ALEXANDRIA, Va., Sept. 24 -- United States Patent no. 12,743,819, issued on Sept. 22, was assigned to NVIDIA Corp. (Santa Clara, Calif.).
"Training-free consistent text-to-image generation" was invented by Yoad Tewel (Tel Aviv, Israel), Rinon Pery Gal (Tel Aviv Yafo, Israel), Yehonatan Kasten (Hinanit, Israel), Yuval Atzmon (Hod Hasharon, Israel) and Gal Chechik (Ramat Hasharon, Israel).
According to the abstract* released by the U.S. Patent & Trademark Office: "Embodiments of the present disclosure relate to training-free consistent text-to-image generation. A pre-trained text-to-image diffusion model is leveraged to generate images depicting a consistent subject for diverse prompts describing scenes. Inputs to the model are a text descr...