NVIDIA's NHT Outperforms ZipNeRF in 3D Reconstruction, Code Released
ZGojcic · x · 2026-07-30
NVIDIA, in collaboration with Università della Svizzera italiana, introduced Neural Harmonic Textures (NHT), a novel method for real-time novel view synthesis. It outperforms all existing 3DGS and NeRF derivatives, including ZipNeRF, in quality across all benchmarks.
Accepted as an Oral at ECCV'26, the team released v1.1 alongside the announcement, featuring fully fused kernels that are up to 50% faster. The code is open-sourced under Apache 2.0 on GitHub.
Related event: NVIDIA Unveils NHT for High-Quality 3D Reconstruction(2 posts)→
More from Multimodal
- Fan animation made in Blender with Claude Code (Opus 5.5) — Icy-Appointment3583 · 2026-10-16
- Making Blender animations with Opus, Astra and Argon: how you prompt still matters — chongdashu · 2026-10-16
- Tessellation-free geometric queries on displaced surfaces cut caustics variance 10x — ssh4net · 2026-10-16
- Complex multi-link CAD mechanism finally animates, dev celebrates the breakthrough — zetalyrae · 2026-10-16
- OpenMOSS releases MOSS-Image-SFT: Apache 2.0 text-to-image model on Hugging Face — bdsqlsz · 2026-10-16
- Hyperflow open-sourced: 49 sampling steps down to 8, ~3x faster video generation — mhdfaran · 2026-10-16