GausSim: Foreseeing Reality by Gaussian Simulator for Elastic Objects

TL;DR

GausSim is a neural network-based Gaussian kernel simulator that accurately models elastic object dynamics using hierarchical structures and physics constraints, validated on a new multi-view deformation dataset.

Contribution

The paper introduces GausSim, a novel hierarchical neural simulator for elastic objects that incorporates physics constraints for realistic and efficient simulations.

Findings

GausSim outperforms existing physics-based baselines in accuracy.

The hierarchical structure reduces computational costs significantly.

GausSim effectively captures complex elastic deformations.

Abstract

We introduce GausSim, a novel neural network-based simulator designed to capture the dynamic behaviors of real-world elastic objects represented through Gaussian kernels. We leverage continuum mechanics and treat each kernel as a Center of Mass System (CMS) that represents continuous piece of matter, accounting for realistic deformations without idealized assumptions. To improve computational efficiency and fidelity, we employ a hierarchical structure that further organizes kernels into CMSs with explicit formulations, enabling a coarse-to-fine simulation approach. This structure significantly reduces computational overhead while preserving detailed dynamics. In addition, GausSim incorporates explicit physics constraints, such as mass and momentum conservation, ensuring interpretable results and robust, physically plausible simulations. To validate our approach, we present a new dataset, READY, containing multi-view videos of real-world elastic deformations. Experimental results demonstrate that GausSim achieves superior performance compared to existing physics-driven baselines, offering a practical and accurate solution for simulating complex dynamic behaviors. Code and model are available at our project page: https://www.mmlab-ntu.com/project/gausim/index.html .