Dynamic motion with physics powered by Viser — equivalent to IssacGym simulation output.

Interactive Demos

Click and Drag / Zoom-in(out) interactive demo to see simulation result everywhere (when not shown please refresh page).

Page 1 / 1

56 outdoor stairs up down

Input Video

Interactive (ours)

58 outdoor parcours

Input Video

Interactive (ours)

36 outdoor long walk

Input Video

Interactive (ours)

09 outdoor walk

Input Video

Interactive (ours)

20 outdoor walk

Input Video

Interactive (ours)

24 outdoor long walk

Input Video

Interactive (ours)

19 indoor walk off mvs

Input Video

Interactive (ours)

27 indoor walk off mvs

Input Video

Interactive (ours)

28 outdoor walk lunges

Input Video

Interactive (ours)

35 indoor walk

Input Video

Interactive (ours)

40 indoor walk big circle

Input Video

Interactive (ours)

48 outdoor walk downhill

Input Video

Interactive (ours)

MPH11

Input Video

Interactive (ours)

MPH16

Input Video

Interactive (ours)

MPH8

Input Video

Interactive (ours)

PKR 5

Input Video

Interactive (ours)

PKR c

Input Video

Interactive (ours)

PKR GRN 2

Input Video

Interactive (ours)

PKR g

Input Video

Interactive (ours)

PKR GRN 1

Input Video

Interactive (ours)

PKR GRN 8

Input Video

Interactive (ours)

Yoga Ball 1

Input Video

Interactive (ours)

PKR b

Input Video

Interactive (ours)

Sandbag 3

Input Video

Interactive (ours)

Soraiii

Input Video

Interactive (ours)

Interactive Comparisons

Inspect VideoMimic reconstructions next to CRISP and the original footage. Each card shows the input video alongside both interactive viewers for direct comparison.

Page 1 / 1

36 outdoor long walk

Input Video

CRISP (ours)

VideoMimic

56 outdoor stairs up down

Input Video

CRISP (ours)

VideoMimic

27 indoor walk off mvs

Input Video

CRISP (ours)

VideoMimic

35 indoor walk

Input Video

CRISP (ours)

VideoMimic

58 outdoor parcours

Input Video

CRISP (ours)

VideoMimic

48 outdoor walk downhill

Input Video

CRISP (ours)

VideoMimic

19 indoor walk off mvs

Input Video

CRISP (ours)

VideoMimic

40 indoor walk big circle

Input Video

CRISP (ours)

VideoMimic

MPH11

Input Video

CRISP (ours)

VideoMimic

MPH16

Input Video

CRISP (ours)

VideoMimic

N3 Office

Input Video

CRISP (ours)

VideoMimic

PKR c

Input Video

CRISP (ours)

VideoMimic

PKR b

Input Video

CRISP (ours)

VideoMimic

PKR 5

Input Video

CRISP (ours)

VideoMimic

Method

CRISP reconstructs physics-ready motion and scene primitives from monocular video, then grounds a contact-aware policy for simulation and downstream control.

Figure 1: CRISP pipeline from video to reconstructed simulation assets
Fig. 1. CRISP pipeline overview.
Figure 2: Contact-guided policy learning and deployment
Fig. 2. Details of Planar Fitting.

Scene-Aware Policy

Click and Drag / Zoom-in(out) inside each Viser viewer to orbit the scene and scroll to zoom. Press play to watch the simulation from your chosen angle.

View the scene-aware policy gallery

Citation

If you find our repository useful, please consider giving it a star ⭐ and citing our paper in your work:

@misc{wang2025crispcontactguidedreal2simmonocular,
  title={CRISP: Contact-Guided Real2Sim from Monocular Video with Planar Scene Primitives},
  author={Zihan Wang and Jiashun Wang and Jeff Tan and Yiwen Zhao and Jessica Hodgins and Shubham Tulsiani and Deva Ramanan},
  year={2025},
  eprint={2512.14696},
  archivePrefix={arXiv},
  primaryClass={cs.CV},
  url={https://arxiv.org/abs/2512.14696},
}

Acknowledgements

We extend our deepest gratitude to Guanya Shi for invaluable feedback and insightful discussions on CRISP. We thank Qitao Zhao for writing suggestions, Arthur Allshire for helping set up baselines, Zhengyi Luo and Yuxuan Kuang for early stage discussions, Luna Shi and Weiyu Li for helpful discussions. The research project is funded by Bosch Research Center for Artificial Intelligence.