PhysCapBench: A Cross-Domain Benchmark for Physically Grounded Video Captioning
cs.CV, cs.AI
Submitted: 2026-11-12
Updated: 2026-11-16
Project page: https://yangzhou24.github.io/OmniWorld
Terminology
Sources
- World Models
- Genie: Generative Interactive Environments
- Cosmos World Foundation Model Platform for Physical AI
- $\tau_0$-WM: A Unified Video-Action World Model for Robotic Manipulation
- Open X-Embodiment: Robotic Learning Datasets and RT-X Models
- DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset
- EgoDex: Learning Dexterous Manipulation from Large-Scale Egocentric Video
- Infinite Worlds with Versatile Interactions
- Panda-70M: Captioning 70M Videos with Multiple Cross-Modality Teachers
- CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer
- Physis-Lang: Self-Evolving Language as a Physical Representation for Video World Model
- FineVLA: Fine-Grained Instruction Alignment for Steerable Vision-Language-Action Policies
- Tarsier: Recipes for Training and Evaluating Large Video Description Models
- OmniFysics-Captioner Technical Report: Grounding Omni-Modal Understanding in the Physical World for Better Captioning
- RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulation
- $\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization
- VideoPhy: Evaluating Physical Commonsense for Video Generation
- Improving and Evaluating Hand-Object Interaction Detection
- Do generative video models understand physical principles?
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