Grasping the Future: Joint 3D Hand-Object Forecasting From Internet Videos
cs.CV
Submitted: 2026-11-09
Updated: 2026-11-09
Code: https://github.com/kabouzeid/JointTransformer
Project page: https://grasping-the-future.github.io
Terminology
Sources
- MotionForesight: Re-purposing Video Models for Future 3D Scene-Flow Prediction
- SAM 3: Segment Anything with Concepts
- Action100M: A Large-scale Video Action Dataset
- SAM 3D: 3Dfy Anything in Images
- Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding
- The Invisible EgoHand: 3D Hand Forecasting through EgoBody Pose Estimation
- Bimanual 3D Hand Motion and Articulation Forecasting in Everyday Images
- DINOv3
- HO-Cap: A Capture System and Dataset for 3D Reconstruction and Pose Tracking of Hand-Object Interaction
- MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs