Scalable GenXR. Generative AI for extended reality in retail

Abstract

GenXR advances AI‑driven 3D and AR content creation by developing a next‑generation platform for the fully automated generation of high‑quality 3D product models from single or few 2D images. The project addresses key limitations of current workflows - most notably the need for extensive manual post‑processing - by introducing novel machine‑learning methods based on generative AI, deep learning, and neural rendering. A systematic benchmarking of state‑of‑the‑art approaches is combined with research into autonomous refinement and perceptual quality optimization. In addition, GenXR explores energy‑efficient model generation to reduce the environmental footprint of large‑scale AI‑based 3D content pipelines. The project targets scalable, accessible, and sustainable solutions applicable across retail and other digital content domains.