Google patents neural network design that normalizes image data twice
Google describes a neural network structure for analyzing images that splits pictures into small patches and processes them through a specific sequence of steps. The design includes two distinct normalization stages, one before and one after converting the image patches into data vectors, before passing the information to a transformer block for final analysis.
In the filing’s own words
One example aspect of the present disclosure is directed to a neural network for machine vision. The neural network may include a stem block that includes a set of stem layers. The neural network may additionally include a visual transformer block. The set of stem layers may include a patch layer, a first normalization layer, an embedding layer, and a second normalization layer. The patch layer subdivides an input image into a set of image patches. The first normalization layer generates a set of normalized image patches by performing a first normalization process on each image patch of the set of image patches. The patch layer feeds forward to the first normalization layer. The embedding layer generates a set of vector embeddings. Each vector embedding of the set of embedding vectors is a projection of a corresponding normalized image patch from the set of normalized image patches onto a visual token. The first normalization layer feeds forward to the embedding layer. The second normalization layer generates a set of normalized vector embeddings by performing a second normalization process on each vector embedding of the set of vector embeddings. The embedding layer feeds forward to the second normalization layer. The transformer block enables one or more machine vision tasks for the input image based on the set of normalized vectors. The second normalization layer feeds forward to the transformer block.
What exactly they claim (claim 1)
1. A computing system comprising: one or more processors; and one or more non-transitory computer-readable media that store instructions that when executed by the one or more processors, cause the computer system to implement a neural network comprising: a stem block that includes a set of stem layers comprising: a patch layer that subdivides an input image into a set of image patches; a first normalization layer that generates a set of normalized image patches by performing a first normalization process on each image patch of the set of image patches, wherein the patch layer feeds forward to the first normalization layer; an embedding layer that generates a set of vector embeddings, wherein each vector embedding of the set of embedding vectors is a projection of a corresponding normalized image patch from the set of normalized image patches onto a visual token and the first normalization layer feeds forward to the embedding layer; and a second normalization layer that generates a set of normalized vector embeddings by performing a second normalization process on each vector embedding of the set of vector embeddings, wherein the embedding layer feeds forward to the second normalization layer; and a transformer block that enables one or more machine-learned tasks for the input image based on the set of normalized vector embeddings, wherein the second normalization layer feeds forward to the transformer block.
Facts
- Company
- Google (Google LLC)
- Official title
- Neural Network Architectures with Multiple Normalization Layers for Machine Vision
- Publication number
- US20260301398A1
- Application number
- 19694705
- Filed
- Jun 1, 2026
- Published
- Oct 1, 2026
- Inventors
- 3
- Claims
- 20
Classified under
G06V 10/82 Physics & computingG06V 10/26 Physics & computingMore from Google in AI & software
- Google patents video search that answers questions using AI descriptions Sep 15, 2026
- Google patents photo tool that removes people and their gear Sep 15, 2026
- Google patents AI that writes live meeting notes and summaries Sep 15, 2026
- Google patents AI tool that turns messy emergency notes into clear messages Sep 3, 2026
- Google patents AI that clicks through apps to finish tasks for you Sep 24, 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.