Amazon patents 3D model creation from single photos
Amazon describes a system that uses a neural network to turn a single 2D photo into a 3D model. It does this by predicting multiple layers of depth information, similar to stacking transparent sheets, to build the object's shape.
In the filing’s own words
The system generates a three-dimensional model with layered depth images based on an input two-dimensional image. For training, layered depth images are derived from existing three-dimensional models. The system trains a machine learning model to predict multiple layered depth images from an input image of an object. The system compares the generated, multiple layered depth images to the derived layered depth images for the object to update the machine learning model during training. At inference time, the system receives an input image for an object. The system applies the machine learning model to the input image to output predicted layered depth images. The system generates a three-dimensional model from the predicted layered depth images.
What exactly they claim (claim 1)
1. A system comprising: a data storage medium to store specific computer-executable instructions; and a computer hardware processor in communication with the data storage medium and configured to execute the specific computer-executable instructions to at least: generate, via a neural network, a first set of layered depth images from a first input image of a first object; generate a first three-dimensional model from the first set of layered depth images; output, with a camera, an output image based at least in part on the first three-dimensional model; determine a loss and a gradient of the neural network from a comparison of the first input image and the output image; update the neural network, wherein to update the neural network the computer hardware processor is configured to execute the specific computer-executable instructions to update a weight or a bias of the neural network based at least in part on the loss and the gradient; generate, via the neural network that is updated, a second set of layered depth images from a second input image of a second object; generate a second three-dimensional model from the second set of layered depth images; and cause presentation of the second three-dimensional model.
Facts
- Company
- Amazon (Amazon Technologies, Inc.)
- Official title
- GENERATIVE MODELING OF THREE-DIMENSIONAL OBJECT WITH LAYERED DEPTH IMAGES
- Publication number
- US20260278972A1
- Application number
- 19189090
- Filed
- Apr 24, 2025
- Published
- Sep 17, 2026
- Inventors
- 4
- Claims
- 20
Classified under
G06T 19/20 Physics & computingG06T 5/50 Physics & computingG06T 5/70 Physics & computingG06T 15/04 Physics & computingG06T 15/205 Physics & computingG06T 17/20 Physics & computingMore from Amazon in AI & software
- Amazon patents system that cites sources for AI answers Sep 29, 2026
- Amazon patents tech to swap words in movies using actor's voice Sep 8, 2026
- Amazon patents ML tool to add graphics to live sports from moving cameras Sep 15, 2026
- Amazon patents shopping assistant that remembers your search filters Oct 6, 2026
- Amazon patents system where AI agents split tasks and talk to each other Oct 20, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.