Systems and methods for image compositing

The image generation system uses descriptive embeddings to enhance the realism and integration of objects into backgrounds, addressing the limitations of conventional methods by preserving object identity and improving geometric consistency.

AU2024203033B2Pending Publication Date: 2026-07-09ADOBE INC
View PDF 3 Cites 0 Cited by

Patent Information

Authority / Receiving Office
AU · AU
Patent Type
Applications
Current Assignee / Owner
ADOBE INC
Filing Date
2024-05-08
Publication Date
2026-07-09

AI Technical Summary

Technical Problem

Conventional image generation systems struggle to realistically composite objects from one image into another, particularly with non-rigid objects, as text prompts are insufficient for capturing detailed object information, and existing diffusion models lack versatility in handling multiple image compositing tasks.

Method used

An image generation system that uses an adapter network to translate image embeddings into descriptive embeddings, which are then used by an image generation model to create composite images, preserving object identity and appearance while harmonizing with the background, and incorporating elements like shadows and lighting.

Benefits of technology

The system generates composite images with enhanced realism and versatility, effectively integrating objects into backgrounds with consistent geometry and lighting, outperforming conventional methods by maintaining object identity and improving generation quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000002_0000
    Figure 00000002_0000
  • Figure 00000076_0000
    Figure 00000076_0000
  • Figure 00000077_0000
    Figure 00000077_0000
Patent Text Reader

Abstract

Systems and methods for image compositing are provided. An aspect of the systems and methods includes obtaining a first image and a second image, wherein the first image includes a target location and the second image includes a target element; encoding the second image using an image encoder to obtain an image embedding; generating a descriptive embedding based on the image embedding using an adapter network; and generating a composite image based on the descriptive embedding and the first image using an image generation model, wherein the composite image depicts the target element from the second image at the target location of the first image.20 24 20 30 33 08 M ay 2 02 4 A B S T R A C T 2 0 2 4 2 0 3 0 3 3 0 8 2 0 2 4 M a y 205205 FIG. 2 Generate a composite image depicting the first image and the element in the target area of the first image Provide the composite image to the user Provide a first image and a second image including an element for compositing into a target area of the first image 200200 215215 210210 2 / 172 / 17 o Provide a first image and a second image including an element for compositing into a target area of the 205 first image Generate a composite 210 image depicting the first image and the element in the target area of the first image Provide the composite image to the user 215 FIG. 2 200 20 24 20 30 33 08 M ay 2 02 4 2 / 1 7 2 0 2 4 2 0 3 0 3 3 0 8 M a y 2 0 2 4 2 1 5 2 0 0
Need to check novelty before this filing date? Find Prior Art