3DforXR: an AI-Driven Multi-Input Pipeline for 3D Object Generation in XR Applications
M. S. Kawka, A. Nikopoulos, Christos Stentoumis, I. Kalisperakis · 2024
We present a simple and user-friendly platform for generating 3D objects ready for use in eXtended Reality (XR) applications. The platform exploits three types of inputs: multiple images, a single image, and a text prompt. We combine and modify up-to-date artificial intelligence-based algorithms to develop a robust pipeline for single-object 3D reconstruction. We also support a post-processing pipeline for editing the mesh and the texture. Non-expert users have tested our work-in-progress platform and provided positive feedback regarding its robustness and effectiveness.