Picasso: Holistic Scene Reconstruction with Physics-Constrained Sampling
cs.CV, cs.AI, cs.RO, cs.SY, eess.SY
Submitted: 2026-02-08
Updated: 2026-07-28
Comments: 15 pages, accepted to Robotics: Science and Systems (RSS) 2026
DOI: 10.15607/RSS.2026.XXII.173
Code: https://github.com/MIT-SPARK/Picasso
Project page: https://libigl.github.io
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Spatial Functa: Scaling Functa to ImageNet Classification and Generation
- Vysics: Object Reconstruction Under Occlusion by Fusing Vision and Contact-Rich Physics
- Strictly Constrained Generative Modeling via Split Augmented Langevin Sampling
- ShapeNet: An Information-Rich 3D Model Repository
- SAM 3D: 3Dfy Anything in Images
- From data to functa: Your data point is a function and you can treat it like one
- Diffusion Models for Constrained Domains
- 3DP3: 3D Scene Perception via Probabilistic Programming
- Bayes3D: fast learning and inference in structured generative models of 3D objects and scenes
- CUPID: Generative 3D Reconstruction via Joint Object and Pose Modeling
- PhysTwin: Physics-Informed Reconstruction and Simulation of Deformable Objects from Videos
- PhysPose: Refining 6D Object Poses with Physical Constraints
- Score-based constrained generative modeling via Langevin diffusions with boundary conditions
- QuantiPhy: A Quantitative Benchmark Evaluating Physical Reasoning Abilities of Vision-Language Models
- A Correct-and-Certify Approach to Self-Supervise Object Pose Estimators via Ensemble Self-Training
- Certifiable 3D Object Pose Estimation: Foundations, Learning Models, and Self-Training
- HoloScene: Simulation-Ready Interactive 3D Worlds from a Single Video
- TEASER: Fast and Certifiable Point Cloud Registration
- TwinTrack: Bridging Vision and Contact Physics for Real-Time Tracking of Unknown Objects in Contact-Rich Scenes
- Box Pose and Shape Estimation and Domain Adaptation for Large-Scale Warehouse Automation
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models