AI & Enterprise
Nvidia unveils ArtiFixer AI model for 3D scene reconstruction
Nvidia has unveiled ArtiFixer, an AI model that generates 3D scenes from multiple photos. The model fills in areas not shown in the reference images, aiming to reduce quality degradation and scene inconsistencies seen in existing 3D reconstruction methods. Built on the Wan 2.1 video-generation model, ArtiFixer has about 16.9 billion parameters. Nvidia released three versions and said ArtiFixer3D+ produces scenes that are both sharp and consistent.
