, it can be a single word such as "cute" or "casual," or multiple words such as "pastel tone" or "soft lighting."

[0128] 17 is a flowchart for explaining the prompt change process of S1506. In explaining S1506, the operation at the first execution and the operation from the second execution onwards will be explained separately.

[0129] First, the operation at the time of first execution will be described. In S1701, the change unit 304 acquires an additional prompt from the prompt impression table 1600 based on the impression difference calculated in S1504 (the impression difference between the target impression and the impression of the prompt acquired in S1501). If the processing of S1701 is executed after the determination of YES in S1505, that is, if the impression distance between the target impression and the impression of the prompt is greater than a predetermined threshold, the change unit 304 refers to the impression difference calculated in S1504. Then, it calculates the distance between the impression difference and the impression of each additional prompt stored in the prompt impression table 1600. Then, the change unit 304 acquires the top N additional prompts from the prompt impression table 1600 in order of the smallest calculated distance value. In this embodiment, the change unit 304 acquires the top two additional prompts. Here, N may be set to a fixed value, or a box (not shown) for specifying the number of additional prompts to acquire may be provided on the image designation screen 701, and only the specified number of additional prompts may be acquired.

[0130] In this embodiment, the change unit 304 acquires an additional prompt based on the impression difference between the target impression and the impression estimated from the prompt, and the distance between the target impression and the impression of the additional prompt, but this is not limited to this. The change unit 304 may also acquire the top N additional prompts from the prompt impression table 1600 in order of shortest distance based on the distance between the target impression and the impression of the additional prompt.

[0131] In S1702, the change unit 304 initializes the prompt currently held in the RAM 103, and acquires a base prompt (hereinafter referred to as a base prompt) before the additional prompt is added.

[0132] (How to get the base prompt (1)) In this embodiment, the change unit 304 acquires the base prompt by reloading the prompt stored in RAM 108 in S1501. Note that the method for acquiring the base prompt is not limited to this. The base prompt may also be determined by any of the acquisition methods (2) to (4) described below.

[0133] (How to get the base prompt (2)) If the prompt acquired in S1501 already includes an additional prompt (character string) registered in the prompt impression table 1600, the change unit 304 initializes the prompt by deleting the additional prompt portion. That is, the character string obtained by deleting the additional prompt portion from the prompt acquired in S1501 is set as the base prompt. In this case, the portion that affects the impression originally included in the prompt acquired in S1501 can be excluded, so even if an additional prompt is added in S1703, inconsistency with the impression created by the original prompt can be prevented.

[0134] (How to get the base prompt (3)) The change unit 304 performs morphological analysis on the prompt acquired in S1501 and initializes the prompt by deleting the parts corresponding to adjectives. That is, the character string obtained by deleting the parts corresponding to adjectives from the prompt acquired in S1501 is set as the base prompt.

[0135] (How to get the base prompt (4)) The modification unit 304 initializes the prompt by syntactically analyzing the prompt acquired in S1501 and deleting the parts corresponding to the modifiers related to the subject. That is, the character string obtained by deleting the parts corresponding to the modifiers related to the subject from the prompt acquired in S1501 is set as the base prompt. In the cases (3) and (4) above, as in the case (2) above, it is possible to prevent inconsistencies between the impression given by the additional prompt and the impression given by the original prompt.

[0136] In S1703, the changing unit 304 changes the prompt by adding the additional prompt acquired in S1701 to the base prompt acquired in S1702.

[0137] (How to change the prompt (1)) In this embodiment, the modifying unit 304 modifies the base prompt to the number of additional prompts acquired by adding each additional prompt to the end of the base prompt, connecting them with a comma. For example, if the base prompt is "a cat" and the additional prompts are "cute" and "pretty," the modifying unit 304 modifies the base prompt to two prompts, "a cat, cute" and "a cat, pretty."

[0138] In the above prompt change method (1), in this embodiment, the same number of prompts as the number of acquired additional prompts are determined by adding one or more acquired additional prompts, but the prompt change method is not limited to this. The change unit 304 may determine the changed prompt using, for example, the following prompt change methods (2) to (6).

[0139] (How to change the prompt (2)) When there are multiple additional prompts, the change unit 304 may change the prompts by the number of combinations of the additional prompts. For example, if the acquired additional prompts are "cute" and "pretty," there are three possible combinations of additional prompts: "cute" and "pretty," which are a single additional prompt, and "cute, pretty," which is a combination of the two. In this case, the change unit 304 can change the prompts from the two acquired additional prompts to three patterns, which is the number of combinations of additional prompts. In this way, by adding multiple prompts that have similar tendencies in terms of impressions together, the influence of the additional prompts can be more strongly reflected.

[0140] (How to change the prompt (3)) In the prompt modification methods (1) and (2) above, the additional prompt is added to the base prompt by connecting it with a comma (,), but the modification method is not limited to this. A template for adding an additional prompt may be stored in HDD 104, and modification unit 304 may add the additional prompt to the template by reflecting it to the base prompt. For example, if the template is "with a (additional prompt) impression" and the additional prompt is "cute," modification unit 304 may add "with a cute impression" to the end of the base prompt. In this way, using a template makes it possible to more reliably modify the prompt as intended.

[0141] How to change the prompt (4) The modification unit 304 may also perform syntactic analysis of the base prompt and add an additional prompt by modifying the subject. For example, if the base prompt is "a cat" and the additional prompt is "cute," syntactic analysis can determine that "cat" is the subject, so the modification unit 304 may insert the additional prompt before "cat" and change it to "a cute cat." This allows the prompt to be modified in a more natural style.

[0142] How to change the prompt (5) Furthermore, the modification unit 304 may add a negative prompt that has a strong influence on an impression opposite to the target impression. A negative prompt is a prompt that instructs the AI ​​model not to include a specific element in a generated image. By setting a negative prompt that has a strong influence on an impression opposite to the target impression, it is possible to suppress the generation of an image that is not suitable for the target impression. In this case, in S1701, the modification unit 304 acquires the top N prompts with the greatest distance between the reference impression difference and the impression of each additional prompt shown in the prompt impression table 1600. Then, in S1703, the modification unit 304 sets the acquired prompts as negative prompts by connecting them with commas.

[0143] How to change the prompt (6) The modification unit 304 may also add a negation word to a prompt that has a strong influence on an impression opposite to the target impression. A negation word refers to a word that negates the word that follows it, such as "not" or "no." Negating a prompt that has a strong influence on an impression opposite to the target impression can suppress the generation of images that are not suitable for the target impression. In this case, in S1701, the modification unit 304 acquires the top N prompts with the greatest distance between the reference impression difference and the impression of each additional prompt indicated in the prompt impression table. Then, in S1703, the modification unit 304 adds a negation word before each acquired prompt and sets them connected by a comma. This concludes the description of the operation of S1506 during the first execution. This concludes the description of the first operation executed in S1506.

[0144] Next, the second and subsequent operations of S1506 will be described. When the second and subsequent operations are executed, it means that the previous prompt change did not generate an image suitable for the target impression. Therefore, the change unit 304 executes a change with processing content different from the prompt change already executed, aiming to obtain a prompt that can generate an image suitable for the target impression. The differences between the second and subsequent operations and the operation at the time of the first execution will be described.

[0145] In S1701, the change unit 304 acquires more additional prompts than in the first execution. In this embodiment, as an example, the change unit 304 acquires two more additional prompts than the number acquired in the previous execution. For example, in the second execution, the change unit 304 acquires the top four additional prompts with the smallest distance between the referenced impression difference and the impression of each additional prompt in the prompt impression table 1600. Note that the number of additional prompts to be acquired may be any number greater than in the previous execution. Furthermore, the change unit 304 may perform control such that the additional prompts acquired in the previous execution are not to be acquired.

[0146] In S1702, the change unit 304 performs the prompt initialization process in the same way as the first execution, and determines the base prompt.

[0147] In S1703, the change unit 304 changes the prompts by the number of combinations of the acquired additional prompts. For example, in the second execution, four additional prompts are acquired, so six patterns of additional prompts are possible.

[0148] The modifying unit 304 may also add an additional prompt. For example, the modifying unit 304 may add an emphasis prompt such as "very" or "extremely" or an inhibition prompt such as "a little" or "slightly" before the additional prompt. In this way, by emphasizing or inhibiting the influence of the additional prompt, the prompt can be fine-tuned to suit the target impression.

[0149] The modification unit 304 may also set a weight for the additional prompt. For example, in Stable Diffusion, an image generation AI, emphasis can be specified by adding parentheses () or square brackets ([]) to the target prompt. Emphasis or suppression can also be specified by adding a colon (:) and a number immediately after the target prompt. In Midjourney, another image generation AI, emphasis or suppression can also be specified by adding a double colon (::) and a number immediately after the target prompt. The modification unit 304 may modify the prompt using such emphasis or suppression expressions specific to image generation AIs. For example, when Stable Diffusion is used as the image generation AI, the modification unit 304 may modify the additional prompt to emphasize it by adding parentheses to the additional prompt or by adding a value greater than 1.0, such as ":1.2," immediately after the prompt. Similarly, the modification unit 304 may modify the additional prompt to suppress it by adding a value less than 1.0, such as ":0.8," immediately after the prompt. In this way, the influence of additional prompts can be emphasized or suppressed, allowing for fine tuning of the prompts to suit the target impression.

[0150] As described above, the following processes are possible for the prompt change process in S1506 from the second time onwards. (1) Obtain more additional prompts than the first time you run the program, and combine the additional prompts to add them to the base prompt. (2) Add prompts to indicate emphasis and inhibition. (3) Add and weight prompts that indicate model-specific emphasis or suppression.

[0151] The above is a description of the operation executed in S1506 from the second time onward. In S1506, the change unit 304 acquires one or more changed prompts. In this embodiment, the change unit 304 may select a predetermined number of changed prompts from one or more changed prompts. Specifically, the change unit 304 may estimate the impression of the changed prompt, calculate the distance from the target impression, and select and output the top N changed prompts in order of shortest distance. In this case, it is possible to prevent the number of changed prompts from becoming enormous, thereby reducing the selection load in prompt selection in S1507, which will be described later, and the image generation load in S1509.

[0152] In S1507, the change unit 304 selects a prompt to actually be used from prompt candidates including the pre-change prompt and one or more post-change prompts. In this embodiment, the change unit 304 displays a prompt selection screen 810 shown in FIG. 8(a) on the display 105 and accepts the user's selection of a prompt.

[0153] If S1507 is executed for the second or subsequent time, and if there is a prompt that is the same as the changed prompt selected in the previous execution, that prompt is excluded from the selection targets. Note that if there is only one type of changed prompt, the change unit 304 may display a prompt selection screen 801 shown in Fig. 8(b) on the display 105 to accept the prompt selection by the user. The prompt selection screen 801 shown in Fig. 8(b) allows the user to select either the before-change or after-change prompt, making it easy to select a prompt and reducing the operational burden on the user.

[0154] Furthermore, if there is only one type of prompt after the change, the change unit 304 may use the changed prompt to perform subsequent processing without displaying the prompt selection screen 801. In this case, the user does not need to perform any operation, thereby reducing the operational burden on the user.

[0155] In S1508, the generation unit 305 generates a random number as an initial value to be input to the image generation AI. Each time a random number is generated, a different value is generated.

[0156] In S1509, the generation unit 305 inputs the prompt acquired in S1501, or the changed or unchanged prompt selected in S1507, and the random number generated in S1508 into the image generation AI to generate an image. Note that the image generation AI may use a known technique for generating an image from a prompt, and detailed description of the image generation AI will be omitted. In this embodiment, Stable Diffusion is used as the image generation AI. Note that other known image generation AIs, such as Midjourney, may also be used, or an unknown image generation AI may also be used. Any technique that generates an image from a prompt in accordance with the content of the prompt may be used. Note that if multiple prompts are acquired, the generation unit 305 generates an image for each prompt and acquires multiple generated images. The generation unit 305 also counts the number of times an image has been generated and records the count in RAM 103. Note that if a different prompt is used through S1506 and S1507, the generation unit 305 resets the count of the number of times an image has been generated and starts counting again. Furthermore, the generating unit 305 may associate the generated image with information about the prompt and random number used when generating the generated image and store the information.

[0157] In S1510, the impression estimation unit 302 estimates the impression of the generated image generated by the generation unit 305 in S1509, and stores the estimated impression in association with the corresponding image.

[0158] In S1511, the evaluation unit 303 calculates the difference and distance between the target impression acquired in S1402 and the estimated impression of the generated image estimated in S1510, and stores the difference and distance in association with the generated image. The calculated difference represents the amount of change required to change the impression of the generated image to one closer to the target impression. Furthermore, the smaller the distance value, the closer the target impression and the impression of the generated image are.

[0159] In S1512, the evaluation unit 303 acquires the number of times the image has been generated that is recorded in the RAM 103, and determines whether or not the number of times the generation is greater than a predetermined threshold (upper limit). If the acquired number of times the generation is greater than the threshold (upper limit), the evaluation unit 303 transitions to S1513, and if the number of times the generation does not exceed the predetermined threshold (upper limit), the evaluation unit 303 transitions to S1508. In this embodiment, the evaluation unit 303 determines whether or not the number of times the generation is greater than five. Note that the threshold (upper limit) for the number of times the generation is performed may be any number equal to or greater than 1. The larger the threshold (upper limit), the more image generation results can be obtained, and the smaller the threshold (upper limit), the fewer image generation results can be obtained, but the processing time can be reduced.

[0160] In S1513, the acquisition unit 301 switches the subsequent processing depending on the prompt change permission information acquired in S1501. If the prompt change permission information acquired by the acquisition unit 301 indicates that a prompt change is permitted (S1513; YES), the process proceeds to S1514, and if the information indicates that a prompt change is not permitted (S1513; NO), the process proceeds to S1515.

[0161] In S1514, the evaluation unit 303 determines whether all distances calculated in S1511 (distances between the estimated impression of the generated image and the target impression) are greater than a predetermined threshold. If all distances calculated in S1511 are greater than the threshold, the evaluation unit 303 transitions to S1506, and if not (if there is at least one generated image for which the distance calculated in S1511 is equal to or less than the threshold), the evaluation unit 303 transitions to S1515.

[0162] Note that if the number of prompt changes in S1506 exceeds a predetermined upper limit and all distances calculated in S1511 are greater than a predetermined threshold, the evaluation unit 303 may proceed to processing different from this flowchart without returning to S1506. That is, if an image close to the target impression cannot be generated even after changing the prompt multiple times, the evaluation unit 303 may, for example, display a warning screen on the display 105 indicating that it is difficult to change the prompt to suit the target impression, and then cancel this flowchart. Alternatively, the evaluation unit 303 may retain and acquire the top N generated images with the smallest distances (the distances between the estimated impression of the generated image and the target impression) calculated in S1511 up to this stage, and proceed to S1515.

[0163] In S1515, the generation unit 305 selects an image to be actually used for the poster from the generated images. In this embodiment, the generation unit 305 displays the image selection screen 901 on the display 105, displays the generated image whose distance is determined to be smaller than the threshold in S1514 as the generated image 902, and accepts the user's image selection. In this embodiment, the generation unit 305 displays all generated images on the image selection screen 901 in order of the shortest distance associated with the generated image. Note that the display order and the number of images to be displayed are not limited to this. For example, the generation unit 305 may select and display a predetermined number of generated images in order of the shortest distance associated with the generated image. At this time, a box for specifying the number of images to be generated may be provided on the image specification screen 701, and only the specified number of images may be selected (not shown). Alternatively, the generated images may be displayed in a random order without referring to the distance associated with the generated image.

[0164] In this embodiment, a determination is made in both S1502 and S1513 as to whether to change the prompt, but the determination may be made in only one of S1502 and S1513. Specifically, when a determination is made in S1502 as to whether to change the prompt, the process may proceed to S1515 without making the determination in S1513. In this case, the image impression estimation performed in S1510 and the image impression difference and distance calculation performed in S1511 may be skipped. Similarly, when a determination is made in S1513 as to whether to change the prompt, the process may proceed to S1508 without making the determination in S1502. In this case, the prompt impression estimation performed in S1503 and the prompt impression distance calculation performed in S1504 may be skipped. By determining whether to change the prompt in only one of S1502 and S1513, it is possible to change the prompt based on the impression of either the input prompt or the generated image while reducing the processing load. This concludes the description of the image generation process executed in S1403. Returning to FIG.

[0165] Through the above-described processing of S1401 to S1403, a target impression is designated, and an image generated using a prompt that is close to the target impression is displayed in the image designation area 605 of the content setting screen 601.

[0166] In S1404, the number of selections is determined so that posters can be generated according to the number of creations specified in the poster creation condition specification unit 201. That is, the skeleton selection unit 214 determines the number of skeletons to select, the color scheme pattern selection unit 215 determines the number of color scheme patterns to select, and the font selection unit 216 determines the number of fonts to select. In this embodiment, the layout unit 217 generates poster data equal to the number of skeletons x the number of color scheme patterns x the number of fonts. The skeleton selection unit 214, the color scheme pattern selection unit 215, and the font selection unit 216 determine the number of skeletons, the number of color scheme patterns, and the number of fonts to select so that the number of posters to be generated is equal to or greater than the number of creations specified in the poster creation condition specification unit 201. For example, the number of skeletons, the number of color scheme patterns, and the number of fonts can each be determined according to the following equation 1.

[0167] TIFF2026014133000002.tif22150

[0168] For example, if the number of creations is 6, the number of selections is 3, and the layout unit 217 generates 27 pieces of poster data, of which the poster selection unit 219 selects 6 pieces of poster data. This allows the poster selection unit 219 to select posters whose overall impression more closely matches the target impression from among the poster data generated in excess of the number of creations. Note that the method for determining the number of selections is not limited to this, and other methods may also be used. The number of selections may also be a fixed value.

[0169] In S1405, the text designation unit 202 and the image designation unit 203 acquire their corresponding settings from the content setting screen 601. In addition, the image acquisition unit 211 acquires the image data designated by the image designation unit 203 or the image data generated by the image generation unit 220, and stores it in the RAM 103.

[0170] In S1406, the image analysis unit 212 performs an analysis process on the image data acquired in S1405 to acquire feature amounts or information indicating the features of the image. Examples of information indicating the features include meta information stored in the image. Examples of feature amounts include image feature amounts that can be acquired by analyzing the image. This information is used in object recognition processing, which is an analysis process. Note that in this embodiment, object recognition processing is executed as the analysis process, but this is not limited to this, and other analysis processes may also be executed. Furthermore, the process of S1406 may be omitted. Below, details of the process performed by the image analysis unit 212 in S1406 will be described.

[0171] The image analysis unit 212 executes object recognition processing on the image acquired in S1405. Here, a known method can be used for the object recognition processing. In this embodiment, objects are recognized by a classifier created by deep learning. The classifier outputs the likelihood of a pixel constituting the image being a pixel constituting each object as a value between 0 and 1, and recognizes that an object exceeding a certain threshold is present in the image. By recognizing the object image, the image analysis unit 212 can acquire the type and position of objects such as faces, pets such as dogs or cats, flowers, food, buildings, ornaments, and landmarks.

[0172] In S1407, the skeleton acquisition unit 213 acquires skeletons that meet the various set conditions. In this embodiment, it is assumed that each skeleton is described in one file and stored in the HDD 104. The skeleton acquisition unit 213 sequentially reads skeleton files from the HDD 104 to the RAM 103, leaves skeletons that meet the set conditions on the RAM 103, and deletes skeletons that do not meet the conditions from the RAM 103. Figure 14(b) is a flowchart of the condition determination process performed by the skeleton acquisition unit 213 in S1407. The condition determination process performed by the skeleton acquisition unit 213 will be described with reference to Figure 14(b).

[0173] In S1421, skeleton acquisition unit 213 determines whether the skeleton size matches the poster size specified by poster creation condition specification unit 201 for the skeleton read into RAM 103. Note that although it is confirmed that the sizes match here, it is also sufficient that the aspect ratios match. In this case, skeleton acquisition unit 213 acquires a skeleton that matches the poster size specified by poster creation condition specification unit 201 by enlarging or reducing the coordinate system of the read skeleton.

[0174] In S1422, the skeleton acquisition unit 213 determines whether the use category specified in the poster creation condition specification unit 201 matches the skeleton category. For skeletons that are used only for specific purposes, the use category is written in the skeleton file, and the skeleton is not acquired unless the corresponding use category is selected. This prevents a skeleton that is designed specifically for a specific purpose, such as one that uses graphics to evoke a school or a sporting goods pattern, from being used in other use categories. Note that if no use category is set on the generation condition setting screen 622, S1422 is skipped.

[0175] In S1423, skeleton acquisition unit 213 determines whether the number of image objects of the loaded skeleton matches the number of images acquired by image acquisition unit 211. If the number of image objects of the loaded skeleton matches the number of images acquired by image acquisition unit 211, the skeleton is left in RAM 103;

[0176] In S1424, the skeleton acquisition unit 213 determines whether the character object of the loaded skeleton matches the character information specified in the text designation unit 202. More specifically, it determines whether the type of character information specified in the text designation unit 202 is present in the skeleton. For example, assume that character strings are specified in the title box 602 and the body box 604 on the content setting screen 601, and that a blank is specified in the subtitle box 603. In this case, all character objects in the skeleton are searched, and if both a character object with "title" set as the type of character information in the metadata and a character object with "body" specified are found, the skeleton is deemed to be suitable; otherwise, it is deemed to be unsuitable. If the character object of the loaded skeleton matches the character information specified in the text designation unit 202, the skeleton acquisition unit 213 leaves the skeleton in RAM 103; if they do not match, the skeleton is deleted from RAM 103.

[0177] As described above, the skeleton acquisition unit 213 stores in RAM 103 skeletons whose skeleton size, usage category, number of image objects, and type of text objects all match the conditions set on the generation condition setting screen 622. In this embodiment, the skeleton acquisition unit 213 examines all skeleton files on HDD 104, but this is not limiting. For example, the poster creation application may store in advance in HDD 104 a database that associates the file paths of skeleton files with search conditions (skeleton size, number of image objects, and type of text objects). In this case, the skeleton acquisition unit 213 can quickly acquire skeleton files by reading only those skeleton files that match the results of a search in the database from HDD 104 to RAM 103. Return to FIG. 14(a).

[0178] In S1408, the skeleton selection unit 214 selects a skeleton that matches the target impression specified by the target impression specification unit 204 from the skeletons acquired in S1407. FIG. 18 is a diagram illustrating a method for the skeleton selection unit 214 to select a skeleton. FIG. 18(a) is a diagram illustrating an example of a table linking skeletons to impressions. The skeleton name column in FIG. 18(a) lists the skeleton file names, and the luxury, familiarity, dynamism, and profoundness columns represent numbers (numeric values) indicating the degree to which each skeleton influences each impression factor. These numerical values ​​indicate that the impression is low (-2), slightly low (-1), neutral (0), slightly high (+1), and high (+2), respectively. First, the skeleton selection unit 214 calculates the distance between the target impression acquired from the target impression specification unit 204 and each skeleton impression shown in the skeleton impression table in FIG. 18(a). For example, if the target impression is "luxury +1, familiarity -1, dynamism -2, profoundness +2," the distance calculated by the skeleton selection unit 214 will be as shown in FIG. 18(b). The smaller the distance value, the closer the target impression and the impression of the skeleton are. Next, the skeleton selection unit 214 selects the top N selection number of skeletons with the smallest distance values ​​in FIG. 18(b). In this embodiment, the skeleton selection unit 214 selects the top two skeletons. That is, the skeleton selection unit 214 selects skeleton 1 and skeleton 4.

[0179] The value of N is determined by the conditions specified in the poster creation condition specification unit 201. When it is a variable value, the selection number N may be determined using the above-mentioned formula 1, or may be determined by other methods. For example, if the number of creations is set to 6 in the creation number box 614 on the creation condition setting screen 622, the poster generation unit 210 generates six posters. The layout unit 217, which will be described later, generates a poster by combining skeletons, color patterns, and fonts selected in the skeleton selection unit 214, color pattern selection unit 215, and font selection unit 216. Therefore, for example, by selecting two skeletons, two color patterns, and two fonts, 2 × 2 × 2 = 8 posters can be generated, thereby satisfying the condition of 6 posters.

[0180] Furthermore, the value range of each impression in the skeleton impression table of FIG. 18(a) does not need to be the same as the value range of the impression specified by the target impression specifying unit 204. In this embodiment, the value range of the impression specified by the target impression specifying unit 204 is -2 to +2, but the value range of the impression in the skeleton impression table may be different from this. In this case, the value range of the skeleton impression table is scaled to match the value range of the target impression before the distance calculation is performed. Furthermore, the distance calculated by the skeleton selecting unit 214 is not limited to Euclidean distance, and any method that can calculate the distance between vectors, such as Manhattan distance or cosine similarity, may be used. Furthermore, impression factors set to off using the radio button 612 are excluded from the distance calculation.

[0181] The skeleton impression table is created in advance by estimating the impression of a poster image generated based on each skeleton, for example, by fixing the color scheme, font, and image and text data to be placed on the skeleton, and then saving it on HDD 104. That is, by estimating the impression of each poster image that uses the same text color and image, but has a different text and image arrangement, the relative characteristics compared to other skeletons are tabulated. In this process, it is desirable to cancel the impression due to the color scheme, image, etc. used, by standardizing the overall estimated impression or averaging the impressions of multiple poster images generated from a single skeleton using multiple color schemes and images. This allows the table to show, for example, how the impression of a skeleton with a small image is determined by elements such as graphics and text, regardless of the image, or how tilted image or text arrangement creates a stronger sense of dynamism.

[0182] FIG. 18(c) is an example of a skeleton corresponding to Skeleton 1 to Skeleton 4 in FIG. 18(a). For example, Skeleton 1 has image objects and text objects arranged in a regular pattern and a small image area, resulting in a low sense of dynamism. Skeleton 2 has circular graphic objects and image objects, resulting in a high sense of familiarity and a low sense of solidity. Skeleton 3 has large image objects and tilted graphic objects placed on top of the image objects, resulting in a high sense of dynamism. Skeleton 4 has images placed across the entire skeleton and minimizes text objects, resulting in a high sense of solidity and a low sense of dynamism. In this way, when a poster image includes text or images, poster images with different target impressions are generated depending on how the text or images are arranged. Note that the method of creating a skeleton impression table is not limited to this; it may be estimated from the characteristics of the arrangement information itself, such as the area and coordinates of the image and title string, or it may be adjusted manually. The skeleton impression table is stored in the HDD 104, and the skeleton selection unit 214 reads the skeleton impression table from the HDD 104 into the RAM 103 and refers to it.

[0183] In S1409, color scheme selection unit 215 selects a color scheme that matches the target impression specified by target impression specification unit 204. Using the same method as in S1407, color scheme selection unit 215 references the impression table corresponding to the color scheme and selects a color scheme according to the target impression. FIG. 19(a) shows an example of a color scheme impression table that links color scheme patterns with impressions. Color scheme selection unit 215 calculates the distance between the impressions indicated by the columns for luxury to dignity in FIG. 19(a) and the target impression, and selects the top N color scheme patterns with the smallest distance values. In this embodiment, the top two color scheme patterns are selected. Note that, like the skeleton impression table, the color scheme impression table can tabulate the impression trends of color scheme patterns by creating and estimating impressions of posters with different color scheme patterns while fixing the skeleton, font, and images other than the color scheme pattern.

[0184] In S1410, the font selection unit 216 selects a font combination that matches the target impression specified by the target impression specification unit 204. The font selection unit 216 references the impression table corresponding to the font in the same manner as in S1407 and selects a font according to the target impression. FIG. 19(b) shows an example of a font impression table that links fonts with impressions. The font selection unit 216 calculates the distance value between the impression indicated by the columns for luxury to dignity in FIG. 19(b) and the target impression, and selects the top N fonts with the smallest distance values. Note that, like the skeleton impression table, the font impression table can tabulate the tendency of font impressions by creating and estimating the impression of posters with different fonts after fixing the skeleton, color pattern, and image other than the font.

[0185] In S1411, the layout unit 217 sets character information, images, color schemes, and fonts for the skeleton selected by the skeleton selection unit 214, and generates a poster.

[0186] The layout processing in step S1411 and the software configuration of the layout unit 217 will be described in detail with reference to FIGS. 20, 21, 22, and 23. FIG.

[0187] 20 is an example of a software block diagram illustrating in detail the layout unit 217. The layout unit 217 includes a color allocation unit 2001, an image layout unit 2002, an image correction unit 2003, a font setting unit 2004, a text layout unit 2005, and a text decoration unit 2006.

[0188] Fig. 21 is a flowchart for explaining the layout processing of S1411 in detail. Fig. 22 is a diagram for explaining the information input to the layout unit 217. Fig. 22(a) is a table summarizing the character information specified in the text specification unit 202 and the image 2201 specified in the image specification unit 203. Fig. 22(b) is an example of a table showing the color scheme patterns acquired from the color scheme pattern selection unit 215, and Fig. 22(c) is an example of a table showing the fonts acquired from the font selection unit 216. Fig. 23 is a diagram for explaining the processing steps of the layout unit 217.

[0189] The layout process in step S1411 will be described in detail with reference to FIG.

[0190] In S2101, the layout unit 217 lists all combinations of skeletons acquired from the skeleton selection unit 214, color patterns acquired from the color pattern selection unit 215, and fonts acquired from the font selection unit 216. The layout unit 217 generates poster data for each combination in turn by performing layout processing from S2102 onwards. For example, if the number of skeletons acquired from the skeleton selection unit 214 is 3, the number of color patterns acquired from the color pattern selection unit 215 is 2, and the number of fonts acquired from the font selection unit 216 is 2, the layout unit 217 generates 3 x 2 x 2 = 12 pieces of poster data. Next, in S2101, the layout unit 217 selects one of the listed combinations and executes the processes of S2102 to S2107.

[0191] In S2102, the color scheme assignment unit 2001 assigns the color scheme pattern acquired from the color scheme pattern selection unit 215 to the skeleton acquired from the skeleton selection unit 214. FIG. 23(a) is a diagram showing an example of a skeleton. In this embodiment, an example will be described in which a color scheme pattern when the color scheme ID in FIG. 22(b) is 1 is assigned to the skeleton 2301 in FIG. 23(a). The skeleton 2301 in FIG. 23(a) is made up of two graphic objects 2302 and 2303, one image object 2304, and three text objects 2305, 2306, and 2307. First, the color scheme assignment unit 2001 assigns colors to the graphic objects 2302 and 2303. Specifically, based on the color scheme number, which is metadata written in the graphic object, the color scheme assignment unit 2001 assigns a corresponding color from the color scheme pattern. Next, the color scheme assignment unit 2001 selects a character object (Text<type=Title> ), for example, the last color in the color scheme pattern is assigned to the character placed in character object 2305. That is, in this embodiment, color 4 is assigned to the character placed in character object 2305. Next, for the character placed in character objects 2306 and 2307 whose metadata is type and whose attribute is other than "title", the character color is set based on the brightness of the background of the character object. In this embodiment, if the brightness of the background of the character object is equal to or lower than a threshold, the character color is set to white; otherwise, the character color is set to black. Figure 23(b) is a diagram showing the state of a skeleton 2308 after the above color scheme assignment process has been performed. The color scheme assignment unit 2001 outputs the colored skeleton data 2308 to the image arrangement unit 2002.

[0192] In S2103, the image arrangement unit 2002 arranges the image data acquired from the image analysis unit 212 in the skeleton data 2308 acquired from the color scheme assignment unit 2001 based on the accompanying analysis information. In this embodiment, the image arrangement unit 2002 assigns image data 2201 to an image object 2304 in the skeleton. Furthermore, if the aspect ratios of the image object 2304 and the image data 2201 differ, the image arrangement unit 2002 performs trimming so that the aspect ratio of the image data 2201 matches the aspect ratio of the image object 2304. More specifically, based on the object position obtained by the image analysis unit 212 analyzing the image data 2201, trimming is performed so that the object area reduced by trimming is minimized. Note that the trimming method is not limited to this, and other trimming methods may be used, such as trimming the center of the image or devising a composition so that the face position forms a triangular composition. The image placement unit 2002 outputs the skeleton data to which the images have been assigned to the image correction unit 2003 .

[0193] In S2104, the image correction unit 2003 obtains skeleton data with images assigned from the image placement unit 2002 and corrects the images placed on the skeleton. In this embodiment, if the image resolution is insufficient, upsampling processing is performed using super-resolution processing. First, the image correction unit 2003 determines whether the image placed on the skeleton meets a certain resolution. For example, assume that an image of 1600px x 1200px is assigned to a 200mm x 150mm area on the skeleton. In this case, the print resolution of the image can be calculated using Equation 2.

[0194] TIFF2026014133000003.tif18150

[0195] Next, if the image print resolution is determined to be less than the threshold, the image correction unit 2003 increases the resolution through super-resolution processing. On the other hand, if the image print resolution is determined to be equal to or greater than the threshold and has sufficient resolution, no image correction is performed. In this embodiment, the super-resolution processing is performed when the image print resolution is less than 300 dpi.

[0196] In S2105, the font setting unit 2004 sets the font acquired from the font selection unit 216 to the image-corrected skeleton data acquired from the image correction unit 2003. FIG. 23C shows an example of a font combination selected by the font selection unit 216. In this embodiment, an example of font assignment will be described in which the font ID "2" in FIG. 22C is assigned to the image-corrected skeleton data. In this embodiment, fonts are set for the character objects 2305, 2306, and 2307 of the skeleton 2308. Note that in posters, a font that is easy to see is often set for the title from the perspective of eye-catchingness, and a font that is easy to read from the perspective of visibility is set for the other characters. Therefore, in this embodiment, the font selection unit 216 selects two types of fonts: a title font and a body font. The font setting unit 2004 sets the title font to the character object 2305 whose attribute is "title," and sets the body font to the other character objects 2306 and 2307. The font setting unit 2004 outputs the skeleton data with the font set to the text layout unit 2005. While the font selection unit 216 selected two types of fonts in this embodiment, this is not a limitation; for example, only the title font may be selected. In this case, the font setting unit 2004 uses a font corresponding to the title font as the body font. That is, if the title is a Gothic font, a typical Gothic font with high readability should be selected for the other text objects, and if the title is a Mincho font, a typical Mincho font should be selected for the other text objects. Naturally, the title font and the body font may be the same. Alternatively, different fonts may be used depending on the degree of prominence desired, such as using a title font for the title and subtitle character objects and a body font for other character objects, or using a title font for a certain font size or larger.

[0197] In S2106, the text layout unit 2005 lays out the text specified by the text designation unit 202 in the skeleton data with font settings obtained from the font setting unit 2004. In this embodiment, each piece of text shown in FIG. 22(a) is assigned by referring to the metadata attributes of the character objects of the skeleton. That is, "Summer Big Thank You Sale," whose attribute is the title, is assigned to character object 2305, and "Blow Away the Midsummer Heat," whose attribute is the subtitle, is assigned to character object 2306. Because no text has been set, nothing is assigned to character object 2307. FIG. 23(c) shows a skeleton 2309, which is an example of skeleton data after processing by the text layout unit 2005. The text layout unit 2005 outputs the skeleton data 2309 with text layout to the text decoration unit 2006.

[0198] In S2107, the text decoration unit 2006 decorates the character objects in the skeleton with text already arranged, obtained from the text arrangement unit 2005. In this embodiment, if the color difference between the title character and its background area is equal to or less than a threshold, a process of adding a border to the title character is performed. This improves the readability of the title. The text decoration unit 2006 outputs the decorated skeleton data, i.e., the poster data for which the layout has been completed, to the poster impression estimation unit 218.

[0199] In S2108, the layout unit 217 determines whether poster data has been generated for all combinations. If the layout unit 217 determines that poster data has been generated for all combinations of skeletons, color patterns, and fonts, it ends the layout process and proceeds to S1412. If it determines that not all poster data has been generated, it returns to S2101 and generates poster data for combinations that have not yet been generated.

[0200] The layout process in step S1411 has been described above. Returning to the description of FIG.

[0201] In S1412, the poster impression estimation unit 218 performs rendering processing on each poster data acquired from the layout unit 217, estimates the impression of the rendered poster image, and associates the estimated impression with the poster data. The rendering processing is a process of converting poster data into image data. For example, even if the color scheme is the same, the layout will change if the skeleton is different, and the actual area in which each color is used will differ. Therefore, this processing is performed at this timing because it is necessary to evaluate not only the individual impression trends of the color scheme and skeleton, but also the impression of the final poster, including images and text. This makes it possible to evaluate not only the impression of individual elements of the poster, such as the color scheme and layout, but also the impression of the final poster, including the layout.

[0202] In S1413, the poster selection unit 219 selects a poster to be output to the display 105 (to be presented to the user) from the poster data and the estimated impression linked to the poster data acquired from the poster impression estimation unit 218. In this embodiment, the poster selection unit 219 selects a poster for which the distance between the target impression and the estimated impression of the poster is equal to or less than a predetermined threshold.

[0203] In this embodiment, Euclidean distance is used as the distance. The smaller the Euclidean distance, the closer the target impression and the estimated impression are. Furthermore, the distance calculated by the poster selection unit 219 is not limited to Euclidean distance; it can be calculated using any method that can calculate the distance between vectors, such as Manhattan distance or cosine similarity.

[0204] Furthermore, if the number of selected posters is less than the number of posters to be created specified by the poster creation condition specification unit 201, the poster selection unit 219 selects the missing posters in ascending order of the distance between the target impression and the estimated impression of the poster. In this embodiment, the poster selection unit 219 also selects the missing posters, but this is not limited to this. For example, if the number of posters selected by the poster selection unit 219 is less than the number of posters to be created, the poster preview screen 1001 ( FIG. 10 ) may display a message indicating that there are not enough posters. Alternatively, the poster selection unit 219 may select the missing posters and then display them on the poster preview screen 1001 so that posters whose distance between the target impression and the estimated impression is equal to or less than a threshold value can be distinguished from posters whose distance is equal to or greater than the threshold value. For example, if there are not enough selected posters, the process may return to S1404 and increase the number of skeletons, color patterns, and fonts to be selected.

[0205] In S1414, the poster display unit 205 renders the poster data selected by the poster selection unit 219 and outputs the poster image to the display 105. That is, the poster preview screen 1001 in FIG.

[0206] The above is a description of the poster generation process that changes the prompt based on the target impression specified by the user and then generates a poster based on the target impression. As described above, the poster creation application of the first embodiment changes the prompt based on information indicating the difference between the target impression specified by the user and the impression estimated from the prompt (the difference in impression and the distance). The prompt is also changed based on information indicating the difference between the target impression and the impression estimated from an image generated using the prompt (the difference in impression and the distance). This allows a different prompt to be determined depending on the target impression specified by the user, and an image generated using the determined prompt can be placed on the poster. Therefore, when generating content to be placed on a poster using a generation AI, the process of the user specifying the prompt and checking the generated image is automated, improving usability. Furthermore, because the content is generated to approximate the target impression specified by the user, the number of attempts required to generate content that meets the user's intentions can be reduced. Therefore, in an information processing device that generates data for a work such as a poster, usability for obtaining content to be placed on the work can be improved.

[0207] <<Modification of the First Embodiment>> 15, in S1508 to S1512, the generation of images with different random numbers (initial values ​​of the images) and the evaluation of the image impressions are repeated up to a predetermined upper limit number of times. After that, in S1513 to S1514, it is determined whether or not the prompt should be changed based on the results of the impression evaluation for all the generated images. However, the flow of the image generation process is not limited to this.

[0208] Fig. 24 is a flowchart for explaining in detail the modified image generation process of the first embodiment. Note that, among the processes in this flowchart, the processes denoted with the same reference numerals as those in the flowchart of Fig. 15 are the same as those in the first embodiment, and therefore their explanation will be omitted.

[0209] Steps S1501 to S1511 in the flowchart of Figure 24 are the same as those in the first embodiment. That is, the poster creation application calculates the impression difference and distance between the target impression specified by the user and the estimated impression of the input prompt, and if the distance value is greater than a predetermined threshold, modifies the input prompt based on the impression difference. The modified prompt and a random number are input to the generation AI to generate an image. The poster creation application estimates the impression of the image generated by the generation AI and calculates the distance between the target impression and the estimated impression of the generated image. Then, the process proceeds to S2401.

[0210] In S2401, the acquisition unit 301 switches the subsequent processing depending on the prompt change permission information acquired in S1501. If the prompt change permission information acquired by the acquisition unit 301 indicates information that permits prompt change (S2401; YES), the process proceeds to S2402, and if the information indicates information that does not permit prompt change (S2401; NO), the process proceeds to S1515.

[0211] In S2402, the evaluation unit 303 determines whether the distance calculated in S1511 is greater than a predetermined threshold. If the distance between the target impression calculated in S1511 and the estimated impression of the generated image is greater than the threshold, the evaluation unit 303 transitions to S2403, and if not, the evaluation unit 303 transitions to S1515.

[0212] In S2403, the generation unit 305 acquires the number of times the image recorded in RAM 103 has been generated and determines whether the number of times the image has been generated is greater than a predetermined threshold. If the acquired number of attempts is greater than the threshold, the evaluation unit 303 transitions to S1506 and executes prompt change processing. If the acquired number of attempts is not greater than the threshold, the evaluation unit 303 transitions to S1508 and generates a random number.

[0213] 24, an image is generated using one random number in S1508 to S1511, and each time an impression of the generated image is evaluated in S2401 to S2402 to determine whether or not to change the prompt. This not only achieves the effects of the first embodiment, but also makes it possible to complete the image generation process with the minimum number of image generation times required.

[0214] <<Second embodiment>> In the first embodiment, the impression of a prompt input by a user is compared with a target impression and evaluated, and the prompt is changed to suit the target impression based on the evaluation result. Furthermore, an example has been described in which an image is generated from the changed prompt and a poster is generated using the generated image. In the second embodiment, a poster creation application style-converts an image specified by a user (hereinafter referred to as a "specified image") so that the impression of the specified image is closer to the target impression. Then, a poster is generated using the converted image. At this time, the prompt used for style conversion is determined so that the impression of the image converted using the prompt (the converted image) is closer to the target impression. As a result, a prompt for converting the image specified by the user into an image suitable for the target impression is determined, and the image is style-converted using the determined prompt and used to generate a poster. Note that the image specified by the user includes image data stored in HDD 104, image data acquired via a network, application stock images, linked stock images acquired from an external image providing service, and AI-generated images generated by an image generation AI.

[0215] <Software block diagram> FIG. 25 is a software block diagram of a poster creation application according to the second embodiment. As shown in FIG. 25, the poster creation application includes a poster creation condition specification unit 201, a text specification unit 202, an image specification unit 203, a target impression specification unit 204, a poster display unit 205, an image generation unit 2502, an image conversion unit 2501, and a poster generation unit 210. Similar to the first embodiment, the poster generation unit 210 includes an image acquisition unit 211, an image analysis unit 212, a skeleton acquisition unit 213, a skeleton selection unit 214, a color scheme pattern selection unit 215, a font selection unit 216, a layout unit 217, a poster impression estimation unit 218, and a poster selection unit 219. The second embodiment differs from the poster generation unit 210 ( FIG. 2 ) of the first embodiment in that an image conversion unit 2501 has been added. The image generation unit 2502 is designated by a different reference numeral because the processing content of the image generation unit 2502 is different from that of the first embodiment. In FIG. 25, the components denoted by the same reference numerals as in FIG. 2 are the same as those in the first embodiment, and therefore the description thereof will be omitted.

[0216] Fig. 26 is a software block diagram of the image conversion unit 2501. As shown in Fig. 26, the image conversion unit 2501 has an image acquisition unit 2601, a prompt acquisition unit 2602, an impression estimation unit 302, an evaluation unit 303, a change unit 304, and a generation unit 2603. The impression estimation unit 302, the evaluation unit 303, and the change unit 304 in Fig. 26 are the same as the impression estimation unit 302, the evaluation unit 303, and the change unit 304 in the image generation unit 220 of the first embodiment, and therefore their description will be omitted.

[0217] The image acquisition unit 2601 acquires an image specified by the image designation unit 203. The images specified by the image designation unit 203 are the same as those in the first embodiment, and include image data stored in the HDD 104, application material images, linked material images, and AI-generated images. In the second embodiment, the image designation unit 203 displays an image designation screen 2801 for designating an image and accepts designation of an image by the user. The image designation screen 2801 will be described later. The image acquisition unit 2601 acquires the image designated by the user from the image designation unit 203, and acquires the desired impression designated by the user from the desired impression designation unit 204.

[0218] The prompt acquisition unit 2602 acquires a prompt based on the image acquired by the image acquisition unit 2601. The method of acquiring a prompt will be described later (FIGS. 31(b) to (d)). The generation unit 2603 converts the image acquired by the image acquisition unit 2601 into an image to be used for a poster, using the prompt acquired by the prompt acquisition unit 2602 and the image generation AI. Furthermore, the image conversion unit 2501 changes the prompt acquired by the prompt acquisition unit 2602 based on the difference between the target impression and the impression of the prompt acquired by the prompt acquisition unit 2602, or between the target impression and the impression of the image generated or converted by the image generation AI. The image conversion unit 2501 outputs the converted image converted by the generation unit 2603 to the image designation unit 203.

[0219] The image generation unit 2502 of the second embodiment acquires a prompt from the image designation unit 203, and generates an image to be used for a poster using the acquired prompt and the image generation AI. The image generation AI may be configured to be included in the poster creation application, or may not be included in the poster creation application and may instead use an external image generation AI service via the data communication unit 108. The image generation unit 2502 outputs the generated image to the image designation unit 203.

[0220] <Example of display screen> 27 to 29 are examples of screens displayed in the poster creation application of the second embodiment. FIG. 27(a) shows a generation condition setting screen 622, and FIG. 27(b) is a diagram showing an example of a content setting screen 2701. In FIG. 27, the components denoted by the same reference numerals as in FIG. 6 are the same as those in the first embodiment, and therefore description thereof will be omitted. The difference from the screen of FIG. 2 is that a conversion button 2704 is provided. The thumbnail of image 2702 displayed in the image designation area 605 is selected when a user performs a designation operation, and a check mark 2703 is displayed. The conversion button 2704 is a button operated to execute image conversion processing on the selected image 2702 with the check mark 2703. When the user presses the conversion button 2704, the image conversion processing shown in FIG. 31 is executed. The image conversion processing will be described later.

[0221] FIG. 28 is a diagram showing an example of an image designation screen 2801. In FIG. 28, components denoted by the same reference numerals as those in FIG. 7 are similar to the image designation screen 701 (FIG. 7) of the first embodiment, and therefore description thereof will be omitted. The second embodiment differs from the image designation screen 701 of the first embodiment in that the check box 719 for setting prompt change permission information is omitted. The second embodiment also differs from the first embodiment in the processing executed when the OK button 2802 is pressed. In the second embodiment, when the OK button 2802 is pressed by the user, the screen displayed on the display 105 transitions to the content setting screen 2701. At this time, thumbnails of one or more images 2702 designated on the image designation screen 2801 are added to the image designation area 605. Note that when the OK button 2802 is pressed while the radio button 706 indicating AI image generation on the image designation screen 2801 is on, the image generation processing shown in FIG. 30 is executed, and then the screen transitions to the content setting screen 2701.

[0222] 29 is a diagram showing an example of an image selection screen 2901. The image selection screen 2901 is a screen on which images converted by the image conversion unit 2501 are displayed, and is displayed on the display 105. When the conversion button 2704 on the content setting screen 2701 is pressed and the image conversion is completed, the screen displayed on the display 105 transitions to the image selection screen 2901.

[0223] An original image 2902 and one or more converted images 2903 are displayed on the image selection screen 2901. The original image 2902 is the image 2702 that was selected when the conversion button 2704 was pressed on the content setting screen 2701. In other words, it is the image before image conversion. The converted image 2903 is an image generated by the image conversion unit 2501. The image conversion unit 2501 generates one or more converted images, and therefore the image selection screen 2901 also displays a list of one or more converted images 2903. When the user operates the pointing device 107 or the like to designate one of the converted images 2903, the designated converted image 2903 is selected and a check mark 2904 is displayed. Note that multiple converted images 2903 may be selected.

[0224] An information display area 2905 is displayed near each converted image 2903. Information related to the image conversion is displayed in the information display area 2905. In this embodiment, the information related to the image conversion includes the prompt and random number used to convert the corresponding converted image 2903. The cancel button 2906 is a button for canceling the selection of the converted image. When the cancel button 2906 is pressed, the selection of the converted image is canceled, and the screen displayed on the display 105 transitions to the content setting screen 2701. When the user presses the OK button 2907, the poster creation application saves the selected converted image 2903 in the HDD 104, and the screen displayed on the display 105 transitions to the content setting screen 2701. Furthermore, a thumbnail of the converted image 2903 that is selected at the time the OK button 2907 is pressed is additionally displayed in the image designation area 605 of the content setting screen 2701. Note that the image 2702 that was selected on the content setting screen 2701 before the image conversion process was performed may be replaced with the converted image 2903 selected on the image selection screen 2901 .

[0225] <Processing flow> The flow of the poster generation process executed by the poster creation application in the second embodiment is similar to the poster generation process in the first embodiment shown in Fig. 14, and therefore a description thereof will be omitted. Here, the image generation process and image conversion process executed in the second embodiment will be described.

[0226] First, the image generation process executed in the second embodiment will be described with reference to Fig. 30. The image generation process is executed when the OK button 2802 is pressed while the radio button 706 for selecting AI image generation on the image specification screen 2801 is set to ON. As described above, the image specification screen 2801 transitions when the add image button 607 is pressed on the content setting screen 2701 displayed in S1403 of the poster generation process.

[0227] FIG. 30 is a flowchart for explaining the image generation processing of the second embodiment in detail. Note that the processes in FIG. 30 denoted by the same reference numerals as those in FIG. 15 are the same as those in the first embodiment, and therefore will not be described again. In the flowchart in FIG. 30, steps S1502 to S1507 and steps S1513 to S1514 shown in FIG. 15 are omitted. That is, in the second embodiment, image generation processing is executed in which processing related to prompt change is omitted. This image generation processing generates multiple images in response to a prompt specified by the user, and from the generated images, an image whose estimated impression is closest to the target impression is selected.

[0228] Next, the image conversion processing executed by the image conversion unit 2501 will be described in detail with reference to Fig. 31. Fig. 31(a) is a flowchart for describing in detail the image conversion processing executed in the poster generation processing of the second embodiment. The image conversion processing shown in Fig. 31(a) is executed when the conversion button 2704 is pressed with an image 2702 selected on the content setting screen 2701 displayed in S1403. Note that the processes in the flowchart of Fig. 31(a) that are given the same reference numerals as in Fig. 15 are the same as those in the image generation processing of the first embodiment (Fig. 15), and therefore their description will be omitted. The image conversion processing is executed by the image acquisition unit 2601, prompt acquisition unit 2602, impression estimation unit 302, evaluation unit 303, change unit 304, and generation unit 2603 of the image conversion unit 2501.

[0229] In S3101, the image acquisition unit 2601 acquires the image 2702 selected in the image designation area 605 on the content setting screen 2701.

[0230] In S3102, the prompt acquisition unit 2602 acquires a prompt that serves as a base when converting the image acquired in S3101. The prompt acquisition unit 2602 also stores the acquired prompt in the RAM 108. In this embodiment, the prompt acquisition unit 2602 acquires a prompt based on the impression of the image, in accordance with the flowchart shown in FIG. 31(b).

[0231] <Prompt acquisition process> The prompt acquisition process in S3102 will now be described. Figures 31(b), (c), and (d) are flowcharts for explaining in detail the prompt acquisition process executed in S3102. In this embodiment, the prompt acquisition process shown in Figure 31(b) is executed as an example.

[0232] (Prompt acquisition process (1)) In the prompt acquisition process shown in FIG. 31(b), in S3110, the impression estimation unit 302 estimates the impression of the image acquired in S3101 and links the estimated impression to the corresponding image (acquired image). In S3111, the evaluation unit 303 calculates the difference (impression difference) between the target impression acquired in S1402 and the impression of the image estimated in S3110, and links the estimated impression to the corresponding image. The calculated impression difference represents the amount of change required to change the impression of the acquired image to the target impression. In S3112, the prompt acquisition unit 2602 acquires a prompt based on a character string (a word or multiple words) stored as an additional prompt in the prompt impression table 1600 shown in FIG. 16, based on the impression difference calculated in S3111. The prompt acquisition unit 2602 calculates the distance between the referenced impression difference and the impression of each prompt shown in the prompt impression table 1600. Then, the prompt acquisition unit 2602 acquires the top N prompts in order of the smallest calculated distance. In this embodiment, the prompt acquisition unit 2602 acquires the top two prompts. Here, N may be set to a fixed value, or a box for specifying the number of prompts to acquire may be provided on the content setting screen 2701, and only the specified number of prompts may be acquired (not shown). Note that in this embodiment, the prompt is acquired based on the distance between the reference impression difference and the impression of the prompt, but this is not limiting. For example, the prompt acquisition unit 2602 may acquire a prompt based on the distance between the target impression and the impression of the prompt.

[0233] The method of obtaining a prompt is not limited to the above method (1). For example, the prompt obtaining unit 2602 may receive a prompt specification from the user in accordance with the flowchart shown in FIG.

[0234] (Prompt acquisition process (2)) In the prompt acquisition process shown in FIG. 31(c), in S3121, the prompt acquisition unit 2602 displays a prompt input screen 3201 shown in FIG. 32(a) on the display 105. The prompt box 3202 accepts the user's specification of a prompt to be used in image conversion. The cancel button 3203 is a button for canceling the specification of the prompt. When the cancel button 3203 is pressed, the image conversion process shown in FIG. 31(a) is canceled, and the screen displayed on the display 105 transitions to the content setting screen 2701. When the user presses the OK button 3204 on the prompt input screen 3201, the prompt specified in the prompt box 3202 is output to the prompt acquisition unit 2602.

[0235] 32(a), a content setting screen 3210 having a prompt box 3211 as shown in FIG. 32(b) may be displayed on the display 105. In this case, when the conversion button 2704 on the content setting screen 3210 is pressed, the prompt specified in the prompt box 3211 is also output to the image conversion unit 2501.

[0236] 32(a) or the content setting screen 3210 of Fig. 32(b), check boxes 3205 and 3212 for the user to set prompt change permission information may be provided, as in the first embodiment. The image conversion unit 2501 may perform control so as not to perform the prompt change process of S1506 when the prompt change permission information is set to off.

[0237] In S3122, the prompt acquisition unit 2602 acquires the prompt that was input via the prompt input screen 3201 or the content setting screen 3210 in S3121 and output to the image conversion unit 2501.

[0238] (Prompt acquisition process (3)) The prompt acquisition unit 2602 may also generate and acquire a base prompt from the image acquired by the image acquisition unit 2601 in S3101 according to the flowchart shown in FIG. 31(d). In the flowchart shown in FIG. 31(d), in S3131, the prompt acquisition unit 2602 generates and outputs a caption for the image from the image acquired by the image acquisition unit 2601 in S3101. Note that any known technique, such as CLIP (Contrasive Language-Image Pretraining), may be used to generate a caption from an image. In S3132, the prompt acquisition unit 2602 performs morphological analysis on the caption generated in S3131 and acquires adjectives included in the caption as a prompt. Note that if the caption does not include an adjective, the acquired prompt is treated as empty. In this case, the estimated impression of the prompt acquired in S1503 of FIG. 31(a) is considered to be 0 for all factors, and processing continues. This concludes the explanation of the prompt acquisition process in S3102 of Figure 31(a). Returning to Figure 31(a).

[0239] Through the processing of S3101 to S3102, an image and a target impression specified by the user are acquired, and a prompt for converting the acquired image is acquired. Next, in S1503, the impression of the acquired prompt is estimated by the impression estimation unit 302, and in S1504, the difference (impression difference) and distance between the target impression acquired in S3102 and the impression of the prompt estimated in S1503 are calculated by the evaluation unit 303. The calculated impression difference is used as a change amount for changing the impression of the prompt to the target impression.

[0240] In S1505, the evaluation unit 303 determines whether the distance calculated in S1504 is greater than a predetermined threshold. If the distance calculated in S1504 is greater than the threshold, the evaluation unit 303 proceeds to S1506, and if not, skips S1506 and S1507 and proceeds to S1508.

[0241] In S1506, the change unit 304 changes the prompt acquired in S3101 to a prompt suitable for the target impression. The prompt change process in S1506 is the same as in the first embodiment. For example, a series of prompts may be determined by adding one or more additional prompts that modify a base prompt, which is a main prompt. As in the first embodiment, any of the methods for determining the base prompt and changing the prompt may be used. In S1507, the change unit 304 selects a prompt to be used for style conversion from the pre-change or post-change prompt. As in the first embodiment, in this embodiment, the prompt selection screen 810 shown in FIG. 8(a) is displayed on the display 105 to accept the user's selection of a prompt. Note that the change unit 304 may select a post-change prompt and perform subsequent processing without displaying the prompt selection screen. In S1508, the generation unit 2603 generates an initial value (random number) to be input to the image generation AI.

[0242] Next, the process proceeds to S3103. In S3103, the generation unit 2603 inputs the prompt acquired in S3102 or the prompt selected in S1507, and the random number generated in S1508 into the image generation AI, and generates a new image by performing style conversion of the image. Here, style conversion refers to a technique for converting the style of the input image to a style to which the user wishes to convert. Style conversion methods can be broadly divided into a method for specifying a prompt representing the style to which the user wishes to convert, and a method for specifying an image representing the style to which the user wishes to convert. In this embodiment, the method for specifying a prompt representing the style to which the user wishes to convert is used. Note that image style conversion can be performed using known techniques, and a detailed description of image style conversion will be omitted.

[0243] When multiple prompts are acquired, the generation unit 2603 performs style conversion on the image using each prompt, and acquires multiple converted images. The generation unit 2603 also counts the number of times that converted images have been generated by changing the random number, and records the count in RAM 103. When style conversion is performed using the changed prompt after steps S1506 and S1507 in FIG. 31, the generation unit 2603 resets the count of the number of times the images have been generated and starts counting again. The generation unit 2603 may also associate the generated converted image with information on the original image, prompt, and random number used when generating the converted image.

[0244] After S3103, the process proceeds to S1510. In S1510, the impression estimation unit 302 estimates the impression of one or more converted images generated by the generation unit 2603 in S3103, and stores the estimated impression in association with the corresponding converted image.

[0245] In S1511, the evaluation unit 303 calculates the difference (impression difference) and distance between the target impression acquired in S3101 and the estimated impression of the converted image estimated in S1510, and stores the difference and distance in association with the converted image. The calculated impression difference represents the amount of change required to change the impression of the converted image to one closer to the target impression. Furthermore, the smaller the distance value, the closer the target impression and the impression of the converted image are.

[0246] In S1512, the evaluation unit 303 acquires the number of transformations of the image stored in the RAM 103, and determines whether the number of transformations is greater than a predetermined threshold (upper limit of the number of transformations). If the acquired number of transformations is greater than the threshold, the evaluation unit 303 transitions to S1514, and if it is not greater, the evaluation unit 303 transitions to S1508. In this embodiment, the evaluation unit 303 sets the upper limit (threshold) of the number of transformations to 5. Note that the threshold may be any number equal to or greater than 1, and the larger the threshold, the more image generation results can be acquired, while the smaller the threshold, the fewer style-converted images can be acquired but the processing time can be reduced.

[0247] In S1514, the evaluation unit 303 determines whether all distances calculated in S1511 (distances between the estimated impression of the converted image and the target impression) are greater than a predetermined threshold. If all distances calculated in S1511 are greater than the threshold (S1514; YES), the evaluation unit 303 transitions to S1506. If there is at least one converted image for which the distance calculated in S1511 is equal to or less than the threshold (S1514; NO), the evaluation unit 303 transitions to S3104. Note that if the evaluation unit 303 determines in S1512 that the number of image conversions has exceeded a certain number and if the evaluation unit 303 determines in S1514 that all distances calculated in S1511 are greater than the threshold, the evaluation unit 303 may perform an operation such as notifying the user. This means that an image close to the target impression could not be generated even after multiple style changes to the image. Therefore, the evaluation unit 303 may display a warning screen on the display 105 indicating that it is difficult to make prompt changes appropriate for the target impression, and then cancel the image conversion process shown in FIG. 31(a). Alternatively, the evaluation unit 303 may retain and acquire the top N converted images with the smallest distances calculated in S1511 up to that point, and proceed to S3104.

[0248] In S3104, the generation unit 2603 selects an image to be actually used for the poster from the converted images. In the second embodiment, the generation unit 2603 displays an image selection screen 2901 on the display 105, and displays a converted image for which the distance calculated in S1511 is determined to be equal to or less than a predetermined threshold as a converted image 2903 on the image selection screen 2901. Then, the generation unit 2603 accepts an image selection by the user. In this embodiment, the generation unit 2603 displays all converted images on the image selection screen 2901 in order of the shortest distance associated with the converted image. Note that the display order and the number of images to be displayed are not limited to this. For example, the generation unit 2603 may select a predetermined number of converted images in order of the shortest distance associated with the converted image, and display them on the image selection screen 2901. In this case, a box for specifying the number of converted images may be provided on the content setting screen 2701, and the number of converted images specified by the user may be selected (not shown). Furthermore, the converted images may be displayed on the image selection screen 2901 in a random order without referring to the distance associated with the converted image.

[0249] The image selected on the image selection screen 2901 is used to generate the poster as the content to be placed on the poster.

[0250] As described above, according to the second embodiment, a prompt for converting an image into a style closer to the target impression is determined based on an image specified by a user or an image generated by a generation AI, and one or more converted images are generated. Furthermore, if multiple converted images are generated, the image to be used for the poster can be selected from the generated converted images. This makes it easier to obtain content to place in a production, improving usability, because the image is converted into a style closer to the target impression in a poster production application that generates data for productions such as posters. Furthermore, it becomes easier to generate posters that are closer to the target impression.

[0251] <<Modification of the Second Embodiment>> In the second embodiment described above, a prompt used for image style conversion was acquired based on the target impression and image specified by the user, and the prompt was then modified to further approximate the target impression, after which the image was style converted. However, this process is not limited to this, and the prompt acquired in the prompt acquisition process of S3102 may be used for image generation rather than image style conversion. In a modified example of the second embodiment, an example will be described in which a poster creation application generates a prompt from an acquired image, modifies the prompt as necessary, and then generates an image using an image generation AI.

[0252] Fig. 33 is a flowchart for explaining in detail the image conversion processing in a modified example of the second embodiment. Note that, among the processes in this flowchart, the processes denoted with the same reference numerals as those in the flowchart of Fig. 31 are the same as the processes explained in the second embodiment, and therefore explanations thereof will be omitted. Note that the image conversion processing shown in Fig. 33 is executed when the conversion button 2704 is pressed with an image 2702 selected on the content setting screen 2701 displayed in S1403.

[0253] In S3301, the image acquisition unit 2601 acquires the image 2702 selected in the image designation area 605 of the content setting screen 2701, similar to S3101 in the image conversion processing of the second embodiment.

[0254] Next, the process proceeds to S3302. In S3302, the prompt acquisition unit 2602 generates a caption for the image acquired in S3301. As described above, any known method may be used for the process of generating a caption from an image. The prompt acquisition unit 2602 acquires the generated caption as a prompt. The prompt acquisition unit 2602 also stores the acquired prompt in RAM 108.

[0255] Thereafter, the process of S1503 to S1508 is executed as in the first and second embodiments. That is, the poster creation application calculates the impression difference and distance between the impression of the prompt acquired in S3302 and the target impression, and if the value of the distance is greater than a predetermined threshold, changes the prompt to one that is closer to the target impression based on the impression difference. The poster creation application also selects a prompt to be used for image generation from the changed prompts and generates a random number to be input to the image generation AI. Next, the process proceeds to S3303.

[0256] In S3303, the generation unit 2603 inputs the prompt acquired in S3302 or S1507 and the random number generated in S1508 to the image generation AI to generate an image. As described above, the image generation AI may use a Stable Diffusion model, a GAN model, Midjourney, or the like.

[0257] Thereafter, the processing of S1510 to S1515 is executed, as in the first and second embodiments. That is, the impression difference and distance between the impression of the image generated in S3303 and the target impression are calculated and linked to the generated image, and image generation is repeated until the number of generation times reaches a predetermined threshold (upper limit number of times). If the number of generation times exceeds the upper limit number and the distance values ​​calculated in S1511 for all generated images are greater than the predetermined threshold, the prompt is changed and image generation is repeated. If the number of image generation times exceeds the upper limit number and an image is generated whose distance value calculated in S1511 is equal to or less than the predetermined threshold, the processing transitions to S1515, where an image selection screen 2901 is displayed on the display 105, and one or more generated images are presented to the user. If the number of image generation times exceeds the upper limit number but no image whose distance value is equal to or less than the predetermined threshold is generated, a warning screen may be displayed and this flowchart may end.

[0258] As described above, according to the modified example of the second embodiment, a caption is generated from an image specified by the user and used as a prompt. The prompt is then changed to one that can generate an image suitable for the desired impression, and the image is then generated using an image generation AI. This allows a new image suitable for the desired impression to be generated while taking into account the content of the image specified by the user. Therefore, a new image that reflects the user's intentions but is not too restricted by the input image can be proposed to the user, improving usability for obtaining the intended content.

[0259] <<Third Embodiment>> In the first and second embodiments, a process for determining a prompt for generating or converting an image based on the impression of the prompt or the impression of an image generated based on the prompt was described. In the third embodiment, an example of performing a prompt change process based on the impression of a poster generated by a poster generation application is described. This allows the prompt change process to be performed with emphasis on the impression of the poster, which is the final product, making it easier to generate a poster that is closer to the intended design.

[0260] <Software block diagram> FIG. 34 is a software block diagram of a poster creation application according to the third embodiment. As shown in FIG. 34, the poster creation application includes a poster creation condition specification unit 201, a text specification unit 202, an image specification unit 3401, a target impression specification unit 204, a poster display unit 205, and a poster generation unit 210. The poster generation unit 210 includes a prompt modification unit 3402, an image generation unit 3403, an image acquisition unit 211, an image analysis unit 212, a skeleton acquisition unit 213, a skeleton selection unit 214, a color scheme selection unit 215, a font selection unit 216, a layout unit 217, a poster impression estimation unit 3404, and a poster selection unit 219. The third embodiment differs from the first embodiment ( FIG. 2 ) in that the prompt modification unit 3402 and the image generation unit 3403 are added. Furthermore, the image specification unit 3401 and the poster impression estimation unit 3404 perform different processing than the image specification unit 203 and the poster impression estimation unit 218 of the first embodiment. The components denoted by the same reference numerals as in FIG. 2 are the same as those in the first embodiment, and therefore the description thereof will be omitted.

[0261] The image designation unit 3401 accepts designation by the user of one or more pieces of image data to be arranged on a poster, as in the first embodiment. In addition, it outputs prompts designated on a setting screen 3501 (FIG. 35), which will be described later, to the image generation unit 3403. The image designation unit 3401 also outputs the sum of the designated number of images and the designated number of prompts to the skeleton acquisition unit 213 as the number of images to be used.

[0262] Image generation unit 3403 generates an image to be used for a poster using the prompt and image generation AI acquired from image specification unit 3401 or prompt change unit 3402. Note that the image generation AI may be configured to be included in the poster creation application, or may not be included in the poster creation application and may instead use an external image generation AI service via data communication unit 108. Image generation unit 3403 outputs the generated image to image analysis unit 212.

[0263] As in the first embodiment, the poster impression estimation unit 3404 estimates an impression for each of the multiple poster data acquired from the layout unit 217 and associates the estimated impression with each poster data. Then, the poster impression estimation unit 3404 outputs one or more poster data associated with the estimated impression to the poster selection unit 219. Furthermore, in the third embodiment, the poster impression estimation unit 3404 outputs the poster data associated with the estimated impression to the prompt change unit 3402.

[0264] The prompt modification unit 3402 acquires the specified prompt from the image designation unit 3401, and acquires the target impression from the target impression designation unit 204. The prompt modification unit 3402 also acquires poster data linked to the estimated impression of the poster and the prompt used for image generation from the poster impression estimation unit 3404. The prompt modification unit 3402 then modifies the acquired prompt so that the estimated impression of the poster becomes closer to the target impression, and outputs the modified prompt to the image generation unit 3403.

[0265] <Example of input screen> 35 is a diagram showing an example of a setting screen 3501 displayed on the display 105 in the poster generation process of the third embodiment. The setting screen 3501 is a screen that combines the generation condition setting screen 622 and the content setting screen 601 described in the first embodiment into a single screen. The components denoted with the same reference numerals as in FIG. 6 are the same as those in the first embodiment, and therefore will not be described again.

[0266] The prompt box 3502 is a box for accepting a user's specification of a prompt to be used as input to the image generation AI. The image specification unit 3401 outputs the prompt input into the prompt box 3502 to the prompt change unit 3402.

[0267] Add prompt button 3503 is a button that is pressed when an additional prompt box 3504 is to be displayed. Similar to prompt box 3502, additional prompt box 3504 accepts prompts specified by the user. The user can specify multiple prompts by inputting prompts into prompt box 3502 and additional prompt box 3504. Note that the method for specifying multiple prompts is not limited to this. For example, character information input into one prompt box 3502 may be divided by a line feed character, and each of the divided pieces of character information may be treated as a separate prompt.

[0268] Reset button 3505 is a button for resetting each setting information on setting screen 3501. When the user presses OK button 3506, poster creation condition specification unit 201, text specification unit 202, image specification unit 3401, and target impression specification unit 204 output the contents set on setting screen 3501 to poster generation unit 210. At this time, poster creation condition specification unit 201 acquires the size of the poster to be created from size list box 613, the number of posters to be created from number of creation box 614, and the purpose category of the poster to be created from category list box 615. Text specification unit 202 acquires character information to be placed on the poster from title box 602, subtitle box 603, and body box 604. Image specification unit 3401 acquires the image file path to be placed on the poster from image specification area 605. Target impression specification unit 204 acquires the target impression of the poster to be created from impression sliders 608 to 611 and radio buttons 612. The poster creation condition specifying section 201, the text specifying section 202, the image specifying section 3401, and the desired impression specifying section 204 may process the values ​​set on the setting screen 3501, as in the first embodiment.

[0269] 36 is a diagram showing an example of an image designation screen 3601 according to the third embodiment. Components denoted by the same reference numerals as those in the image designation screen 701 of FIG. 7 are the same as those in the first embodiment, and therefore will not be described. The image designation screen 3601 is the same as the image designation screen 701 of the first embodiment except for the radio button 706 for designating AI image generation, the prompt box 716, and the check box 719 for switching prompt change permission information on / off.

[0270] FIG. 37 is a diagram showing an example of a poster preview screen 3701 of the third embodiment. Components denoted by the same reference numerals as those in the poster preview screen 1001 of FIG. 10 are the same as those in the first embodiment, and therefore will not be described again. The poster preview screen 3701 of the third embodiment is provided with an information display area 3702. The information display area 3702 is an area that displays information related to image generation. In this embodiment, the prompt and random number used in generating the generated image arranged on the corresponding poster are displayed as information related to image generation.

[0271] <Processing flow> Fig. 38 is a flowchart showing the poster generation process executed by the poster creation application in the third embodiment. Of the processes in this flowchart, the same processes as those in the poster generation process of the first embodiment shown in Fig. 14(a) are given the same reference numerals, and duplicated explanations will be omitted. The following explanation will focus on the differences from the first embodiment.

[0272] 35 on the display 105. The user inputs each setting via the UI screen of the setting screen 3501 using the keyboard 106 or the pointing device 107.

[0273] In S1402, as in the first embodiment, the poster creation condition specifying unit 201 and the target impression specifying unit 204 acquire their corresponding settings from the setting screen 3501. That is, the poster creation condition specifying unit 201 acquires the poster size, number of posters to be created, and usage category designated by the user. The target impression specifying unit 204 acquires the target impression designated by the user.

[0274] Next, the process proceeds to S1404. In S1404, as in the first embodiment, the number of selections is determined so that posters can be generated according to the number of creations specified in the poster creation condition specification unit 201. That is, the skeleton selection unit 214 determines the number of skeletons to select, the color scheme selection unit 215 determines the number of color scheme patterns to select, and the font selection unit 216 determines the number of fonts to select. The method of determining the number of selections is the same as in the first embodiment.

[0275] Next, the process proceeds to S3802. In S3802, the text designation unit 202 and the image designation unit 3401 acquire their corresponding settings from the setting screen 3501. In addition, the image acquisition unit 211 acquires image data. Specifically, the image acquisition unit 211 reads the image files, in-application materials, and externally linked materials in the HDD 104 that are specified by the image designation unit 3401 into the RAM 103. In addition, the image designation unit 3401 acquires the prompts specified in the prompt boxes 3502 and 3504 on the setting screen 3501, and stores them in the RAM 108.

[0276] Thereafter, the process proceeds to S1407. In S1407, the skeleton acquisition unit 213 acquires a skeleton that meets the various setting conditions. The skeleton acquisition method is the same as in the first embodiment. S1408 to S1410 are the same as in the first embodiment. That is, the skeleton selection unit 214 selects, from the skeletons acquired in S1407, a skeleton that matches the desired impression specified in the desired impression designation unit 204. In S1409, the color scheme selection unit 215 selects a color scheme that matches the desired impression specified in the desired impression designation unit 204. In S1410, the font selection unit 216 selects a font combination that matches the desired impression specified in the desired impression designation unit 204.

[0277] Then, the process proceeds to S3803. In S3803, the image generation unit 3403 generates a random number as an initial value to be input to the image generation AI.

[0278] In S3804, the image generation unit 3403 inputs the prompt acquired in S3802 or the changed prompt acquired in S3810, and the random number generated in S3803 into the image generation AI to generate an image. As with the image generation process in the first embodiment, the image generation AI may use a known technique for generating an image from a prompt, and detailed description of the image generation AI will be omitted. For example, Stable Diffusion is used as the image generation AI. If multiple prompts are acquired, image generation is performed for each prompt to obtain multiple generated images. The image generation unit 3403 also counts the number of times an image has been generated and stores this count in RAM 103. If a new random number is generated in S3802 or if the prompt is changed in S3810, the image generation unit 3403 resets the count of the number of times the image has been generated and starts counting again. The image generation unit 3403 also associates the generated image with information about the prompt and random number used to generate the generated image.

[0279] In S3805, the image generation unit 3403 acquires the number of times the image recorded in RAM 103 has been generated, and determines whether the number of times the image has been generated is greater than a predetermined threshold (upper limit). If the acquired number of times the image has been generated is greater than the threshold (upper limit), the image generation unit 3403 transitions to S3806, and if the number of times the image has been generated does not exceed the predetermined threshold (upper limit), the image generation unit 3403 transitions to S3803. In this embodiment, the image generation unit 3403 determines whether the number of times the image has been generated is greater than five. Note that the threshold for the number of times the image has been generated (upper limit) can be any number equal to or greater than 1.

[0280] In S3806, the image analysis unit 212 performs analysis processing on the image data acquired in S3802 and the image data generated in S3804 to acquire information indicating feature amounts. The analysis processing is similar to the processing described in S1406 in the first embodiment, and therefore a description thereof will be omitted.

[0281] In S3807, the layout unit 217 generates a poster by setting character information, images, color schemes, and fonts for the skeleton selected by the skeleton selection unit 214. In the layout processing of the third embodiment, the layout unit 217 lists the combinations of images generated in S3804, combines them with the images acquired in S3802, and lists all combinations of skeletons, color patterns, fonts, and images. Note that if there are multiple prompts acquired in S3802, one image generated for each prompt is selected and combined.

[0282] For example, consider the case where generated images 11 and 12 are generated using different random numbers for prompt 1, and generated images 21 and 22 are generated using different random numbers for prompt 2. In this case, layout unit 217 obtains four image combinations: generated images 11 and 21, generated images 11 and 22, generated images 12 and 21, and generated images 12 and 22. If image 3 obtained in S3802 is also present, possible image combinations are enumerated by combining image 3 with each of the four combinations of generated images 11 to 22. Layout unit 217 combines the four image combinations obtained in this way with the skeleton, color scheme, and font selected in S1408 to S1410.

[0283] In this embodiment, the layout unit 217 lists all the combinations of generated images, but this is not limiting. For example, the layout unit 217 may display the generated images on the image selection screen 901 (FIG. 9) described in the first embodiment, and allow the user to specify the generated image to be used. In this case, the number of combinations of generated images can be narrowed down to one, thereby reducing the number of combinations of skeletons, color patterns, and fonts. The layout unit 217 executes layout processing for each combination in turn to generate poster data. The layout processing is the same as that shown in FIG. 21.

[0284] Next, the process proceeds to S1412. In S1412, the poster impression estimation unit 3404 executes rendering processing on each poster data acquired from the layout unit 217, and estimates the impression of the rendered poster image. The poster impression estimation unit 3404 associates the estimated impression of the poster with the poster data.

[0285] Next, the process proceeds to S3808. In S3808, the poster impression estimation unit 3404 calculates the difference (impression difference) and distance between the target impression acquired in S3801 and the impression of the generated poster estimated in S1412, and links them to the corresponding poster data. The calculated impression difference represents the amount of change required to bring the impression of the generated poster closer to the target impression. Furthermore, the smaller the distance value, the closer the target impression and the impression of the generated poster are.

[0286] In S3809, the poster impression estimation unit 3404 determines whether all distances calculated in S3808 are greater than a predetermined threshold. That is, if the distances between the estimated impressions and the target impressions of each of the multiple posters generated by the layout unit 217 are all greater than the predetermined threshold, the process proceeds to S3810; otherwise, the process proceeds to S3811. Note that if the number of prompt changes in S3810 exceeds a certain number and all distances calculated in S3808 are greater than the threshold, the poster creation application may proceed to a different process, such as a notification process for the user. In other words, this indicates that a poster suitable for the target impression could not be generated even after multiple prompt changes. Therefore, the poster creation application may display a warning screen on the display 105 indicating that it is difficult to change the prompt to one closer to the target impression and then abort the poster generation process. Alternatively, the poster creation application may retain and acquire the top N generated posters with the smallest distance values ​​among the distances calculated in S3808 up to this stage, and proceed to S3811.

[0287] In S3810, the prompt change unit 3402 changes the prompt acquired in S3802 to a prompt suitable for the target impression. The processing of the prompt change unit 3402 in S3810 is the same as the prompt change processing in S1506 in the first embodiment. The prompt change unit 3402 acquires an additional prompt based on the difference in impression calculated in S3808.

[0288] In S3811, the poster selection unit 219 selects a poster to be output to the display 105 (to be presented to the user) from the poster data acquired from the poster impression estimation unit 3404. In this embodiment, the poster selection unit 219 selects a poster for which the distance value is determined to be equal to or less than the threshold value in S3809. Other operations are the same as those described in S1513 of the first embodiment.

[0289] In S3812, the poster display unit 205 renders the poster data selected by the poster selection unit 219 and outputs the poster image to the display 105. That is, a poster preview screen 3701 shown in FIG.

[0290] As described above, the poster generation process of the third embodiment allows the prompt to be changed based on the impression of the generated poster. This allows the prompt to be changed with emphasis on the impression of the poster, which is the final product, making it easier to generate a poster that is closer to the intended design. Therefore, in a poster creation application that generates data for a product such as a poster, usability for obtaining the intended content to be placed on the poster can be improved.

[0291] Although preferred embodiments of the present disclosure have been described above with reference to the accompanying drawings, the present disclosure is not limited to such examples. In the above-described embodiments, the content generated by the generation AI is described as an image. However, the content is not limited to images and may be text information to be placed on a poster. For example, the present disclosure can be applied by using a generation AI capable of generating the title, subtitle, and text and catchphrases used as the main text of a poster. Furthermore, for example, the third embodiment illustrates an example of changing a prompt for generating an image based on the impression of the poster. However, as shown in the second embodiment, a prompt for style conversion of an image may be determined. Furthermore, a template may be selected and presented instead of generating a poster. Furthermore, the display screen, processing flow, layout method, etc. are merely examples. It is clear that a person skilled in the art can conceive of various modifications or alterations within the scope of the disclosed technical ideas, and these are naturally understood to fall within the technical scope of the present disclosure.

[0292] <<Other embodiments>> The above-described embodiments can also be realized by executing the following process. That is, software (programs) that realize the functions of the above-described embodiments are supplied to a system or device via a network or various storage media, and a computer (CPU, MPU, etc.) of the system or device reads and executes the program. The program may be executed by a single computer or by multiple computers in cooperation with each other. Furthermore, all of the above-described processes do not need to be implemented by software; some or all of the processes may be implemented by hardware such as an ASIC. Furthermore, the CPU is not limited to a single CPU that performs all processes; multiple CPUs may perform processes in cooperation with each other as appropriate. The functions of the above-described embodiments are not necessarily realized by a computer reading and executing program code. The above-described functions may also be realized by an operating system (OS) running on a computer that performs some or all of the actual processes based on the instructions of the program code.

[0293] The disclosure of the above-described embodiment includes the following configurations.

[0294] (Configuration 1) An information processing device that generates data of a production, a receiving means for receiving a user's designation of a target impression, which is an impression that the production is ultimately required to retain; a determination means for determining a prompt for a generation AI (Artificial Intelligence) to generate content to be placed in the production, an information processing device, characterized in that a first prompt determined by the determination means when the receiving means receives a first target impression is different from a second prompt determined by the determination means when the receiving means receives a second target impression different from the first target impression;

[0295] (Configuration 2) 2. The information processing apparatus according to configuration 1, wherein the determining means determines the first prompt or the second prompt based on an impression estimated from a prompt and the target impression.

[0296] (Configuration 3) 2. The information processing device according to claim 1, wherein the determining means determines the first prompt or the second prompt based on an impression estimated from content generated using a prompt and the target impression.

[0297] (Configuration 4) The information processing device described in configuration 1, characterized in that the determination means determines the first prompt or the second prompt based on the impression estimated from the production in which content generated using the prompt is placed and the target impression.

[0298] (Configuration 5) The accepting means further accepts a prompt specified by the user; 5. The information processing device according to claim 2, wherein the determining means determines the first prompt or the second prompt by changing a character string included in the prompt received by the receiving means.

[0299] (Configuration 6) The receiving means further receives a designation of the content by the user; further comprising an acquisition unit that acquires a prompt based on the content accepted by the acceptance unit; 5. The information processing device according to claim 2, wherein the determining means determines the first prompt or the second prompt by changing a character string included in the prompt acquired by the acquiring means.

[0300] (Configuration 7) The receiving means further receives a designation of the content by the user; 2. The information processing apparatus according to configuration 1, wherein the determining means determines a conversion prompt for converting the content accepted by the accepting means as the first prompt or the second prompt.

[0301] (Configuration 8) the content is an image, 8. The information processing device according to configuration 7, wherein the conversion prompt is a prompt for changing the style of the image.

[0302] (Configuration 9) An information processing device according to any one of configurations 1 to 8, further comprising a selection means for selecting a prompt for generating the content to be placed in the production from one or more of the first prompts or second prompts determined by the determination means.

[0303] (Configuration 10) The information processing device described in configuration 9, characterized in that the selection means displays a screen for the user to select a prompt for generating the content to be placed in the production from one or more of the first prompts or second prompts determined by the determination means, and accepts the selection by the user.

[0304] (Configuration 11) 11. The information processing device according to configuration 10, wherein the prompts accepted by the accepting means are further displayed on the screen in a selectable manner.

[0305] (Configuration 12) An information processing device according to any one of configurations 1 to 11, further comprising a selection means for selecting the content to be placed in the production from one or more contents generated by the generation AI using one or more of the first prompts or second prompts determined by the determination means.

[0306] (Configuration 13) The information processing device described in configuration 12, characterized in that the selection means displays a screen for the user to select the content to be placed in the production from one or more contents generated by the generation AI using one or more of the first prompts or second prompts determined by the determination means, and accepts the selection by the user.

[0307] (Configuration 14) An information processing device according to any one of configurations 1 to 13, further comprising a selection means for selecting data of one or more of the productions in which one or more contents generated by the generation AI are placed using one or more of the first prompts or the second prompts determined by the determination means.

[0308] (Configuration 15) The information processing device described in configuration 14, characterized in that the selection means displays a screen for the user to select one of the data of one or more of the productions in which one or more contents generated by the generation AI are placed using one or more of the first prompts or second prompts determined by the determination means, and accepts the selection by the user.

[0309] (Configuration 16) The information processing device described in configuration 13 or 15, characterized in that the screen further displays at least one of the first prompt or the second prompt used to generate the content and the initial value set in the generation AI.

[0310] (Configuration 17) 17. The information processing device according to any one of configurations 1 to 16, wherein the first prompt and the second prompt include different character strings.

[0311] (Configuration 18) The information processing device according to any one of configurations 1 to 16, characterized in that the prompt determined by the determination means includes a base prompt that is a prompt used as a base and one or more additional prompts to be added to the base prompt, and the first prompt and the second prompt are different from each other in the additional prompts.

[0312] (Configuration 19) 19. The information processing apparatus according to configuration 18, wherein the determining means determines the additional prompt based on a difference between the target impression and an impression estimated from the base prompt.

[0313] (Configuration 20) impression information, which is information in which a character string and a value indicating an impression of the character string are linked in advance, is stored in advance; 20. The information processing apparatus according to claim 19, wherein the determining means acquires a character string to be the additional prompt from the impression information.

[0314] (Configuration 21) 19. The information processing apparatus according to configuration 18, wherein when the accepting means accepts the user's designation of the prompt, the determining means treats the prompt accepted by the accepting means as the base prompt.

[0315] (Configuration 22) The information processing device according to any one of configurations 18 to 20, characterized in that when the accepting means accepts the user's designation of the content, the determining means acquires the base prompt based on an impression estimated from the content accepted by the accepting means.

[0316] (Configuration 23) When the accepting unit accepts a designation of an image as the content from the user, 21. The information processing device according to any one of configurations 18 to 20, wherein the determining means acquires the base prompt based on a caption generated from the image accepted by the accepting means.

[0317] (Configuration 24) 24. The information processing device according to any one of configurations 1 to 23, wherein the product is a poster.

[0318] (Configuration 25) An information processing device that generates data of a production, A receiving means for receiving a designation of a target impression, which is an impression that the creation is ultimately required to retain, and a designation of a prompt for a generation AI to generate content to be placed in the creation; a change instruction means for receiving an instruction from a user as to whether or not to permit the prompt change received by the receiving means; a presentation means for presenting a changed prompt or content generated by the changed prompt in a manner that is visible to the user when an instruction to permit the change of the prompt is received from the change instruction means; An information processing device comprising:

[0319] (Configuration 26) An information processing device that generates data of a production, a receiving means for receiving a designation of a target impression that is an impression that the production is ultimately required to maintain, and a designation of content to be placed in the production; conversion instruction means for receiving from a user an instruction to convert the content received by the receiving means; a presentation means for presenting the content received by the reception means and the converted content so as to be visible to the user when the conversion instruction means receives an instruction to convert the content; An information processing device comprising:

[0320] (Configuration 27) An information processing method for generating data of a production, comprising: a receiving step of receiving a user's designation of a target impression, which is an impression that the production is ultimately required to retain; A determination step of determining a prompt for the generation AI to generate content to be placed in the production, an information processing method characterized in that a first prompt determined in the determination step when a first target impression is received in the reception step is different from a second prompt determined in the determination step when a second target impression different from the first target impression is received in the reception step.

[0321] (Configuration 28) A program for causing a computer to execute an information processing method for generating data of a production, The information processing method includes: a receiving step of receiving a user's designation of a target impression, which is an impression that the production is ultimately required to retain; A determination step of determining a prompt for the generation AI to generate content to be placed in the production, A program in which a first prompt determined in the determination step when a first target impression is received in the reception step is different from a second prompt determined in the determination step when a second target impression different from the first target impression is received in the reception step.

[0322] (Configuration 29) An information processing method for generating data of a production, comprising: A receiving step of receiving a designation of a target impression, which is an impression that the creation is ultimately required to retain, and a designation of a prompt for a generation AI to generate content to be placed in the creation; a change instruction step of receiving, from the user, an instruction as to whether or not to permit the prompt change received in the receiving step; a presentation step of presenting the changed prompt or content generated by the changed prompt to the user when an instruction to permit the change of the prompt is received in the change instruction step; An information processing method comprising:

[0323] (Configuration 30) A program for causing a computer to execute an information processing method for generating data of a production, The information processing method includes: A receiving step of receiving a designation of a target impression, which is an impression that the creation is ultimately required to retain, and a designation of a prompt for a generation AI to generate content to be placed in the creation; a change instruction step of receiving, from the user, an instruction as to whether or not to permit the prompt change received in the receiving step; a presentation step of presenting the changed prompt or content generated by the changed prompt to the user when an instruction to permit the change of the prompt is received in the change instruction step; Including, the program.

[0324] (Configuration 31) An information processing method for generating data of a production, comprising: a receiving step of receiving a designation of a target impression that is an impression that the production is ultimately required to maintain, and a designation of content to be arranged in the production; a conversion instruction step of receiving, from a user, an instruction to convert the content received in the receiving step; a presentation step of presenting the content received in the reception step and post-conversion content to the user when an instruction to convert the content is received in the conversion instruction step; An information processing method comprising:

[0325] (Configuration 32) A program for causing a computer to execute an information processing method for generating data of a production, The information processing method includes: a receiving step of receiving a designation of a target impression that is an impression that the production is ultimately required to maintain, and a designation of content to be arranged in the production; a conversion instruction step of receiving, from a user, an instruction for converting the content received in the receiving step; a presentation step of presenting the content received in the reception step and the converted content to the user when an instruction to convert the content is received in the conversion instruction step; Programs including.

Claims

1. An information processing device that generates data of a production, a receiving means for receiving a user's designation of a target impression, which is an impression that the production is ultimately required to retain; a determination means for determining a prompt for a generation AI (Artificial Intelligence) to generate content to be placed in the production; an information processing device characterized in that a first prompt determined by the determination means when the receiving means receives a first target impression is different from a second prompt determined by the determination means when the receiving means receives a second target impression different from the first target impression;

2. 2. The information processing apparatus according to claim 1, wherein the determining means determines the first prompt or the second prompt based on an impression estimated from a prompt and the target impression.

3. The information processing apparatus according to claim 1 , wherein the determining means determines the first prompt or the second prompt based on an impression estimated from content generated using a prompt and the target impression.

4. The information processing device according to claim 1, characterized in that the determination means determines the first prompt or the second prompt based on an impression estimated from the production in which content generated using the prompt is placed and the target impression.

5. The accepting means further accepts a prompt specified by the user; 5. The information processing device according to claim 2, wherein the determining unit determines the first prompt or the second prompt by changing a character string included in the prompt received by the receiving unit.

6. The receiving means further receives a designation of the content by the user; further comprising an acquisition unit that acquires a prompt based on the content accepted by the acceptance unit; 5. The information processing apparatus according to claim 2, wherein the determining unit determines the first prompt or the second prompt by changing a character string included in the prompt acquired by the acquiring unit.

7. The receiving means further receives a designation of the content by the user; 2. The information processing apparatus according to claim 1, wherein the determining unit determines a conversion prompt for converting the content received by the receiving unit as the first prompt or the second prompt.

8. the content is an image, 8. The information processing apparatus according to claim 7, wherein the conversion prompt is a prompt for changing the style of the image.

9. 2. The information processing apparatus according to claim 1, further comprising a selection unit for selecting a prompt for generating the content to be placed in the production from one or more of the first prompts or the second prompts determined by the determination unit.

10. The information processing device according to claim 9, characterized in that the selection means displays a screen for the user to select a prompt for generating the content to be placed in the production from one or more of the first prompts or second prompts determined by the determination means, and accepts the selection by the user.

11. 11. The information processing apparatus according to claim 10, wherein the prompts accepted by the accepting means are further displayed on the screen in a selectable manner.

12. The information processing device described in claim 1, further comprising a selection means for selecting the content to be placed in the production from one or more contents generated by the generation AI using one or more of the first prompts or second prompts determined by the determination means.

13. The information processing device described in claim 12, characterized in that the selection means displays a screen for the user to select the content to be placed in the production from one or more contents generated by the generation AI using one or more of the first prompts or second prompts determined by the determination means, and accepts the selection by the user.

14. The information processing device described in claim 4, further comprising a selection means for selecting one or more pieces of production data in which one or more contents generated by the generation AI are placed using one or more of the first prompts or second prompts determined by the determination means.

15. The information processing device described in claim 14, characterized in that the selection means displays a screen for the user to select one of the one or more pieces of production data in which one or more contents generated by the generation AI are placed using one or more of the first prompts or second prompts determined by the determination means, and accepts the user's selection.

16. The information processing device according to claim 13 or 15, characterized in that the screen further displays at least one of the first prompt or the second prompt used to generate the content and an initial value set in the generation AI.

17. The information processing apparatus according to claim 1 , wherein the first prompt and the second prompt include different character strings.

18. 2. The information processing device according to claim 1, wherein the prompts determined by the determination means include a base prompt that is a prompt used as a base and one or more additional prompts that are added to the base prompt, and the first prompt and the second prompt have different additional prompts.

19. 19. The information processing apparatus according to claim 18, wherein the determining means determines the additional prompt based on a difference between the target impression and an impression estimated from the base prompt.

20. impression information, which is information in which a character string and a value indicating an impression of the character string are linked in advance, is stored in advance; 20. The information processing apparatus according to claim 19, wherein the determining means acquires a character string to be the additional prompt from the impression information.

21. 19. The information processing apparatus according to claim 18, wherein when the accepting unit accepts the prompt specified by the user, the determining unit treats the prompt accepted by the accepting unit as the base prompt.

22. 19. The information processing device according to claim 18, wherein when the accepting unit accepts the user's designation of the content, the determining unit acquires the base prompt based on an impression estimated from the content accepted by the accepting unit.

23. When the accepting unit accepts a designation of an image as the content from the user, 19. The information processing apparatus according to claim 18, wherein the determining means acquires the base prompt based on a caption generated from the image accepted by the accepting means.

24. 2. The information processing apparatus according to claim 1, wherein the work is a poster.

25. An information processing device that generates data of a production, A receiving means for receiving a designation of a target impression, which is an impression that the creation is ultimately required to retain, and a designation of a prompt for a generation AI to generate content to be placed in the creation; a change instruction means for receiving an instruction from a user as to whether or not to permit the prompt change received by the receiving means; a presentation means for presenting a changed prompt or content generated by the changed prompt in a manner that is visible to the user when an instruction to permit the change of the prompt is received from the change instruction means; An information processing device comprising:

26. An information processing device that generates data of a production, a receiving means for receiving a designation of a target impression that is an impression that the production is ultimately required to maintain, and a designation of content to be placed in the production; conversion instruction means for receiving from a user an instruction to convert the content received by the receiving means; a presentation means for presenting the content received by the reception means and the converted content so as to be visible to the user when the conversion instruction means receives an instruction to convert the content; An information processing device comprising:

27. An information processing method for generating data of a production, comprising: a receiving step of receiving a user's designation of a target impression, which is an impression that the production is ultimately required to retain; determining a prompt for a generation AI to generate content to be placed in the production; an information processing method characterized in that a first prompt determined in the determination step when a first target impression is received in the reception step is different from a second prompt determined in the determination step when a second target impression different from the first target impression is received in the reception step;

28. A program for causing a computer to execute an information processing method for generating data of a production, The information processing method includes: a receiving step of receiving a user's designation of a target impression, which is an impression that the production is ultimately required to retain; determining a prompt for a generation AI to generate content to be placed in the production; a first prompt determined in the determination step when a first target impression is received in the reception step, and a second prompt determined in the determination step when a second target impression different from the first target impression is received in the reception step, the first prompt being different from the second prompt determined in the determination step.

29. An information processing method for generating data of a production, comprising: A receiving step of receiving a designation of a target impression, which is an impression that the creation is ultimately required to retain, and a designation of a prompt for a generation AI to generate content to be placed in the creation; a change instruction step of receiving, from the user, an instruction as to whether or not to permit the prompt change received in the receiving step; a presentation step of presenting the changed prompt or content generated by the changed prompt to the user when an instruction to permit the change of the prompt is received in the change instruction step; An information processing method comprising:

30. A program for causing a computer to execute an information processing method for generating data of a production, The information processing method includes: A receiving step of receiving a designation of a target impression, which is an impression that the creation is ultimately required to retain, and a designation of a prompt for a generation AI to generate content to be placed in the creation; a change instruction step of receiving, from the user, an instruction as to whether or not to permit the prompt change received in the receiving step; a presentation step of presenting the changed prompt or content generated by the changed prompt to the user when an instruction to permit the change of the prompt is received in the change instruction step; Including, the program.

31. An information processing method for generating data of a production, comprising: a receiving step of receiving a designation of a target impression that is an impression that the production is ultimately required to maintain, and a designation of content to be arranged in the production; a conversion instruction step of receiving, from a user, an instruction to convert the content received in the receiving step; a presentation step of presenting the content received in the reception step and post-conversion content to the user when an instruction to convert the content is received in the conversion instruction step; An information processing method comprising:

32. A program for causing a computer to execute an information processing method for generating data of a production, The information processing method includes: a receiving step of receiving a designation of a target impression that is an impression that the production is ultimately required to maintain, and a designation of content to be arranged in the production; a conversion instruction step of receiving, from a user, an instruction for converting the content received in the receiving step; a presentation step of presenting the content received in the reception step and the converted content to the user when an instruction to convert the content is received in the conversion instruction step; Programs including.

Images