Data recognition apparatus and method, and training apparatus and method

This invention describes a method for recognizing data, such as images or audio, by processing it through a series of steps. First, it extracts key characteristics from the raw data. Then, it groups similar parts of these characteristics together. Finally, it combines these grouped parts into a unique digital signature, called an embedding vector. This unique signature can then be used to identify the input data, for example, by matching it to a specific registered user.

Why it matters: Filed before the widespread maturation of deep learning frameworks and pre-trained models. The ability to generate highly discriminative embedding vectors for user recognition has advanced significantly, making such systems more accurate and efficient to build today.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedComplex machine learning model software

3 AI reinvention directions for this patent

See all three

AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.

Reinvent this with AI

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

This one

+10