Get the newsletter
Adobe · AI / ML · Filed Mar 27, 2025 · USPTO published Oct 1, 2026 · Oct 1 release

Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo

Moving an object in a photo sounds simple, but getting it to look like it belongs somewhere new is the hard part. Adobe has filed a patent for an AI system that clones an object, shape and all, into any spot you choose, then generates a version that fits its new surroundings.

A computer screen displays an image editing interface with multiple spoons of coffee beans cloned into a new region. Drawing from patent filing US 2026/0301257 A1.
A computer screen displays an image editing interface with multiple spoons of coffee beans cloned into a new region.All 27 drawings ↓
Publication number
US 2026/0301257 A1(US20260301257A1)
Applicant
ADOBE INC.
Filed · Published
Mar 27, 2025 · Oct 1, 2026
Inventors
Ajay Bedi, Nishant Gupta
Status when we published
Examiner raised objections; the company can respond
Document
20 claims
US classification
345/649
Examiner
MCCULLEY, RYAN D (Art Unit 2611)

What Adobe's generative object cloning actually does

A designer drops a chair into a corner of a room photo. The chair looks wrong: wrong shadow, wrong angle, wrong fit. Getting it to look natural takes real skill and a lot of manual work.

Adobe's patent describes a system that automates this. You pick the object you want to copy, pick where you want it to go, and the system figures out the object's exact outline. It places a silhouette of that shape in the new location, then hands both the original photo and that silhouette to an AI image model. The AI fills in the silhouette with a version of the object that fits its new context, adjusting for lighting, angle, and surroundings.

The result is a synthetic variant of the original object, not a pixel-for-pixel copy. It's built to look like it was always there.

From the filing · CLAIM 1
… computing a target mask based on the source object and the target region, wherein the target mask has a shape of the source object and is located within the target region …

Translation: The system traces the object you want to copy and places that outline wherever you want it to go.

How the mask guides the AI to place a cloned object

The system takes an input image that contains two things: a source object (the thing you want to clone) and a target region (where you want it to appear).

From there, the pipeline has three steps:

  • Mask computation: The system generates a target mask, a shape cutout that matches the outline of the source object and is placed inside the target region. Think of it as a stencil dropped into the destination area.
  • Image generation: An image generation model (an AI that creates or edits images) takes both the original photo and the stencil as inputs. It fills the stencil area with a new version of the object, one that matches the new location's lighting and context rather than just copying pixels.
  • Output: The result is an output image where the target region now contains a synthetic variant of the source object, generated to fit, not pasted.

The key distinction here is that the AI does not paste the original object. It generates a plausible version of it in the new position, which means shadows, angles, and surface details can adapt automatically.

What this means for everyday photo editing in Adobe tools

For anyone who edits photos regularly, this is the difference between cutting and pasting versus actually compositing. Current tools require manual masking, perspective correction, and lighting adjustments. This patent describes doing all of that in a single model pass.

If this capability lands in Adobe Photoshop or Firefly, it would put a professional-grade compositing shortcut in front of a much wider audience, including people who don't know what compositing is. The practical use cases range from product photography and interior design mockups to removing and rearranging elements in any photo without leaving obvious traces.

Adobe's 25th filing we've tracked on our controllable AI image watchlist since May builds on earlier applications like turning text into 3D worlds and prompts from cursor hover.

Source: US 2026/0301257 A1 at the USPTO (PDF). Full patent text and figures are in the official publication.

Get our take in your Top Stories

Liked this breakdown? Add Patentlyze as a preferred source on Google, and our plain-English take shows up more often in your Top Stories the next time Adobe patent news breaks.

There are more where this came from

We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.

The drawings · 27 sheets

27 drawing sheets from US 2026/0301257 A1 · click any drawing to enlarge

AI image cloning system overview., drawing sheet 1 from US 2026/0301257 A1
Sheet 1 AI image cloning system overview.
Steps for AI image cloning., drawing sheet 2 from US 2026/0301257 A1
Sheet 2 Steps for AI image cloning.
Selecting an object for cloning., drawing sheet 3 from US 2026/0301257 A1
Sheet 3 Selecting an object for cloning.
Generating cloned objects in an image., drawing sheet 4 from US 2026/0301257 A1
Sheet 4 Generating cloned objects in an image.
Multiple cloned objects in an image., drawing sheet 5 from US 2026/0301257 A1
Sheet 5 Multiple cloned objects in an image.
Different objects cloned into an image., drawing sheet 6 from US 2026/0301257 A1
Sheet 6 Different objects cloned into an image.
Method for generating a synthetic object., drawing sheet 7 from US 2026/0301257 A1
Sheet 7 Method for generating a synthetic object.
Image processing computer components., drawing sheet 8 from US 2026/0301257 A1
Sheet 8 Image processing computer components.
Show the other 19 sheets
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 9
Sheet 9
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 10
Sheet 10
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 11
Sheet 11
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 12
Sheet 12
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 13
Sheet 13
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 14
Sheet 14
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 15
Sheet 15
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 16
Sheet 16
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 17
Sheet 17
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 18
Sheet 18
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 19
Sheet 19
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 20
Sheet 20
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 21
Sheet 21
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 22
Sheet 22
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 23
Sheet 23
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 24
Sheet 24
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 25
Sheet 25
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 26
Sheet 26
Adobe Patents an AI Tool That Copies Objects Into New Parts of a Photo patent drawing, sheet 27
Sheet 27
Patent filing page

Watchlists this filing belongs to

All watchlists →
Reader comments

No comments yet

Start the discussion

Real name or a handle, either is fine. Comments are read by a person before they appear, so allow a little time. Keep it about the filing.