Patent · US Active

Content synthesis using generative Artificial Intelligence model

US12271978B1 · kind B1 · utility

0Cited by

0References

20Claims

0Family size

Assignee

AI Incorporated · CA

Inventors

Rahim Entezari · London, GB
Patrick Esser · Atlanta, US
Robin Rombach · London, GB
Andreas Blattmann · Waldkirch, DE

Key dates

Filing date	Sep 11, 2024
Grant date	Apr 8, 2025
Priority date	—
Expiry date	Sep 11, 2044

Classification

Technology area (CPC G)Physics
CPC primaryG06F40/40
WIPO fieldComputer technology
WIPO sectorElectrical engineering

Abstract

A method including receiving a prompt describing a desired characteristic of an image. The method further including generating, using a set of encoding models, a prompt encoding based on the prompt. The method further including generating, using a first transformer block of a diffusion transformer model, a first prompt embedding and a first image embedding based on the prompt encoding and a noise input. The method further including generating, using a second transformer block of the diffusion transformer model, a second image embedding based on the first image embedding and the first prompt embedding. The method further including generating the image based on the second image embedding.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.