Information processing systems and programs

Abstract

This invention provides an information processing system and program that can assist in creating prompts to be input into an image generation AI system. [Solution] The information processing system 10 comprises a control means 20 and a storage means 21. The control means 20 comprises an instruction acquisition unit 30 that acquires an image generation instruction, a prompt generation unit 31 that generates a prompt to be input to the generation AI system 15, a prompt transmission unit 32 that transmits the prompt, an AI output information acquisition unit 33 that acquires AI output information from the generation AI system 15, and an information output unit 34 that presents the AI ​​output information. The prompt generation unit 31 generates a concretization instruction prompt that includes an instruction to supplement information so as to concretize the image generation instruction as an image generation prompt, transmits the concretization instruction prompt to the generation AI system 15 to acquire an image generation prompt, and the information output unit 34 presents the acquired image generation prompt to the user.

Description

Technical Field

[0001] The present invention relates to an information processing system and a program that complement a prompt to be sent to an image generation AI system and point out missing elements.

Background Art

[0002] In recent years, services that generate text in an interactive format with users have been provided by generation AI systems using large language models (LLMs). Also, image generation AI systems that generate images by inputting text that describes the image and conditions of the desired image are provided. In order to obtain a desired output in such a generation AI system, it is necessary for the user to create an appropriate prompt and input it into the generation AI system.

[0003] However, it is difficult for inexperienced users to create appropriate prompts. For this reason, systems that generate prompts for generation AI systems are also known. Examples of systems that generate prompts to be input to a generation AI system using an LLM include those cited in Patent Document 1.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Images generated by AI image generation systems exhibit significant variation and fluctuations depending on the input prompts. Therefore, when inputting prompts to an AI image generation system, it is beneficial to clearly specify information that might not be explicitly stated in human communication, as well as obvious prerequisites, to increase the likelihood of outputting the desired image. Inexperienced users may not know how to properly configure prompts for an AI image generation system, so appropriate support is desirable.

[0006] This invention has been made in view of the above-mentioned problems, and aims to provide an information processing system and program that can assist in creating prompts to be input to an image generation AI system. [Means for solving the problem]

[0007] To solve the aforementioned problems, the information processing system according to the present invention is an information processing system that generates an image generation prompt based on an image generation instruction input by a user to an image generation AI system, and comprises control means and storage means, wherein the control means comprises an instruction acquisition unit that acquires the image generation instruction, a prompt generation unit that generates a prompt to be input to a generation AI system using LLM, a prompt transmission unit that transmits the prompt to the generation AI system, an AI output information acquisition unit that acquires AI output information from the generation AI system, and an information output unit that presents the AI ​​output information to the user, wherein the prompt generation unit generates a concretization instruction prompt that includes an instruction to supplement information so as to concretize the image generation instruction as the image generation prompt, the prompt transmission unit transmits the concretization instruction prompt to the generation AI system, and the information output unit presents the image generation prompt to the user, in which information has been supplemented to the image generation instruction acquired by the AI ​​output information acquisition unit.

[0008] Furthermore, in the information processing system according to the present invention, the prompt generation unit generates the concretization instruction prompt so as to include an instruction to extract the elements missing from the image generation instruction, and the information output unit presents the user with information on the elements missing from the image generation instruction acquired by the AI ​​output information acquisition unit.

[0009] Furthermore, in the information processing system according to the present invention, the prompt generation unit presents the user with information on the missing elements of the image generation instruction, and if it obtains the missing elements from the user, it regenerates the concretization instruction prompt based on the information that adds the missing elements to the image generation instruction.

[0010] Furthermore, in the information processing system according to the present invention, the prompt generation unit includes, as an instruction to supplement the information, an instruction to specify a particular nationality when the image generation instruction includes an instruction for a person and the nationality of the person is not specified.

[0011] Furthermore, in the information processing system according to the present invention, the prompt generation unit includes an instruction to supplement background information based on the instructions included in the image generation instruction, as an instruction to supplement the information.

[0012] Furthermore, in the information processing system according to the present invention, the prompt generation unit includes an instruction to enhance visibility when the size of the generated image is small, as an instruction to supplement the information.

[0013] Furthermore, the information processing system according to the present invention includes, in the prompt generation unit, an instruction to exclude certain inappropriate elements from the concretization instruction prompt.

[0014] To solve the aforementioned problems, the present invention provides a program usable from a user terminal that causes a computer to execute an instruction acquisition process for acquiring the image generation instruction, a prompt generation process for generating a prompt to be input to a generation AI system using LLM, a prompt transmission process for sending the prompt to the generation AI system, an AI output information acquisition process for acquiring AI output information from the generation AI system, and an information output process for presenting the AI ​​output information to the user. The prompt generation process generates a concretization instruction prompt that includes an instruction to supplement information so as to concretize the image generation instruction as an image generation prompt, the prompt transmission process transmits the concretization instruction prompt to the generation AI system, and the information output process presents the image generation prompt, with information supplemented to the image generation instruction acquired in the AI ​​output information acquisition process, to the user. [Effects of the Invention]

[0015] According to the information processing system and program of the present invention, users can easily generate image generation prompts to obtain stable images with less fluctuation and variability in the generated images. [Brief explanation of the drawing]

[0016] [Figure 1] This diagram shows the configuration of the information processing system according to this embodiment. [Figure 2] A detailed diagram of the control system is shown. [Figure 3] This is a flowchart in an information processing system. [Modes for carrying out the invention]

[0017] Embodiments of the present invention will be described in detail with reference to the drawings. The information processing system 10 of this embodiment is a system that, when generating an image with an image generation AI system, complements the image generation instructions input by the user, extracts missing elements to create more specific image generation prompts, and assists in generating an image closer to the user's wishes. The information processing system 10 uses a generation AI system 15 that employs an LLM (Large-Scale Language Model) to complement the image generation instructions and extract missing elements. An LLM is a language model constructed using a large amount of text data and deep learning technology. As a result, the generation AI system 15 performs natural language processing and is capable of natural dialogue and text generation similar to that of a human. The generation AI system 15 can receive prompts from the information processing system 10 and transmit AI output information corresponding to them.

[0018] As shown in Figure 1, the information processing system 10 can connect to multiple user terminals 12 via a communication network 17. This allows multiple users to log in and use the information processing system 10. Each user terminal 12 has a display means 13, such as a monitor. The user terminals 12 are computers such as PCs, or mobile information terminals such as smartphones. However, the user terminals 12 are not limited to these.

[0019] The information processing system 10 can connect to the generation AI system 15 and the image generation AI system 16 via the communication network 17. The information processing system 10 can obtain AI output information by sending prompts to the generation AI system 15. The information processing system 10 can also obtain generated images by sending image generation prompts to the image generation AI system 16. The information processing system 10 may be able to connect to multiple generation AI systems 15 and image generation AI systems 16, or the information processing system 10 itself may include the generation AI system 15 and the image generation AI system 16.

[0020] In FIG. 1, the information processing system 10 is connected to a plurality of user terminals 12, a generation AI system 15, and an image generation AI system 16 via a single communication network 17. However, the information processing system 10 may be connected to the user terminal 12, the generation AI system 15, and the image generation AI system 16 via separate communication networks respectively. For example, the information processing system 10 can be connected to the user terminal 12 via an in-house LAN and to the generation AI system 15 and the image generation AI system 16 via the Internet.

[0021] The information processing system 10 can be composed of a computer such as a server. The information processing system 10 includes at least a control means 20, a storage means 21, and a communication means 22. The control means 20 can be composed of a processor such as a CPU. The storage means 21 can be configured using, for example, RAM, DRAM, HDD, SSD, or external cloud or distributed storage. The communication means 22 can be configured using, for example, a NIC or a wireless LAN connection device.

[0022] As shown in FIG. 2, the control means 20 includes an instruction acquisition unit 30 that transmits and receives information to and from the user terminal 12 to acquire an image generation instruction input by the user, a prompt generation unit 31 that generates a prompt to be input to the generation AI system 15, a prompt transmission unit 32 that transmits the prompt to the generation AI system 15 or the image generation AI system 16, an AI output information acquisition unit 33 that acquires AI output information from the generation AI system 15 or the image generation AI system 16, and an information output unit 34 that transmits the AI output information to the user terminal 12 and presents it to the user.

[0023] Next, the flow of a user using the generation AI system 15 using the information processing system 10 will be described. As shown in FIG. 3, first, the user inputs an image generation instruction on the user terminal 12 (S1). The image generation instruction is a description in language of the image that the user wants to generate. The image generation instruction is, for example, as follows. "Female image: A businesswoman in her late 30s, with a confident expression. He appears to be wearing a suit and holding a tablet. "It evokes the image of a team leader or manager."

[0024] The instruction acquisition unit 30 receives and acquires the image generation instruction entered by the user. Next, the prompt generation unit 31 generates a concretization instruction prompt to further concretize the acquired image generation instruction as an image generation prompt (S2).

[0025] The concretization prompt takes user-entered image generation instructions, concretizes abstract expressions as much as possible, verbalizes the requests that can be inferred from the image generation instructions, and, if the image generation instructions include a purpose, accurately understands that purpose and includes it in the image generation prompt. The concretization instruction prompt contains instructions that supplement information to concretize the image generation instruction as an image generation prompt. The information that should be supplemented in these instructions is typically information that is easily omitted when creating image generation instructions, such as things that are not usually communicated between people or assumptions that are taken for granted.

[0026] Instructions to supplement information specifically include instructions to specify a particular nationality, such as Japan, when an image generation instruction includes an instruction for a person but does not specify the person's nationality.

[0027] Furthermore, instructions to supplement information may include instructions to supplement background information based on other instructions included in the image generation instructions. In addition, instructions to supplement background information may include instructions to specify a simple background when there are no specific background instructions in the image generation instructions.

[0028] Furthermore, instructions to supplement information include instructions to improve visibility when the generated image size is small. For example, if the generated image size is smaller than a certain size, a simple background or simple composition is specified.

[0029] The specific instruction prompt includes instructions to exclude certain inappropriate elements. For example, if the generated image generation prompt contains elements such as sexual depictions, violence, discriminatory language, or copyright infringement, the prompt will instruct the system to output the result after excluding these elements.

[0030] The concretization prompt includes instructions to extract elements that are missing from the image generation instruction. These instructions list possible missing elements, such as purpose, image style, details of people, intended use of the image, mood or atmosphere of the image, and elements that should not be included in the image. If any of these are missing from the generated image generation prompt, the instructions include a description that asks the user to present up to one element, along with advice for improvement and multiple examples.

[0031] The concretization instruction prompt further includes instructions for outputting the information generated from the above instruction sentence in a specific format. The specified format involves outputting the Japanese image generation prompt, the English translation of the Japanese image generation prompt, the content and examples of any missing elements, and the content with any inappropriate elements excluded, separately.

[0032] When a concretization instruction prompt is generated, the prompt transmission unit 32 transmits the concretization instruction prompt to the generation AI system 15 (S3). The generation AI system 15 generates an image generation prompt based on the instructions in the concretization instruction prompt (S4). As mentioned above, the concretization instruction prompt includes not only instructions to supplement information and generate an image generation prompt, but also instructions to extract missing elements, generate improvement proposals, and exclude inappropriate elements, so the generation AI system 15 also generates information for these purposes. Furthermore, the generation AI system 15 is instructed by the concretization instruction prompt to output the image generation prompt in English, so the generation AI system 15 translates the image generation prompt into English (S5). The generated information is transmitted from the generation AI system 15 to the information processing system 10, and the AI ​​output information acquisition unit 33 acquires the AI ​​generated information (S6).

[0033] The AI ​​output information generated by the generation AI system 15 will look like this, for example: Image generation prompt (Japanese) • Depicts a Japanese businesswoman in her late 30s. With a confident expression, her gaze is directed towards the camera. She is wearing a dark gray suit and a white blouse. • Hold the tablet in your right hand and assume a natural pose. The surrounding scenery depicts a corner of an office, with blurred desks and laptops in the background. • The background is simple, while reflecting a modern office space. • Emphasize a bright, professional, and leadership-oriented demeanor. Image generation prompt (English) ·Depict a Japanese businesswoman in her late 30s ·Confidence exuded through her facial expression, looking at the camera ·Wearing a dark gray suit and a white blouse ·The surroundings portray a corner of an office; blurred desks and a laptop in the background ·Background reflects a modern office space while keeping it simple ·Emphasize a bright, professional, and leadership-evoking mood Advice regarding missing elements Specifying the style and touch in detail will increase the likelihood of generating something closer to your desired image. For example, consider styles such as "anime," "semi-realistic," or "pop art."

[0034] In this example, the image generation instructions do not explicitly state the woman's nationality, clothing details, or background. Therefore, the generation AI system 15 supplements this information based on the instructions in the concretization instruction prompt to generate the image generation prompt. Furthermore, the image generation instructions lack information about the overall style of the image. Therefore, the generation AI system 15 extracts these missing elements based on the instructions in the concretization instruction prompt and presents them along with an example. In addition, the Japanese image generation prompt has been translated into English.

[0035] The image generation instruction specifies the size of the image to be generated, and if this size is small, for example, 150px x 150px, the concretization instruction prompt includes instructions to specify a simple background and a simple composition if the size of the image to be generated is smaller than a certain size. Therefore, the generation AI system 15 includes a specification in the image generation prompt to make the background a simple one, such as a solid color or a gradient.

[0036] Next, the information output unit 34 of the information processing system 10 determines whether or not there is missing information in the acquired AI output information (S7). If there is no missing information, the information output unit 34 sends the generated image generation prompt to the user terminal 12 and displays it for the user (S8). If the user modifies the instructions in the image generation prompt (S9), the process from step S5 is repeated, sending the modified image generation prompt to the generation AI system 15 for translation into English. If the instructions are not modified in S9, the process proceeds to the next step.

[0037] If the AI ​​output information acquired in S7 contains information about missing elements, the information output unit 34 sends the information about the missing elements along with the generated image generation prompt to the user terminal 12 and displays it for the user (S10). If the user modifies the instructions in the image generation prompt (S11), the process from step S5 is repeated, sending the modified image generation prompt to the generation AI system 15 for translation into English. If the instructions are not modified in S11, and the user modifies the missing elements presented (S12), the process from step S2 is repeated, generating a concretization instruction prompt based on the information with the missing elements added to the initial image generation instruction. If no missing elements are added in S12, the process proceeds to the next step.

[0038] Next, the prompt transmission unit 32 sends an image generation prompt to the image generation AI system 16 (S13). The AI ​​output information acquisition unit 33 acquires the image generated by the image generation AI system 16 according to the image generation prompt, and the information output unit 34 sends the generated image to the user terminal 12 for display (S15). Alternatively, the information processing system 10 may only generate the image generation prompt, and the user may use the user terminal 12 to acquire the image generated by sending the image generation prompt to the image generation AI system 16.

[0039] In this way, the information processing system 10 assists in creating image generation prompts to be input to the image generation AI system 16 by generating concrete instruction prompts to be input to the generation AI system 15 using LLM based on the image generation instructions input by the user, supplementing the image generation instructions and further indicating any missing elements. This allows the user to easily generate image generation prompts that produce stable images with less fluctuation and variation in the generated image.

[0040] Furthermore, the program used in the information processing system 10 of this embodiment can also be provided independently. In this case, the program is a program that can be used from a user terminal 12 and causes the computer to execute an instruction acquisition process to acquire an image generation instruction, a prompt generation process to generate a prompt to be input to the generation AI system 15 using LLM, a prompt transmission process to send the prompt to the generation AI system 15, an AI output information acquisition process to acquire AI output information from the generation AI system 15, and an information output process to present the AI ​​output information to the user. The prompt generation process generates a concretization instruction prompt that includes an instruction to supplement information so as to concretize the image generation instruction as an image generation prompt, the prompt transmission process sends the concretization instruction prompt to the generation AI system 15, and the information output process presents the user with an image generation prompt in which information has been supplemented to the image generation instruction acquired in the AI ​​output information acquisition process.

[0041] Although embodiments of the present invention have been described above, the application of the present invention is not limited to these embodiments and can be applied in various ways within the scope of its technical concept. [Explanation of Symbols]

[0042] 10 Information Processing Systems 12 user terminals 13 Display means 15. Generative AI System 16 Image Generation AI System 17 Communication Networks 20 Control means 21 Memory means 22 Means of communication 30 Instruction acquisition part 31 Prompt generation unit 32 Prompt transmission unit 33 AI output information acquisition section 34 Information Output Unit

Claims

1. An information processing system that creates an image generation prompt based on an image generation instruction input by a user to an image generation AI system, Equipped with control means and storage means, The control means is An instruction acquisition unit that acquires the aforementioned image generation instruction, A prompt generation unit that generates prompts to be input to a generation AI system using LLM, A prompt transmission unit that transmits the prompt to the generation AI system, An AI output information acquisition unit that acquires AI output information from the aforementioned AI generation system, An information output unit that presents the AI ​​output information to the user, Equipped with, The prompt generation unit generates a concretization instruction prompt that includes an instruction to supplement information so as to concretize the image generation instruction as the image generation prompt. The prompt transmission unit transmits the concretization instruction prompt to the generation AI system. The information output unit is an information processing system that presents the user with the image generation prompt, in which information has been added to the image generation instruction acquired by the AI ​​output information acquisition unit.

2. The prompt generation unit generates the concretization instruction prompt so as to include an instruction to extract the elements missing from the image generation instruction, The information processing system according to claim 1, wherein the information output unit presents to the user information on elements missing from the image generation instruction acquired by the AI ​​output information acquisition unit.

3. The information processing system according to claim 2, wherein the prompt generation unit presents information to the user regarding the missing elements in the image generation instruction, and if it obtains the missing elements from the user, it regenerates the concretization instruction prompt based on the information that adds the missing elements to the image generation instruction.

4. The information processing system according to claim 1, wherein the prompt generation unit includes, as an instruction to supplement the information, an instruction to specify a particular nationality when the image generation instruction includes an instruction for a person and the nationality of the person is not specified.

5. The information processing system according to claim 1, wherein the prompt generation unit includes an instruction to supplement background information based on the instructions included in the image generation instruction, as an instruction to supplement the information.

6. The information system according to claim 1, wherein the prompt generation unit includes an instruction to improve visibility when the size of the generated image is small, as an instruction to supplement the information.

7. The information system according to claim 1, wherein the prompt generation unit includes an instruction to exclude certain inappropriate elements in the concretization instruction prompt.

8. A program that can be used from the user terminal, On the computer, The instruction acquisition process for acquiring the aforementioned image generation instruction, A prompt generation process that generates prompts to be input to a generation AI system using LLM, A prompt transmission process that transmits the aforementioned prompt to the generating AI system, An AI output information acquisition process that acquires AI output information from the aforementioned AI generation system, Information output processing that presents the AI ​​output information to the user, Make it run, In the prompt generation process, a concretization instruction prompt is generated that includes an instruction to supplement information so as to concretize the image generation instruction as an image generation prompt. In the prompt transmission process, the concretization instruction prompt is transmitted to the generation AI system. The information output process includes a program that presents the user with an image generation prompt in which information has been added to the image generation instruction obtained in the AI ​​output information acquisition process.

Images