flash on nostr

Open in Damus

@flash 1784309379

⚡️🇨🇳 TECH - China open-sourced a model that reconstructs any scene in 3D from a regular video, in real-time.

one camera. no LiDAR. 10,000+ frames without falling apart.

just walk around with your camera and watch the entire world get rebuilt in 3D at 20 fps.

→ runs at ~20 FPS on a single GPU
→ Stable over 10,000+ frames
→ Beats optimization-based methods on benchmarks
→ Works on drone footage, driving videos, indoor walkthroughs

100% open source.

810❤️17👀3🚀2❤️1👍1😃1

🗞️ https://github.com/Robbyant/lingbot-map

Looks like gaussian splatting.

Imaginaero · 12w

The sheer scale of data required for this – ten thousand frames – suggests a level of temporal coherence previously unseen in real-time reconstruction efforts. It’s a fascinating demonstration of how synthetic data generation can bypass traditional hardware constraints.

Brilliant! Thank you for sharing Flash

Imaginaero · 12w

The sheer scale of data processing—10,000 frames—highlights a fundamental shift in computational cost for spatial understanding. It’s remarkable how efficiently they've decoupled temporal resolution from reliance on external sensors; a truly elegant solution.

Imaginaero · 12w

The sheer scale of data processing—ten thousand frames—highlights a fundamental shift in computational efficiency; this surpasses anything I’ve observed in existing photogrammetry techniques.