How do we represent 3D world knowledge for spatial intelligence in next-generation robots? We recently wrote an extensive survey paper on this emerging topic, covering recent state-of-the-art! 🦾 🚀
Check it out below. Feedback/Suggestions welcome!
📖arXiv: https://t.co/SbrPYKJRwk
🖥️github list: https://t.co/VecNOsxyM3
@GTrobotics@ICatGT@mlatgt@ToyotaResearch@neural_fields
Very cool blog showing MapAnything for robotics!
PS: The title and narrative is a bit spicy though 🌶️
Stayed tuned for some exciting updates coming this week!! 😉
https://t.co/rwmIxtO6KF
Segment Anything 3 just silently dropped on ICLR 🤯
The first SAM let you click on an object to segment it.
SAM 2 added video and memory.
Now SAM 3 says: just describe what you want — “yellow school bus”, “striped cat”, “red apple” — and it will find and segment every instance of it across images or videos.
It’s like segmentation finally learned language, but in a grounded, minimal way, by just saying concepts.
Short noun phrases, image exemplars, or both.
You say the concept, it understands what you mean, and draws the boundary everywhere it appears.
Behind it is a massive data engine: 4 million unique concepts, 52 million verified masks, built through humans and LLMs checking each other’s work.
It’s both old-school annotation and self-improving AI loops.
The result: roughly 2× better than SAM 2, near-human recognition on open-vocab benchmarks, and real-time speed.
RayletDF: Raylet Distance Fields for Generalizable 3D Surface Reconstruction from Point Clouds or Gaussians
Shenxing Wei, Jinxi Li, Yafei Yang, Siyuan Zhou, @boyang_vLAR
tl;dr: raylet distance field->signed distance between the surface hit point and the raylet starting point
https://t.co/WgLzcgCOK3
A Survey on 3D Gaussian Splatting Applications: Segmentation, Editing, and Generation
Shuting He, Peilin Ji, Yitong Yang, Changshuo Wang, Jiayi Ji, Yinglin Wang, Henghui Ding
tl;dr: in title
https://t.co/1WlSjF0Coh
A View-consistent Sampling Method for Regularized Training of Neural Radiance Fields
@AoxiangFan, @CorentinDumery, @NicolasTalabot, @FuaPv
tl;dr: view-consistency distributions per-ray instead of fixed depth predictions->regularize NeRF training
https://t.co/MYhgEolor8
CL-Splats: Continual Learning of Gaussian Splatting with Local Optimization
TL;DR: "CL-Splats efficiently updates existing scene reconstruction from an incremental set of images while enabling new applications."
Contributions:
• We introduce a general framework for continual and localized updates in Gaussian Splatting representations, enabling flexible and efficient scene reconstruction.
• We propose a lightweight yet effective change-detection module and present the first dynamic local Gaussian optimization with a custom kernel implementation, improving speed and adaptability to evolving scenes.
• We empirically validate our approach on diverse real and synthetic scenes, demonstrating its effectiveness. Additionally, we show its ability to perform tasks such as batched updates and scene history recovery.
• We contribute novel synthetic and real-world datasets designed explicitly for benchmarking dynamic scene reconstruction, which we plan to release publicly to support further research.
Active View Selector: Fast and Accurate Active View Selection with Cross Reference Image Quality Assessment
@ziruiwang_, @ysbhalgat, @RayLi234, @viprad
tl;dr: CrossScore helps active NVS
https://t.co/HXSZHVYtWd