Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

FreeGave: 3D Physics Learning from Dynamic Videos by Gaussian Velocity

About

In this paper, we aim to model 3D scene geometry, appearance, and the underlying physics purely from multi-view videos. By applying various governing PDEs as PINN losses or incorporating physics simulation into neural networks, existing works often fail to learn complex physical motions at boundaries or require object priors such as masks or types. In this paper, we propose FreeGave to learn the physics of complex dynamic 3D scenes without needing any object priors. The key to our approach is to introduce a physics code followed by a carefully designed divergence-free module for estimating a per-Gaussian velocity field, without relying on the inefficient PINN losses. Extensive experiments on three public datasets and a newly collected challenging real-world dataset demonstrate the superior performance of our method for future frame extrapolation and motion segmentation. Most notably, our investigation into the learned physics codes reveals that they truly learn meaningful 3D physical motion patterns in the absence of any human labels in training.

Jinxi Li, Ziyang Song, Siyuan Zhou, Bo Yang• 2025

Related benchmarks

TaskDatasetResultRank
Future frame extrapolationDynamic Indoor Scene Dataset
PSNR35.019
24
Future frame extrapolationDynamic Object Dataset
PSNR31.987
22
Novel view interpolationDynamic Indoor Scene Dataset
PSNR32.287
22
Novel view interpolationDynamic Object Dataset
PSNR39.393
20
Future frame extrapolationNVIDIA Dynamic Scene Truck
PSNR29.954
12
Novel view interpolationNVIDIA Dynamic Scene Truck
PSNR28.584
12
Future frame extrapolationNVIDIA Dynamic Scene Skating
PSNR28.391
12
Novel view interpolationNVIDIA Dynamic Scene Skating
PSNR26.589
12
Future frame extrapolationDynamic Multipart (test)
PSNR34.667
9
Unsupervised Object Segmentationsynthetic indoor scene dataset
AP99.75
7
Showing 10 of 13 rows

Other info

Follow for update