Department: Vision & Imaging
Available slots: 1 of 2 startups
3D Gaussian Splatting (3DGS) has opened a new level of photorealism for capturing and rendering real-world objects and spaces. Historically, production use was constrained by heavy models, slow processing, and high costs. Large files and limited mobile performance created a gap between demo quality and real-world deployment across many sectors. These constraints limited accessibility and scalability.
Startups building 3D experiences for web, mobile, or AR/VR need models that are both photorealistic and practical. Without compact representations and efficient reconstruction, even compelling use cases stay locked in proof-of-concept. Users expect instant loading, smooth interaction, and experiences that work on any device.
Our research at Fraunhofer HHI addresses this by pioneering next-generation 3DGS across the full pipeline – from data acquisition and scene reconstruction to efficient representations, compression, and real-time rendering. We develop state-of-the-art methods that produce compact 3DGS representations and reduce model sizes while preserving visual fidelity. This enables instant streaming and seamless experiences across mobile, AR/VR, and low-bandwidth environments.
This is a rare opportunity to build on research from leading experts in 3D vision and collaborate with a team focused on making photorealism production-ready. Startups can lower infrastructure costs, accelerate development, and deliver high-quality 3D experiences that run smoothly on mobile and beyond.
.svg)
.svg)
%201.svg)