Hardware
Scale AV Perception Across Vehicle Platforms with NVIDIA Omniverse NuRec
NVIDIA Omniverse NuRec reconstructs real‑world driving scenes through 3D Gaussian splatting to produce the Physical AI NuRec Dataset—over 1,500 neural‑reconstructed videos (~20 s each) captured from s
NVIDIA Omniverse NuRec reconstructs real‑world driving scenes through 3D Gaussian splatting to produce the Physical AI NuRec Dataset—over 1,500 neural‑reconstructed videos (~20 s each) captured from six camera views, available on Hugging Face. Developers pair a reconstructed scene with a target vehicle rig, render new camera perspectives, optionally refine temporally consistent frames via NVIDIA Harmonizer, and train perception models; internal tests show significant improvements in object‑detection precision and recall for new carlines compared to zero‑shot baselines. The open‑source nurec-skills repository packages this workflow—scene download, rendering, Harmoner application—and includes documentation on setup, supported hardware, validation, and rendering instructions.
Related
- Integrate Physical AI Capabilities into Existing Apps with NVIDIA Omniverse Libraries
- Integrate NVIDIA Omniverse RTX Sensor Simulation Into Existing Apps
- A Scalable PyTorch Abstraction for Multi-GPU Gaussian Splatting
Source: NVIDIA Developer | 2026-08-31