Information processing device, information processing method, and program

Abstract

When you ask a generation AI to create decorative text images, sometimes the resulting decorations are so elaborate that the original text becomes illegible. [Solution] Based on a prompt containing an input string received from the user, the generated AI analyzes whether it can decipher the decorative character image it has created. Then, based on the analysis results, it outputs a decorative character image that can decipher the input string.

Description

Technical Field

[0001] The present disclosure relates to a technology for generating decorative characters using generative AI.

Background Art

[0002] In recent years, using generative AI, decorative characters preferred by users have been generated for any character specified by the user. Examples of decorative characters include ligatures, which combine multiple characters and represent them as one character, and swashes, which are design glyphs with extended lines at the beginning or end of a word. Although decorative characters may seem difficult to read due to the applied decoration, it is important that they can be correctly recognized as characters. Patent Document 1 discloses a technique for obtaining the character code of a character with a decoration such as a shaded character by performing OCR processing on the character when converting electronic documents generated by various application software into a predetermined format and centrally managing them.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Decorative lettering is often used in advertising and promotional materials such as posters and flyers to convey information such as store names and product / service concepts while creating a visual impact. Typically, when users consider decorative lettering to be used in posters, they check and select candidate lettering to see if it is suitable for their image (design). However, some decorative lettering generated by AI may be illegible to humans. If illegible decorative lettering is used in advertising and promotional materials, the original purpose of conveying information cannot be achieved. When generating candidate decorative lettering with AI, in addition to the above-mentioned check and selection process, users also need to judge whether the decorative lettering presented by the AI ​​is even legible, increasing the effort required of the user. [Means for solving the problem]

[0005] The information processing device relating to this disclosure is characterized by comprising: a receiving means for receiving the specification of an input string from a user; an acquisition means for acquiring a decorative character image generated by a generating AI based on a prompt including the specified input string, and the result of an analysis process determining whether or not the input string can be deciphered from the decorative character image; and an output means for outputting a decorative character image generated by the generating AI that is determined to be decipherable from the input string, based on the acquired result of the analysis process. [Effects of the Invention]

[0006] According to this disclosure, it is possible to reduce the effort required of users to judge the legibility of decorative characters presented by the generating AI. [Brief explanation of the drawing]

[0007] [Figure 1] A diagram showing an example of the overall configuration of a content creation system. [Figure 2] A diagram showing an example of the hardware configuration of an information processing device. [Figure 3]A flowchart showing the operation flow of the content editing server according to Embodiment 1. [Figure 4] (a) to (d) are diagrams showing examples of layout editing screens. [Figure 5] A diagram showing an example of the readability level setting screen. [Figure 6] (a) to (e) are diagrams showing examples of decorative lettering images. [Figure 7] A flowchart showing the operation flow of the content editing server according to Embodiment 2. [Figure 8] A diagram showing an example of the user permission settings screen. [Modes for carrying out the invention]

[0008] The embodiments described below will be explained with reference to the drawings. Note that the following embodiments are not limiting to the present invention, and not all combinations of features described in each embodiment are essential to the solution of the present invention. Furthermore, various other forms that do not depart from the spirit of the present invention are also included, and parts of the following embodiments can be combined as appropriate.

[0009] [Embodiment 1] <System Configuration> Figure 1 shows the overall configuration of the content creation system according to this embodiment, and an example of the functional configuration of the client terminal 10 and the servers 20-40 that provide cloud services, which are components of the system. The content creation system shown in Figure 1 consists of a client terminal 10, a content editing server 20, an image generation server 30, and a storage server 40. The content editing server 20 in this embodiment provides a cloud service that creates content by laying out objects such as text, images, and shapes in a design template (hereinafter simply referred to as "template"). Users can access the content editing server 20 from the client terminal 10 and create content such as posters, flyers, and photo albums by laying out text with decorative characters generated by the generation AI as objects in the template they have selected.

[0010] The client terminal 10 and each server 20-40 communicate with each other via a network. While the network is described using the internet as an example, it could also be a LAN such as an intranet, a WAN, a wireless communication network, a telephone line, a dedicated digital line, an ATM, etc. Alternatively, a communication network implemented using a combination of these could also be used. In the content creation system shown in Figure 1, there are three servers 20-40 corresponding to each service provided. However, a server configuration where, for example, the functions of the content editing server 20 are shared among multiple information processing devices is also possible. Alternatively, a single server could provide multiple services, such as the content editing server 20 also having the functions of the image generation server 30.

[0011] ≪Overview of Client Terminals≫ The client terminal 10 is an information processing device equipped with a web browser 111, such as a notebook PC, smartphone, or tablet device. The web browser 111 is software used to browse websites on the internet and to use content creation services provided by the content editing server 20. The web browser 111 receives client programs from the web application 121 for service provision of the content editing server 20 and enjoys content creation services by interpreting scripting languages ​​such as HTML and JavaScript. The web browser 111 can also print design content such as posters received from the web application 121 on a printing device (not shown) according to user instructions. The web browser 111 has a request processing unit 112 and a GUI control unit 113. The request processing unit 112 processes requests from the content editing server 20, for example, to generate design content or decorative characters to be used in said design content, and to receive the output generated in response to said requests. The GUI control unit 113 controls the GUI (Graphical User Interface) corresponding to the web application 121 on the display of the user interface 201. The user provides instructions to the web application 121 by using input devices such as a keyboard and mouse to input or select text while viewing various UI screens displayed on the screen.

[0012] ≪Overview of the Content Editing Server≫ The content editing server 20 is an information processing device that has a web application 121 for users to create design content such as posters, flyers, and photo albums. The web application 121 consists of a response processing unit 122, a settings management unit 123, a character image analysis unit 124, and a GUI control unit 125.

[0013] The response processing unit 122 receives user instructions from the client terminal 10 and performs response processing according to the received user instructions. For example, if the received instruction is a search instruction for design content templates such as posters or flyers, the response processing unit 122 accesses the storage server 40 and retrieves the template related to the search instruction from the templates held by the storage server 40. The user can specify the search conditions for templates by entering text or selecting template images on the template search screen generated by the GUI control unit 125 (described later) and displayed by the web browser 111. Also, for example, if the received instruction is a request to generate decorative characters, the response processing unit 122 sends a request to generate decorative character images to the image generation server 30, receives the decorative character images as output from the image generation server 30 and transfers them to the client terminal 10. The response processing unit 122 also sends content containing the desired decorative characters, created based on the user instructions from the client terminal 10, to the client terminal 10.

[0014] The data management unit 123 stores information about the layout when the user edits a desired template on a web browser, data of decorative character images generated by the generation AI, etc., in the storage server 40, linking them to the user.

[0015] The character image analysis unit 124 performs an analysis process on the decorative character image generated by the generative AI to determine whether the underlying character string can be deciphered. For this analysis process, for example, a learned model obtained by training pairs of various decorative character images and their correct data (corresponding character strings) as training data is used. Specifically, the learned model is input with the decorative character image to be processed and a prompt instructing an analysis of whether the decorative character image can be deciphered to the extent that it meets a predetermined deciphering level, and a result of whether deciphering is possible is obtained. Also, as an analysis process, for example, OCR (Optical Character Recognition) processing may be performed on the decorative character image to be processed, and a process of determining whether the recognized character string matches the character string underlying the decorative character image may be performed. At this time, OCR processing may be executed after applying known image processing techniques such as layout analysis and region extraction of the decorative character image. Furthermore, both the method using the learned model and the method using OCR processing may be implemented, and the logical sum (OR operation) or logical product (AND operation) of the analysis results of both may be taken. The character image analysis unit 124 executes such analysis processing using the resources within the content editing server 20, or requests a dedicated external server (not shown) to execute part or all of it.

[0016] The GUI control unit 125 generates various UI screens such as a home screen, a template search screen, a layout editing screen, and an error message screen according to the URL specified by the user via the web browser 111, and causes them to be displayed on the web browser 111 via the response processing unit 122.

[0017] ≪Overview of Image Generation Server≫ The image generation server 30 is an information processing device that provides an image generation service using generative AI, and is composed of a response processing unit 131 and a character image generation unit (generative AI) 132.

[0018] The response processing unit 131 receives a generation request from the content editing server 20 and performs response processing to transmit data of the decorative character image generated according to the received generation request to the content editing server 20. The generation AI 132 is a generation AI that generates decorative character images and the like using a prompt as an input. As the image generation AI, for example, "Copilot" can be used, but it is not limited to this, and other image generation AIs may also be used. The generation AI 132 is realized by a learned model (content generation model) obtained by performing machine learning on various data by means such as deep learning so that the target content can be obtained.

[0019] <<Overview of Storage Server>> The storage server 40 is an information processing device that provides storage and management services for various data, setting information, etc. used and generated by the web application 121, and is composed of a response processing unit 141 and a write / read unit 142.

[0020] The response processing unit 141 receives requests for data storage or data acquisition from the content editing server 20, the image generation server 30, or the client terminal 10, and performs response processing according to the received requests. The write / read unit 142 performs processing to write data to and hold it in a large-capacity storage medium such as an HDD (not shown) and read out the held data based on instructions from the response processing unit 141. Specifically, it performs writing or reading of various template images, template images after user editing, files indicating the layout of objects in the template images, decorative character images generated by the generation AI, information related to the user, etc.

[0021] <Hardware Configuration> Figure 2 shows an example of the hardware configuration of the information processing unit common to the client terminal 10 and each server 20-40. The user interface 201 is an operation display unit that performs input and output of information and signals, consisting of a display, keyboard, mouse, buttons, touch panel, etc. The network interface 202 connects to a network such as a LAN and communicates with other computers and network devices. The communication method can be either wired or wireless. The CPU 203 is a central processing unit that executes programs read from ROM 204, RAM 205, secondary storage device 206, etc. ROM 204 is a non-volatile memory in which embedded programs and data are recorded. RAM 205 is a volatile memory that provides a temporary memory area. The secondary storage device 206 is a large-capacity storage device such as an HDD or flash memory. Computers that do not have this hardware can be connected to and operated from other computers using remote desktop or remote shell. Each part is connected via input / output interface 207.

[0022] <Content editing server operation flow> Next, we will explain the process performed by the content editing server 20, specifically the process of presenting only legible decorative character images generated by the generation AI 132 to the user, which is a characteristic feature of this embodiment, with reference to the flowchart in Figure 3. The series of processes shown in the flowchart in Figure 3 are initiated when the user selects a desired template on the poster creation UI screen (not shown) displayed on the client terminal 10 by the web browser 111. In the following explanation, the symbol "S" represents a step.

[0023] In S301, the response processing unit 122 receives a request to generate decorative characters from the client terminal 10. Figure 4(a) shows an example of a layout editing screen, which is a UI screen for placing desired objects on a template selected by the user. As mentioned above, this layout editing screen is generated by the GUI control unit 125 of the Web application 121, and the Web browser 111 acquires it and displays it using the GUI control unit 113. The user can instruct the generation AI 132 to generate a decorative character image by entering the string they want to represent with decorative characters into the text box 404 in the left pane 401 of the layout editing screen 400 in Figure 4(a) and pressing the generate button 405. At the top of the left pane 401, there are tabs 403 corresponding to the type of target object to be laid out on the template specified by the user. The decorative character tab is for instructing the generation of the desired decorative character image, the materials tab is for searching and generating the desired materials, and the photos tab is for selecting and specifying the desired photos. By switching between these tabs, an operation area corresponding to the type of target object is displayed below the tab. The right pane 402 of the layout editing screen 400 displays the template 406 selected by the user in the template search screen (not shown). The user can create design content such as posters by placing objects that have been searched, generated, specified, etc., under the selected tab onto the template 406, for example, by dragging and dropping them with the mouse. When the decorative text tab is selected in the layout editing screen 400 in Figure 4(a), an empty text box 404 and a generate button 405 are displayed first. As described above, the user enters the string they want to express with decorative text into the text box 404 and presses the "Generate" button 405. Then, the request processing unit 112 of the web browser 111 sends a request to the content editing server 20 to generate a decorative text image based on the entered string (hereinafter referred to as the "input string"), along with information about the input string. In this step, the response processing unit 122 receives this generation request.Currently, the text box 404 in Figure 4(a) contains the string "Milk," and the following explanation will use the example of generating a decorative image of "MilK."

[0024] In S302, the response processing unit 122 determines the next process to be executed depending on whether the input string related to the generation request received in S301 has a real entity (i.e., whether the input string is not empty). If the input string is empty, S311 is executed next. On the other hand, if the input string is not empty, S303 is executed next.

[0025] In S303, the response processing unit 122 creates a prompt to instruct the generation AI 132 to generate a decorative character image of the input string related to the generation request received in S301. Here, a prompt is an instruction that specifies to the generation AI what kind of content should be generated. The prompt needs to include specific information so that the generation AI can correctly understand and process the instruction. In this embodiment, for example, a prompt is created that includes the input string itself, such as "Please create a ligature font for 'Milk'". Here, as a method for creating the prompt, for example, several prompt samples may be kept in the storage server 40 and created by combining them with the input string. Alternatively, a generation AI dedicated to creating prompts may be used. Furthermore, various logo images / photographs may be prepared in advance, and for example, if the user selects a specific logo image after entering a string, a prompt that takes into account the style of that logo image may be created. For example, if the base colors of the logo image selected by the user are red and white, a prompt such as "Please create a ligature font for 'Milk' with red and white as the base colors" may be created. Furthermore, the system may be configured to allow users to specify any color through palette selection, thereby creating prompts based on a particular color.

[0026] In S304, the response processing unit 122 sends a request to generate a decorative character image to the image generation server 30, along with the prompt created in S303. Upon receiving this generation request, the generation AI 132 in the image generation server 30 generates a decorative character image according to the prompt, and the response processing unit 122 receives the data of the generated decorative character image.

[0027] In S305, the response processing unit 122 determines the next process to be executed depending on whether it was able to receive the data for the decorative character image corresponding to the sent generation request. If the decorative character image is received, S306 is executed next. On the other hand, if the decorative character image is not received, S311 is executed next.

[0028] In S306, the next process to be executed is determined by whether the analysis process for all received decorative character images has been completed. If there are any character images that have not been analyzed, S307 is executed next. On the other hand, if the analysis process for all received decorative character images has been completed, S310 is executed next.

[0029] In S307, the character image analysis unit 124 analyzes the decorative character images of interest from the received images to determine whether the underlying input string can be deciphered, according to a pre-set decipherment level. Figure 5 is an example of a UI screen (decipherment level setting screen) for the user to set the decipherment level that the character image analysis unit 124 refers to in the analysis process. In area 501 of the decipherment level setting screen 500 shown in Figure 5, the user sets the criteria for determining what degree of match between the string extracted from each decorative character image and the underlying input string is considered decipherable. In this embodiment, the user selects either "exact match" or "partial match" using radio buttons. If "partial match" is selected, the user can further specify an arbitrary match rate as a percentage. The user selects and specifies the decipherment level considering factors such as the high probability of misrecognizing the lowercase letter "l" as the number "1" and the priority of design, and then presses the OK button 502 after completing these operations. In this embodiment, the difference between uppercase and lowercase letters of the alphabet does not affect the result (i.e., even if the input string is the alphabet letter "M" and the string extracted from the decorative character image subject to analysis (recognized string) is "m", it will be judged as a "match"). Now, in the readability level setting screen 500 of Figure 5, "partial match" is selected and "75%" is specified as the match rate. In this case, if the match rate between the string extracted from the decorative character image and the input string on which the decorative character image is based is 75% or higher, it will be judged as readable. Figures 6(a) to (e) show examples of decorative character images generated by the generation AI using the prompt "Please create a ligature font of "Milk"" as input. In the case of the decorative character images in Figures 6(a) to (c), it is generally difficult to extract each of the four letters of the alphabet "Milk". However, in the case of the decorative character image in Figure 6(d), it is possible to extract the string "MLK" consisting of three letters of the alphabet, which means that three of the four letters of "Milk" are readable.Furthermore, the decorative character image in Figure 6(e) can read the string "MIILK," which consists of five letters of the alphabet. This means that three of the four letters in "Milk" are legible. In other words, if the readability level is set to "partial match" and the match rate to "75%", then both the decorative character images in Figure 6(d) and (e) have three out of four characters matching, resulting in a match rate of 75%, and are judged to be legible. Note that in the above example, the satisfaction of the match rate is determined only by the presence or absence of each character that makes up the input string, but the match rate with the input string may also be determined by considering factors such as the number of recognized characters and their order.

[0030] In S308, the next process to be executed is determined based on the results of the analysis process described above for the string image of interest in S307. If the analysis process determines that the input string is legible, S309 is executed next. On the other hand, if the input string is determined to be illegible, the process returns to S306 and continues.

[0031] In S309, the character image analysis unit 124 adds the character string image of interest to a list for presenting the generated decorative character images to the user (hereinafter referred to as the "presentation list"). Specifically, the response processing unit 122 instructs the storage server 40 to save the data of the decorative character image of interest and the presentation list to which the decorative character image has been added. Upon receiving this instruction, the storage server 40 saves the data of the character string image of interest and the presentation list to which the character string image has been added by the write / read unit 142. The presentation list thus created holds information such as that shown in Table 1 below.

[0032] [Table 1]

[0033] In the case of Table 1 above, the "InputText" field stores the input string used to generate the decorative character image as its field value. The "Response" field stores information indicating the result of the analysis process. The "Path" field contains the path information within the storage server 40 where the decorative character image data is stored. In the example in Table 1 above, multiple values ​​can be entered for the "Path" field, but the value of "Path" could also be managed using a separate data structure, such as a key-value type. When a string image of interest is added to the presentation list, the process returns to S306 and continues.

[0034] Once the analysis process for all received decorative character images is complete, in S310, the response processing unit 122 sends a response containing a suggestion list. That is, the suggestion list obtained through the processing up to this point is sent to the client terminal 10, which was the source of the request to generate decorative character images, and instructs it to display the images included in the suggestion list on the layout editing screen. Upon receiving this instruction and the suggestion list, the GUI control unit 113 in the client terminal 10 retrieves the data of the decorative character images specified in the suggestion list by accessing the storage server 40, etc., and displays them on the layout editing screen.

[0035] In S311, error handling is performed. Specifically, for example, the GUI control unit 125 creates an error message screen prompting the user to input a string or retry, and the response processing unit 122 sends this to the client terminal 10 as part of the error handling. In response, the GUI control unit 113 on the client terminal 10 performs display processing, such as displaying the received error message screen as a pop-up. The above describes the operation flow of the content editing server 20 according to this embodiment.

[0036] <Verification of decorative character images on the client terminal> Here, referring to the diagram, we will explain, using the aforementioned specific example, how the decorative character images based on the presentation list are presented to the user on the Web browser 111 of the client terminal 10.

[0037] First, for comparison, let's refer to Figure 4(b) to explain the case of the conventional technology. In the layout editing screen 400' shown in Figure 4(b), all the string images generated by the generation AI 132 based on the input string "Milk" (in this case, the five decorative character images shown in Figures 6(a) to (e) above) are displayed in the left pane 401. When a user sees such a layout editing screen, before or simultaneously searching for a decorative character image that suits their preferences, they need to determine whether each of the displayed decorative character images is legible. This is a time-consuming task for the user, and the more decorative character images displayed, the more time it takes to make a judgment.

[0038] Next, the method according to this embodiment will be explained with reference to Figure 4(c). In the layout editing screen 400" shown in Figure 4(c), only those string images listed in the presentation list from all the string images generated by the generation AI 132 based on the input string "Milk" are displayed in the left pane 401. That is, of the five decorative string images shown in Figures 6(a) to (e) above, only the decorative string image 421 corresponding to Figure 6(d) and the decorative string image 422 corresponding to Figure 6(e), which are determined to be legible, are presented to the user. In the layout editing screen 400" shown in Figure 4(c), the user does not need to pay attention to legibility compared to the layout editing screen 400' shown in Figure 4(b), so they can concentrate on selecting a decorative string image that suits their preference. Figure 4(d) shows the state in the layout editing screen 400" shown in Figure 4(c) where the user has selected the decorative string image 421 and placed it on the template 406 in the right pane 402. In this way, the user can create the desired design content by performing editing operations such as placing other objects.

[0039] <Variation> In the above embodiment, only the decorative character images generated by the generation AI that met a pre-set readability level were stored in the storage server 40, but this is not limited to this. Generally, it is said that images generated by the generation AI are not reproducible (it is difficult to generate an image that is exactly the same as the first generated image as the second generated image). Therefore, images that do not meet the readability level may also be stored in the storage server 40. For example, if the user cannot find a decorative character image that satisfies them from the presented images, an analysis process with a lower readability level is performed on all the saved decorative character images based on the user's request. Then, the system may be configured to present images that meet the lowered readability level. Alternatively, only the data of decorative character images laid out on a template by the user may be stored.

[0040] Furthermore, in the above-described embodiment, the decorative character images generated by the generation AI that meet the readability level were listed and presented to the user. However, for example, a minimum number of decorative character images that should be included in the list may be defined. If the number of decorative character images that meet the readability level falls below a certain number, the process of returning to S304 and repeatedly having the generation AI generate decorative character images may be performed to ensure that a certain number or more of decorative character images can be presented to the user.

[0041] Furthermore, in the above-described embodiment, a presentation list was created that could identify decorative character images that met the readability level among those generated by the generation AI, and this list was presented to the user. The data of the listed decorative character images was stored in the storage server 40. However, the system is not limited to this. For example, a presentation list containing data of decorative character images that have been determined to meet the readability level may be created and sent to the client terminal 10. In this case, each time a character is determined to be readable in S308, the presentation list containing the data of the decorative character image of interest may be updated in S309, and the presentation list may be stored in the RAM 205 or secondary storage device 206 of the content editing server 20.

[0042] Furthermore, although the above-described embodiment describes a method of generating decorative character images using a generation AI via a web application on a server that provides content editing services, the embodiment is not limited to this. For example, the generation of decorative character images using a generation AI may be performed via an application installed on the client terminal 10.

[0043] As described above, according to this embodiment, only legible decorative character images generated by the generation AI are presented to the user. Therefore, the user does not need to judge whether the presented decorative character images are legible or not, and can concentrate on checking and selecting which decorative character images are suitable for the design content such as posters that they are creating.

[0044] [Embodiment 2] In the case of typical AI generators, there is often a limit on the number of times they can be used within a certain period, which is tied to the user. In that case, a user who has reached the limit on the number of times they can use the AI ​​generator will not be able to have the AI ​​generate the desired decorative text images to create design content until the next certain period. Embodiment 2 describes a method that allows users to use decorative text images generated by other users when they are unable to receive the service of decorative text images provided by the AI ​​generator, for example, because they have exceeded the limit on the number of times they can use the AI ​​generator.

[0045] <Content editing server operation flow> Since the system configuration and other aspects are the same as in Embodiment 1, we will omit the explanation and focus on the differences, specifically the operation flow of the content editing server. Figure 7 is a flowchart showing the operation flow of the content editing server 20 in this embodiment, which corresponds to the flowchart in Figure 3 of Embodiment 1. When performing the series of processes shown in this flowchart, for example, each user may be prompted to indicate whether or not they allow other users to freely use the decorative character images they have generated. Specifically, each user who has the generation AI generate decorative character images via the Web browser 111 is asked to set whether or not to allow other users to freely use them via a predetermined UI screen. Figure 8 is an example of a UI screen (usage permission setting screen) for a user to set whether or not to allow other users to freely use the decorative character images they have generated. In area 801 of the usage permission setting screen 800 shown in Figure 8, the user selects whether or not to allow other users to freely use the decorative character images they have generated using radio buttons from "Do not allow at all", "Allow at all", and "Set individually". If "Set individually" is selected, the user presses the reference button 802 to specify whether to allow or not allow each individual decorative character image. Then, when the user presses the OK button 803, attribute information corresponding to the selected radio button is assigned to each decorative character image generated by the AI ​​based on the user's request, and the data of the decorative character image is saved to the storage server 40. Table 2 below shows an example in which a flag indicating whether or not it can be freely used (flag value "1" = usable, "0" = not allowed) is assigned as attribute information to each decorative character image.

[0046] [Table 2]

[0047] In the example shown in Table 2 above, other users can freely use the decorative text image "001.jpg," but they cannot freely use the decorative text image "002.jpg."

[0048] The following describes the operation flow of the content editing server 20 according to this embodiment, following the flowchart in Figure 7. Similar to the flowchart in Figure 3, the series of processes shown in Figure 7 are initiated when the user selects a desired template on a UI screen (not shown) for poster creation, displayed on the client terminal 10 via the web browser 111. In the following description, the symbol "S" represents a step.

[0049] S701 to S704 correspond to S301 to S304 in the flow of Figure 3 of Embodiment 1, respectively, and there is no particular difference, so the explanation is omitted. In S705, if the response processing unit 122 is able to receive the decorative character image data corresponding to the sent generation request, S706 is executed next; otherwise, S711 is executed next. S706 to S710 correspond to S306 to S310 in the flow of Figure 3 of Embodiment 1, respectively, and there is no particular difference, so the explanation is omitted.

[0050] In S711, the response processing unit 122 determines the next process to be executed depending on whether the reason for not being able to receive the decorative character image is that the number of times the currently active user has used the generation AI has exceeded the upper limit. If the decorative character image could not be received because the number of times the generation AI has been used has reached the upper limit and no decorative character image was generated, then S712 is executed next. On the other hand, if the reason for not being able to receive the decorative character image is something other than reaching the upper limit, then S715 is executed next.

[0051] In S712, the response processing unit 122 accesses the storage server 40 and sends a request to acquire decorative character images (hereinafter referred to as "free images") that are permitted for free use by other users. Upon receiving this acquisition request, the response processing unit 141 of the storage server 40 refers to the input string, which is the value of "InputText" in Table 2 mentioned above, and attribute information (license flag) to search for decorative character image data whose degree of matching with the input string satisfies the aforementioned readability level. Then, the response processing unit 141 of the storage server 40 acquires the found decorative character image data via the write / read unit 142 and returns it to the content editing server 20.

[0052] In S713, the response processing unit 122 determines the next process to be executed depending on whether it was able to retrieve free image data from the storage server 40. If one or more free images are retrieved, S714 is executed next. On the other hand, if one or more free images are not retrieved, S715 is executed next.

[0053] In S714, the response processing unit 122 adds one or more acquired free images to the aforementioned suggestion list. That is, it instructs the storage server 40 to save the data of one or more free images and the suggestion list to which those free images have been added. At this time, information such as a flag indicating that an image is a free image may be added to the suggestion list. After the execution of S714, a response of the suggestion list is made in S710.

[0054] In S715, error handling is performed. Specifically, for example, the GUI control unit 125 creates a message screen prompting the user to input a string or retry, or a message screen informing the user that no free images are available, and the response processing unit 122 sends this to the client terminal 10. In response, the GUI control unit 113 on the client terminal 10 performs display processing, such as displaying the received message screen as a pop-up.

[0055] The above describes the operation flow of the content editing server 20 according to this embodiment. In this way, even if the currently using user is unable to generate decorative character images with the generation AI for some reason, in S710, when S711 to S714 are passed, the response processing unit 122 can respond with a presentation list containing free images.

[0056] <Variation> If the decorative character image cannot be received because the number of uses of the generation AI has reached its limit (Yes in S711), the system may ask the user whether they wish to use a free image before proceeding to S712. Specifically, the system may confirm the user's intention by displaying a message screen on the web browser 111 such as, "The number of uses of the generation AI has reached its limit, so a decorative character image cannot be generated. Would you like to use a decorative character image generated by another user in the past?" The system may then proceed to execute S712 only if the user's intention to use a free image is confirmed. If the user responds that they do not wish to use a free image, the system may display an error message screen such as, "Please wait a while and try again" in S715.

[0057] <Other Embodiments> The present invention can also be realized by supplying a program that implements one or more of the functions of the above-described embodiments to a system or device via a network or storage medium, and by having one or more processors in the computer of that system or device read and execute the program. It can also be realized by a circuit (e.g., an ASIC) that implements one or more functions.

[0058] Furthermore, this disclosure includes the following configurations and methods.

[0059] [Configuration 1] A means of receiving input strings from the user, An acquisition means for acquiring a decorative character image generated by a generation AI based on a prompt containing the specified input string, and the result of an analysis process to determine whether or not the input string can be deciphered from the decorative character image. Based on the results of the analysis process obtained, an output means outputs a decorative character image generated by the generation AI that is determined to be able to read the input string, An information processing device characterized by having the following features.

[0060] [Configuration 2] The information processing apparatus according to Configuration 1, characterized in that the output means does not output any decorative character images generated by the generating AI that are determined to be illegible to the input string, based on the results of the analysis process obtained.

[0061] [Configuration 3] The acquisition means is, The decorative character image generated by the generation AI is acquired, The analysis process is performed on the acquired decorative character image, and the results of the analysis process are obtained. An information processing apparatus according to configuration 1 or 2, characterized by the above.

[0062] [Structure 4] The receiving means receives the specification of the input string and the specification of the readability level of the input string in the analysis process, The acquisition means performs the analysis process based on the specified reading level. The information processing apparatus according to configuration 3, characterized by the features described herein.

[0063] [Composition 5] The information processing device according to configuration 4, characterized in that the recognition level is the degree of match between the string extracted from the decorative character image generated by the generation AI and the input string.

[0064] [Composition 6] The information processing device according to configuration 5, characterized in that the degree of matching includes an exact match between the extracted string and the input string, and a partial match between the extracted string and the input string.

[0065] [Composition 7] The information processing apparatus according to configuration 6, characterized in that the receiving means further accepts the specification of the matching rate when the reading level is specified as partial match.

[0066] [Structure 8] The information processing apparatus according to any one of configurations 4 to 7, characterized in that the analysis process is a process of determining whether the recognition level is met using a trained model obtained by training a pair of decorative character images and their correct data as training data.

[0067] [Composition 9] The information processing apparatus according to any one of configurations 4 to 7, characterized in that the analysis process is a process of determining whether the readability level is met by comparing the recognized string obtained by performing OCR (Optical Character Recognition) processing on the decorative character image generated by the generating AI with the input string on which the decorative character image is based.

[0068] [Configuration 10] The information processing apparatus according to any one of configurations 1 to 9, characterized in that the output means creates a list of decorative character images generated by the generation AI that are determined to be able to read the input string, and outputs the list to an external device used by the user to specify the input string.

[0069] [Composition 11] The output means outputs the data of the decorative character image generated by the generation AI to an external device that stores the data. The aforementioned list includes information indicating the storage location of the data for the decorative character images listed in the aforementioned list. The information processing apparatus according to configuration 10, characterized by the above.

[0070] [Composition 12] The information processing apparatus according to configuration 11, characterized in that the output means outputs data of the decorative character image generated by the generation AI, which is determined to be able to read the input string, to an external device that stores data.

[0071] [Composition 13] The acquisition means is, Create the prompt containing the specified input string, By having an external device that provides an image generation service using the aforementioned generation AI generate a decorative character image based on the created prompt and receiving the resulting output, a decorative character image with the specified input string decorated is obtained. An information processing device according to any one of configurations 1 to 12, characterized by the above.

[0072] [Composition 14] The external device that stores data stores past data of decorative character images generated by the generation AI. If the acquisition means is unable to acquire the decorative character image generated by the generation AI using the prompt containing the specified input string as input, it acquires data of decorative character images from the past data in which the input string can be deciphered, and creates the list. The information processing apparatus according to configuration 11, characterized by the features described above.

[0073] [Composition 15] The information processing device according to configuration 14, characterized in that if the acquisition means is unable to acquire the decorative character image generated by the generation AI using the prompt containing the specified input string as input, it is because the number of times the user has used the generation AI has exceeded the upper limit.

[0074] [Composition 16] The information processing device according to configuration 14 or 15, characterized in that the aforementioned past data is data of decorative character images generated by the generating AI based on the instructions of other users, which are permitted to be freely used by other users.

[0075] [Method 1] A reception step that accepts input strings from the user, An acquisition step of obtaining a decorative character image generated by the generation AI based on a prompt containing the specified input string, and the result of an analysis process to determine whether or not the input string can be deciphered from the decorative character image, Based on the results of the analysis process obtained, an output step is to output a decorative character image generated by the generation AI that is determined to be able to read the input string, An information processing method characterized by including

[0076] [Composition 17] A program for causing a computer to function as an information processing device as described in any one of the configurations 1 to 16.

Claims

1. A means of receiving input strings from the user, An acquisition means for acquiring a decorative character image generated by a generating AI based on a prompt containing the specified input string, and the result of an analysis process to determine whether or not the input string can be deciphered from the decorative character image. Based on the results of the analysis process obtained, an output means outputs a decorative character image generated by the generating AI that is determined to be able to read the input string, An information processing device characterized by having the following features.

2. The information processing apparatus according to claim 1, characterized in that the output means does not output any decorative character images generated by the generating AI that are determined to be illegible to the input string, based on the results of the analysis process obtained.

3. The acquisition means is, The decorative character image generated by the generation AI is acquired, The analysis process is performed on the acquired decorative character image, and the results of the analysis process are obtained. The information processing apparatus according to feature 1.

4. The receiving means receives the specification of the input string and the specification of the readability level of the input string in the analysis process, The acquisition means performs the analysis process based on the specified reading level. The information processing apparatus according to claim 3.

5. The information processing apparatus according to claim 4, characterized in that the recognition level is the degree of agreement between the string extracted from the decorative character image generated by the generating AI and the input string.

6. The information processing apparatus according to claim 5, characterized in that the degree of matching includes an exact match between the extracted string and the input string, and a partial match between the extracted string and the input string.

7. The information processing apparatus according to claim 6, characterized in that the receiving means further accepts the specification of the matching rate when the reading level is specified as the partial match.

8. The information processing apparatus according to claim 4, characterized in that the analysis process is a process of determining whether the recognition level is met using a trained model obtained by training a pair of decorative character images and their correct data as training data.

9. The information processing apparatus according to claim 4, characterized in that the analysis process is a process of determining whether the recognition level is met by comparing the recognized string obtained by performing OCR (Optical Character Recognition) processing on the decorative character image generated by the generating AI with the input string that was the basis for the decorative character image.

10. The information processing apparatus according to claim 1, wherein the output means creates a list of decorative character images generated by the generation AI that are determined to be able to read the input string, and outputs the list to an external device used by the user to specify the input string.

11. The output means outputs the data of the decorative character image generated by the generation AI to an external device that stores the data. The aforementioned list includes information indicating the storage location of the data for the decorative character images listed in the aforementioned list. The information processing apparatus according to feature 10.

12. The information processing apparatus according to claim 11, characterized in that the output means outputs data of the decorative character image generated by the generation AI, which is determined to be able to read the input string, to an external device that stores data.

13. The acquisition means is, Create the prompt containing the specified input string, By having an external device that provides an image generation service using the aforementioned generation AI generate a decorative character image based on the created prompt and receiving the resulting output, a decorative character image with the specified input string decorated is obtained. The information processing apparatus according to feature 1.

14. The external device that stores data stores past data of decorative character images generated by the generation AI. If the acquisition means is unable to acquire the decorative character image generated by the generation AI using the prompt containing the specified input string as input, it acquires data of decorative character images from past data in which the input string can be deciphered, and creates the list. The information processing apparatus according to feature 11.

15. The information processing apparatus according to claim 14, wherein if the acquisition means is unable to acquire the decorative character image generated by the generation AI using the prompt containing the specified input string as input, it is because the number of times the user has used the generation AI has exceeded the upper limit.

16. The information processing apparatus according to claim 14, characterized in that the aforementioned past data is data of the decorative character image generated by the generating AI based on the instructions of other users, which is permitted to be freely used by other users.

17. A reception step that accepts input strings from the user, An acquisition step of obtaining a decorative character image generated by the generating AI based on a prompt containing the specified input string, and the result of an analysis process to determine whether or not the input string can be deciphered from the decorative character image, Based on the results of the analysis process obtained, an output step is to output a decorative character image generated by the generating AI that is determined to be able to read the input string, An information processing method characterized by including

18. A program for causing a computer to execute the information processing method described in claim 17.

Images