POP production support system using AI-powered automatic layout and text coordinate synthesis.

Abstract

We provide a product listing image generation device that enables on-site, field-led promotional activities without requiring specialized skills. [Solution] A listing image generation device comprising an imaging device that captures images of products and generates image data, a storage device that stores image data, various template data and a program, display devices 4 and 6 that display information, and a processing unit that executes a program, wherein the processing unit performs background removal processing to automatically extract the outline of the product portion from the overall image data of the product acquired by the imaging device and erase the background area, performs layout determination processing to automatically resize and arrange the partial image data of the product that has undergone background removal processing on template data consisting of multiple product image layout frames for specific auction sites or specific e-commerce sites read from the storage device, and performs control to assign a unique identification number to the generated output image data and physically link it with management data in the storage device for batch storage.

Description

Technical Field

[0001] The present invention relates to a system for generating a POP advertisement that utilizes generative AI to automatically perform background generation, product image clipping, and optimal text layout (layout), and combines accurate information required at the site such as a store.

Background Art

[0002] In conventional image generation AI, Japanese typos and "strings that do not make sense" are often drawn, which is not suitable for commercial POPs. In addition, product photos taken by ordinary people with smartphones etc. have a cluttered background, and specialized image editing skills are required to seamlessly combine them with a high-quality design background.

Summary of the Invention

Problems to be Solved by the Invention

[0003] In conventional image generation AI, since the background and text are generated as an integrated image, Japanese typos and strings that do not make sense are often drawn, which is not suitable for commercial POPs that require accuracy. In addition, simple product photos taken with smartphones etc. have a cluttered background, and specialized image editing skills are required to seamlessly combine them with a high-quality design background. The present invention aims to completely eliminate the risk of typos by generative AI and enable users without specialized knowledge to immediately create high-quality POPs by utilizing product photos.

Means for Solving the Problems

[0004] The POP production support system according to the present invention includes an input means (1) for receiving an advertising concept, a plurality of text information, and a subject image from a user, and a background clipping means (12) for removing the background from the subject. Further, a background image generation means (2) for generating a background image that does not include text information using generative AI, and a layout determination means (14) for calculating the optimal arrangement coordinates of the subject and text via generative AI based on the composition of the background and the priority of the text are provided.

[0005] This system assigns the generating AI only non-routine creative processing such as "background rendering" and "suggestion of placement coordinates," and in the final image export, the text superimposing and compositing means (3) directly superimposes physical font data onto the image based on the calculated optimal coordinates. This physically prevents the occurrence of typographical errors, which are a drawback of generating AI.

[0006] Furthermore, the system includes modification means (5, 13) that provide instructions for regeneration of the displayed image using natural language, enabling interactive reconstruction of only the background and layout while maintaining specific information. [Effects of the Invention]

[0007] (1) Because the text is synthesized as physical data, it is possible to create accurate promotional materials that completely eliminate typos and inconsistencies inherent in AI generation. (2) By extracting and compositing only the subject from simple photos taken in the store, it becomes possible to easily create POP displays with professional-level appeal regardless of the shooting environment. (3) Timely promotional activities tailored to inventory status and campaigns can be carried out by the field staff without the need for specialized editing skills. [Brief explanation of the drawing]

[0008] [Figure 1] This block diagram shows the overall flow of this POP production support system. [Figure 2] This figure shows an example of the display screen of this POP production support system. [Explanation of symbols]

[0009] 1. Input means 2. Background image generation means 3 Text polymerization synthesis means 4 Display means 5 Corrective measures 6. Display section for generated images 7. Theme input field 8. Main text input field 9. Second candidate input field 10 Other text input fields 11 Image registration field 12. Image cropping input field (background cropping method) 13 Correction Instruction Input Field 14. Layout determination means

Claims

[Claim 1] A POP production support system characterized by comprising: an input means (1) that receives from a user an advertisement concept instruction, multiple pieces of text information to be displayed in the advertisement, and a subject image to be superimposed; a background removal means (12) that removes the background area from the subject image and extracts only the subject; a background image generation means (2) that generates background image data without text information using a generation AI based on the concept instruction; a layout determination means (14) that calculates the optimal coordinates for placing the subject and the text information using a generation AI based on the composition of the background image data and the priority of the text information; a text superimposition and synthesis means (3) that superimposes and synthesizes the subject and the text information on the background image data as physical font data without the generation AI based on the calculated optimal coordinates; a display means (4, 6) that displays the synthesized image; and a modification means (5, 13) that instructs the layout determination means or the background image generation means to regenerate the displayed image.

Images