FracGen: Physics-Informed Fracture-Aware Video Generation from a Single Image

Summary

FracGen is a fracture-aware video generation model that produces plausible, controllable fracture dynamics from a single image of an intact object, conditioned on physics signals. To train it, the authors built FracSim, a fracture-aware simulation framework that augments material point method (MPM) simulation with a continuum damage model, generating paired fracture videos and dense pixel-aligned physical fields at no extra cost beyond standard rendering. FracGen exploits these maps in two ways: it is trained to jointly predict physical fields alongside RGB video, encouraging the model to capture physical state rather than surface appearance, and it is supervised with physics-informed losses enforcing consistency among predicted maps. The model captures material-specific fracture behavior without costly test-time simulation or per-scene tuning, offering fine-grained control over where objects tear, how fast cracks propagate, and how much objects deform before breaking. The authors also introduce a benchmark for evaluating the physical plausibility of generated fracture videos and show experimentally that FracGen outperforms existing video generation baselines in both physical and visual fidelity.