Nvidia patents method to train AI to see objects hidden behind others
Nvidia describes a technique for training computer vision models to identify the full shape of objects, even when parts are hidden behind other items. The method uses a two-step process where the model first learns from examples of hidden objects, then creates synthetic images to practice recognizing these hidden shapes more accurately.
In the filing’s own words
The disclosed method for training a machine learning model for image segmentation includes performing, based on amodal training data, one or more training operations to train a machine learning model to generate a trained machine learning model, where the trained machine learning model is trained to process a first image and a first bounding box prompt and generate a first predicted amodal mask; generating, based on modal training data and using the trained machine learning model, unoccluded object data, occluded object data, and one or more synthetic images; and performing, based on the unoccluded object data, the occluded object data, and the one or more synthetic images, one or more training operations to retrain the trained machine learning model to generate a retrained machine learning model, where the retrained machine learning model processes a second image and a second bounding box prompt to generate a second predicted amodal mask.
What exactly they claim (claim 1)
1. A computer-implemented method for training a machine learning model for image segmentation, the method comprising: performing, based on amodal training data, one or more first training operations to train a machine learning model to generate a trained machine learning model, wherein the trained machine learning model is trained to process a first image and a first bounding box prompt and generate a first predicted amodal mask; generating, based on modal training data and using the trained machine learning model, unoccluded object data, occluded object data, and one or more synthetic images; and performing, based on the unoccluded object data, the occluded object data, and the one or more synthetic images, one or more second training operations to retrain the trained machine learning model to generate a retrained machine learning model, wherein the retrained machine learning model processes a second image and a second bounding box prompt to generate a second predicted amodal mask.
Facts
- Company
- Nvidia (NVIDIA CORPORATION)
- Official title
- AMODAL OBJECT SEGMENTATION AND SYNTHETIC AMODAL DATA GENERATION
- Publication number
- US20260260357A1
- Application number
- 19410875
- Filed
- Dec 5, 2025
- Published
- Sep 3, 2026
- Inventors
- 5
- Claims
- 20
Classified under
G06T 7/11 Physics & computingG06T 11/00 Physics & computingG06T 2207/20076 Physics & computingG06T 2207/20081 Physics & computingG06T 2207/20084 Physics & computingG06T 2207/20092 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.