Nvidia patents camera system that spots defects using two angles
Nvidia describes a method for checking objects for flaws by taking photos from two different angles. The system uses a neural network to compare these images, highlighting specific features to make defects easier to see against the background.
In the filing’s own words
First image data of an object captured from a first viewpoint and second image data of the object captured from a second viewpoint is obtained. The first and second image data are combined into a representation that accentuates a feature of the object by determining a set of feature maps comprising features extracted from the first and second image data, generating a set of feature vectors by providing the set of feature maps from the set of convolutional layers to a set of fully connected layers of the neural network, and concatenating the set of feature vectors to increase contrast between the feature and one or more background features. An indication of whether the object exhibits a defect is generated, based, at least in part, on the representation.
What exactly they claim (claim 1)
1. One or more processors comprising circuitry to: obtain: first image data of an object from a first camera from a first viewpoint; and second image data of the object from a second camera from a second viewpoint; combine the first and second image data into a composite representation of the object that accentuates a feature of the object by causing the one or more processors to: determine, by providing the first and second image data as input to a set of convolutional layers of a neural network, a set of feature maps comprising a plurality of features extracted from the first and second image data; generate a set of feature vectors by providing the set of feature maps from the set of convolutional layers to a set of fully connected layers of the neural network; and concatenate the set of feature vectors to increase contrast between the feature and one or more background features; and generate, based, at least in part, on the composite representation, an indication of whether the object exhibits a defect.
Facts
- Company
- Nvidia (NVIDIA Corporation)
- Official title
- DEFECT DETECTION USING ONE OR MORE NEURAL NETWORKS
- Publication number
- US20260320775A1
- Application number
- 19467353
- Filed
- Feb 2, 2026
- Published
- Oct 15, 2026
- Inventors
- 2
- Claims
- 20
Classified under
G06T 7/001 Physics & computingG06N 3/04 Physics & computingG06T 7/586 Physics & computingG06T 7/97 Physics & computingG06T 2207/10016 Physics & computingG06T 2207/20221 Physics & computingMore from Nvidia in AI & software
- Nvidia patents AI coach that analyzes your gameplay to suggest strategies Oct 13, 2026
- Nvidia patents AI system that writes and voices character conversations Sep 10, 2026
- Nvidia patents system that uses AI to find bugs in game videos Sep 17, 2026
- Nvidia patents AI method to make digital characters move realistically Sep 8, 2026
- Nvidia patents AI system that builds image generators from text prompts Sep 15, 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.