Neural Material: Learning Elastic Constitutive Material and Damping Models from Sparse Data

The accuracy and fidelity of deformation simulations are highly dependent upon the underlying constitutive material model. Commonly used linear or nonlinear constitutive material models only cover a tiny part of possible material behavior. In this work we propose a unified framework for modeling deformable material. The key idea is to use a neural network to correct a nominal model of the elastic and damping properties of the object. The neural network encapsulates a complex function that is hard to explicitly model. It injects force corrections that help the forward simulation to more accurately predict the true behavior of a given soft object, which includes non-linear elastic forces and damping. Attempting to satisfy the requirement from real material interference and animation design scenarios, we learn material models from examples of dynamic behavior of a deformable object's surface. The challenge is that such data is sparse as it is consistently given only on part of the surface. Sparse reduced space-time optimization is employed to gradually generate increasingly accurate training data, which further refines and enhances the neural network. We evaluate our choice of network architecture and show evidence that the modest amount of training data we use is suitable for the problem tackled. Our method is demonstrated with a set of synthetic examples.

[1]  Paolo Cignoni,et al.  Elastic textures for additive fabrication , 2015, ACM Trans. Graph..

[2]  Steve Marschner,et al.  Modeling and estimation of internal friction in cloth , 2013, ACM Trans. Graph..

[3]  A. J. Fossard,et al.  Modeling and estimation , 1995 .

[4]  Mihai Anitescu,et al.  Real-Time Stochastic Optimization of Complex Energy Systems on High-Performance Computers , 2014, Computing in Science & Engineering.

[5]  Sepp Hochreiter,et al.  Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs) , 2015, ICLR.

[6]  Andrew P. Witkin,et al.  Spacetime constraints , 1988, SIGGRAPH.

[7]  Jamshid Ghaboussi,et al.  Autoprogressive training of neural network constitutive models , 1998 .

[8]  Ming C. Lin,et al.  Learning-Based Cloth Material Recovery from Video , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[9]  Miguel A. Otaduy,et al.  High-resolution interaction with corotational coarsening models , 2016, ACM Trans. Graph..

[10]  L. Liang,et al.  A deep learning approach to estimate chemically-treated collagenous tissue nonlinear anisotropic stress-strain responses from microscopy images. , 2017, Acta biomaterialia.

[11]  Jernej Barbic,et al.  Deformable object animation using reduced optimal control , 2009, ACM Trans. Graph..

[12]  Jamshid Ghaboussi,et al.  Neural network constitutive model for rate-dependent materials , 2006 .

[13]  Mathieu Desbrun,et al.  Numerical coarsening of inhomogeneous elastic materials , 2009, SIGGRAPH 2009.

[14]  Wojciech Matusik,et al.  Data-driven finite elements for geometry and material design , 2015, ACM Trans. Graph..

[15]  James H. Garrett,et al.  Knowledge-Based Modeling of Material Behavior with Neural Networks , 1992 .

[16]  Ronald Fedkiw,et al.  Invertible finite elements for robust simulation of large deformation , 2004, SCA '04.

[17]  Hans-Peter Seidel,et al.  Animating deformable objects using sparse spacetime constraints , 2014, ACM Trans. Graph..

[18]  Jessica K. Hodgins,et al.  Estimating cloth simulation parameters from video , 2003, SCA '03.

[19]  Cosmin G. Petra,et al.  An Augmented Incomplete Factorization Approach for Computing the Schur Complement in Stochastic Optimization , 2014, SIAM J. Sci. Comput..

[20]  Michael A. Saunders,et al.  LSQR: An Algorithm for Sparse Linear Equations and Sparse Least Squares , 1982, TOMS.

[21]  M. Otaduy,et al.  Capture and modeling of non-linear heterogeneous soft tissue , 2009, ACM Trans. Graph..

[22]  Hans-Peter Seidel,et al.  Measurement‐Based Interactive Simulation of Viscoelastic Solids , 2004, Comput. Graph. Forum.

[23]  Steve Marschner,et al.  Data‐Driven Estimation of Cloth Simulation Models , 2012, Comput. Graph. Forum.

[24]  Jernej Barbic,et al.  FEM simulation of 3D deformable solids: a practitioner's guide to theory, discretization and model reduction , 2012, SIGGRAPH '12.

[25]  Miguel A. Otaduy,et al.  Modeling and Estimation of Energy‐Based Hyperelastic Objects , 2016, Comput. Graph. Forum.

[26]  Bin Wang,et al.  Deformation capture and modeling of soft objects , 2015, ACM Trans. Graph..

[27]  Drakos Stefanos,et al.  On Neural Network Constitutive Models for Geomaterials , 2015 .

[28]  Hongyi Xu,et al.  Nonlinear material design using principal stretches , 2015, ACM Trans. Graph..

[29]  Huamin Wang,et al.  Data-driven elastic models for cloth: modeling and measurement , 2011, ACM Trans. Graph..

[30]  Hongyi Xu,et al.  Example-based damping design , 2017, ACM Trans. Graph..

[31]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[32]  Wojciech Matusik,et al.  Dynamics-aware numerical coarsening for fabrication design , 2017, ACM Trans. Graph..

[33]  Dinesh K. Pai,et al.  Exponential Rosenbrock-Euler Integrators for Elastodynamic Simulation , 2018, IEEE Transactions on Visualization and Computer Graphics.

[34]  Dinesh K. Pai,et al.  Scanning physical interaction behavior of 3D objects , 2001, SIGGRAPH.

[35]  Matthias Teschner,et al.  Robust and Efficient Estimation of Elasticity Parameters using the linear Finite Element Method , 2007, SimVis.