Check out Marcel's amazing talk and synthetic dataset! Given 3 photos from your cell phone, you can synthesize high-quality novel views 🤯

#3DV2025AMA Fourth guest on the Ask Me Anything series: Jia-Bin Huang Jia-Bin Huang from University of Maryland & Meta! Jia-Bin was the engine behind all the fun tweets of last year 3DV 🌟 🕒 You have now 24 HOURS to ask him anything — drop your questions in the comments below!

Interested in making image latent diffusion models more efficient? Check out Morteza's NeurIPS paper!

Our NeurIPS work Omnigrasp's poster session will be at this Friday! 11 a.m. PST — 2 p.m. PSTEast Exhibit Hall A-C #4108 Feel free to reach out if you are interested in the work!

🤖 Introducing Human-Object Interaction from Human-Level Instructions! First complete system that generates physically plausible, long-horizon human-object interactions with finger motions in contextual environments, driven by human-level instructions. 🔍 Our approach: - LLMs

Interested in: ✅Humanoids mastering scalable motor skills for everyday interactions ✅Whole-body loco-manipulation w/ diverse tasks and objects ✅Physically plausible HOI animation Meet InterMimic #CVPR2025 ArXiv: arxiv.org/abs/2502.20390 Project: sirui-xu.github.io/InterMimic 🧵[1/9]

🚀 Check out our #CVPR2025 paper BimArt! 👐🔧 We generate 3D bimanual interactions with articulated objects by: ✅ Predicting contact maps from object trajectories ✅ Using an articulation-aware feature representation Project Webpage: vcai.mpi-inf.mpg.de/projects/bimar… VCAI - MPI for Informatics

🚨Introduced at #GTC25: Newton is an open-source extensible physics engine being developed by NVIDIA, Google DeepMind, and Disney Research to advance robot learning and development. ⚡️ Learn how physics engines and #physicalAI models are driving the future of #humanoidrobots.

This year, an important mission is to let simulated humanoids solve long-term manipulation tasks in challenging dynamic, cluttered environments. This is a simple demonstration achieved by TokenHSI (CVPR 2025). The humanoid here is controlled by a single, unified model.

📢 Taking a single image of a human grasping an object and having real robot hands grasp the object accordingly? Such one-shot functional grasping is now possible with FunGrasp, published in RA-L! Happy to finally have transferred our method D-Grasp to the real world 🥳

Want to grasp almost "anything" with a dexterous hand? Our recent work, RobustDexGrasp, makes it possible. Trained entirely in simulation, it achieves 94.6% success on 500+ real objects with diverse shapes, materials, and weights, while robustly adapting to various disturbances.

A robot hand grasp over 500 totally new objects without fail? Zero-shot, single-view & super reliable ⬇️ + Paper Grasping random objects is hard for robots, especially when shapes, weights, and materials vary. RobustDexGrasp solves this with a smart new way of seeing and

📢 Interested in generating realistic hand motions from text prompts? If you are at CVPR this week, come check out our paper LatentHOI, a method that uses latent diffusion and generalizes well to unseen object geometry! 📍 Poster Session 4, #155 📅 Saturday

If you are at #RSS2025, check out our workshop on Generative AI for Human-Robot Interaction. We have a stacked lineup of speakers and panelists!

Sadly, I am no longer a professor at ETH (@eth_en) due to very severe #longCovid and #MECFS. ethrat.ch/de/ernennungen….

🎉Excited to share RobustDexGrasp is accepted by #CoRL2025! A dexterous grasping policy trained with 35 objects in simulation, achieving 94.6% real-world success on 500+ diverse objects. The code has been released. See you in Seoul! 🌐 Project Website: zdchan.github.io/Robust_DexGras…

Glad to share that FunGrasp has won the Best Poster Award on the 2nd Workshop on Dexterous Manipulation #corl2025!