Contact-Centric Deformation Learning

Cristian Romero, Dan Casas, Maurizio M. Chiaramonte and Miguel A. Otaduy
ACM Transactions on Graphics (Proc. of SIGGRAPH), 2022


We propose a novel method to machine-learn highly detailed, nonlinear contact deformations for real-time dynamic simulation. We depart from previous deformation-learning strategies, and model contact deformations in a contact-centric manner. This strategy shows excellent generalization with respect to the object's configuration space, and it allows for simple and accurate learning. We complement the contact-centric learning strategy with two additional key ingredients: learning a continuous vector field of contact deformations, instead of a discrete approximation; and sparsifying the mapping between the contact configuration and contact deformations. These two ingredients further contribute to the accuracy, efficiency, and generalization of the method. We integrate our learning-based contact deformation model with subspace dynamics, showing real-time dynamic simulations with fine contact deformation detail.


@article {romero2022contactcentric,
    author  = {Romero, Cristian and Casas, Dan and Chiaramonte, Maurizio M. and Otaduy, Miguel A.},
    title   = {Contact-Centric Deformation Learning},
    number  = "4",
    volume  = "41",
    journal = {ACM Transactions on Graphics (Proc. of ACM SIGGRAPH)},
    year    = {2022}

Description and Results

We present a learning-based method to augment a subspace deformable simulation with contact-driven deformation detail. We learn contact deformations in a contact-centric manner, which allows us to significantly reduce the sampling of configurations of the deformable object, and subsequently learn highly complex deformations.

Instead of learning a discrete approximation of the contact deformation field, we learn the continuous field directly, inspired by recent work on implicit surface modeling. Our method generalizes continuity and differentiability to unseen configurations.

We demonstrate that our contact-centric approach shows excellent generalization with respect to the object’s subspace state.

Also, to learn effectively from sparse data, we sparsify the mapping between the contact configuration and the resulting contact deformations.

For additional results and comparisons, check the supplementary video.


We wish to thank the anonymous reviewers for their helpful comments. We also thank Mickeal Verschoor and Suzanne Sorli for help with the hand demo. This work was funded in part by the European Research Council (ERC Consolidator Grant 772738 TouchDesign).


Cristian Romero –
Dan Casas –
Maurizio M. Chiaramonte –
Miguel A. Otaduy –