Techniques for creating digital collages

Patent metadata

Field	Value
Publication number	US-12499597-B2
Application number	US-202318372625-A
Country	US
Kind code	B2
Filing date	Sep 25, 2023
Priority date	Sep 25, 2023
Publication date	Dec 16, 2025
Grant date	Dec 16, 2025

Abstract

Systems and methods are disclosed for reflowing documents to display semantically related content. Embodiments may include receiving a request to view a document that includes body text and one or more images. A trimodal document relationship model identifies relationships between segments of the body text and the one or more images. A linearized view of the document is generated based on the relationships and the linearized view is caused to be displayed on a user device.

First claim

What is claimed is:

1 . A method, comprising:
receiving a selection of a collage template to generate a digital collage, the collage template comprising a digital frame of a set of digital frames; retrieving a reference image associated with the digital frame using an image fitting module; receiving a selection of a digital image from a graphical user interface (GUI); identifying a visual object within a target region of the digital image to be placed in the digital frame based on the reference image using a machine learning model; determining a defined position for the visual object within the digital frame based on a position of a visual object in a reference region of the reference image represented as coordinates in a coordinate system; and inserting the target region with the visual object at the defined position within the digital frame of the collage template, wherein the entire digital frame is filled with the target region.

2 . The method of claim 1 , comprising:
retrieving visual object data for the reference image associated with the digital frame using the machine learning model; and identifying the visual object within the target region of the digital image based on the visual object data for the reference image using the machine learning model; wherein the visual object data for the reference image comprises a reference segmented avatar representing a visual object within the reference image, the visual object comprising a person with a set of body segments, the reference segmented avatar comprising a set of key points, line segments, and angles between line segments representing the set of body segments for the person.

3 . The method of claim 1 , comprising generating a target segmented avatar representing the visual object within the digital image, the visual object comprising a person with a set of body segments, the target segmented avatar comprising a set of key points, line segments, and angles between line segments representing the set of body segments for the person.

4 . The method of claim 1 , comprising matching a portion of a reference segmented avatar for a visual object within the reference image with a portion of a target segmented avatar for the visual object within the digital image based on a shared number of key-points, a shared number of line segments, or a shared number of angles between line segments.

5 . The method of claim 1 , comprising identifying the visual object within the target region of the digital image based on a shared set of key points, line segments, and angles between line segments for a reference segmented avatar and a target segmented avatar.

6 . The method of claim 1 , comprising adjusting a visual property of the target region to place the visual object in the defined position within the digital frame based on a shared set of key points, line segments, and angles between line segments for a reference segmented avatar and a target segmented avatar.

7 . The method of claim 6 , wherein adjusting the visual property of the target region of the digital image comprises adjusting a size parameter, an angle parameter, an alignment parameter, a position parameter, a light parameter, a color parameter, a sharpness parameter, a filter parameter, a crop parameter, an orientation parameter, a transform parameter, a skew parameter, an aspect ratio parameter, an effect parameter, a spot removal parameter, an eye parameter, or a style parameter.

8 . The method of claim 1 , wherein the visual property comprises a size for the target region, further comprising:
calculating a reference average length across line segments in a reference segmented avatar for a visual object within the reference image; calculating a target average length across line segments in a target segmented avatar for the visual object within the digital image; and adjusting a size parameter of the target region using a ratio of the reference average length and the target average length.

9 . The method of claim 1 , comprising adjusting a parameter of the digital frame of the collage template to fit the modified digital image, wherein the parameter represents a shape adjustment, a size adjustment, a border adjustment, a text adjustment, an edge adjustment, a sticker adjustment, or a background adjustment.

10 . The method of claim 1 , comprising:
generating the digital collage from the collage template; and presenting the digital collage on a graphical user interface (GUI) of an electronic display of a client device.

11 . A system, comprising:
a memory component; and one or more processing devices coupled to the memory component, the one or more processing devices to perform operations comprising: presenting a graphical user interface (GUI) for an image editing system, the GUI to include GUI elements for creating a digital collage; receiving a selection of a collage template to generate a digital collage, the collage template comprising a digital frame of a set of digital frames; retrieving a reference image associated with the digital frame using an image fitting module; receiving a selection of a digital image from the GUI; identifying a visual object within a target region of the digital image to be placed in the digital frame based on the reference image using a machine learning model; determining a defined position for the visual object within the digital frame based on a position of a visual object in a reference region of the reference image represented as coordinates in a coordinate system; and inserting the target region with the visual object at the defined position within the digital frame of the collage template, wherein the entire digital frame is filled with the target region.

12 . The system of claim 11 , comprising:
retrieving visual object data for the reference image associated with the digital frame using the machine learning model; and identifying the visual object within the target region of the digital image based on the visual object data for the reference image using the machine learning model; wherein the visual object data for the reference image comprises a reference segmented avatar representing a visual object within the reference image, the visual object comprising a person with a set of body segments, the reference segmented avatar comprising a set of key points, line segments, and angles between line segments representing the set of body segments for the person.

13 . The system of claim 11 , comprising generating a target segmented avatar representing the visual object within the digital image, the visual object comprising a person with a set of body segments, the target segmented avatar comprising a set of key points, line segments, and angles between line segments representing the set of body segments for the person.

14 . The system of claim 11 , comprising matching a portion of a reference segmented avatar for a visual object within the reference image with a portion of a target segmented avatar for the visual object within the digital image based on a shared number of key-points, a shared number of line segments, or a shared number of angles between line segments.

15 . The system of claim 11 , comprising:
identifying the visual object within the target region of the digital image based on a shared set of key points, line segments, and angles between line segments for a reference segmented avatar and a target segmented avatar; and adjusting a visual property of the target region to place the visual object in the defined position within the digital frame based on the shared set of key points, line segments, and angles bet

Assignees

• Adobe Inc

Inventors

• Tagra Sanjeev
• Soni Sachin
• Mondal Prasenjit
• Jain Ajay

Classifications

• G06V40/10 Primary

  Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands · CPC title

• G06T2200/24

  involving graphical user interfaces [GUIs] · CPC title

• G06T2207/30196

  Human being; Person · CPC title

• G06T7/11

  Region-based segmentation · CPC title

• G06T2207/20081

  Training; Learning · CPC title