OVA is coordinating a new and innovative research project that aims to transform industrial VR training by integrating NeRF (Neural Radiance Field) technology with traditional 3D content generation methods. This €790.000 project, carried out in collaboration with Fraunhofer IIS and DGG (Darmstadt Graphics Group GmbH), has been funded by National Research Council Canada (NRC), XR Interaction Innovation Network, and the German Federal Ministry for Economic Affairs and Climate Action (as part of Central Innovation Programme for small and medium-sized enterprises (ZIM)).
This project emerged as a result of an initial bridge between StellarX and RapidPipeline, DGG’s 3D optimization software, which optimizes meshes and converts them into various formats. OVA’s exploration of NeRFs and gaussian splat led to partner Fraunhofer being added to the picture and, finally, to the elaboration of a full fledged R&D project with NRC IRAP’s support.
The main goal of this German-Canadian joint project is to explore an asset importation workflow that allows users to capture images with a simple device, like a smartphone camera, and convert them into 3D meshes, NeRFs, or potentially gaussian splats assets to be used in immersive environments. Key objectives in the roadmap include developing interfaces for seamless NeRF to mesh conversion, optimizing real-time VR simulation environments, and ensuring efficient performance on VR hardware.
This technology would be tremendously helpful in use cases requiring specialized 3D models (such as industrial hardware, for example) that cannot be found online. By automating the capture and modeling of detailed, real-world objects and environments, we eliminate the need for lengthy and expensive customized asset creation. High-quality 3D content becomes more accessible, which significantly improves the efficiency, realism, and practicality of immersive training scenarios for industrial applications.
We are stoked to lead this R&D project, and hope to integrate its developments into StellarX.
.png){kind=small}