Reproduce “Breaking Bad: A Dataset for Geometric Fracture and Reassembly” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Silvia Sellán, Yun-Chun Chen, Ziyi Wu, Animesh Garg, Alec Jacobson

Breaking Bad creates a benchmark for reassembling fractured 3D objects from geometric pieces rather than semantic parts. It precomputes fracture modes and projects varied impacts onto them, generating 1,047,400 patterns from 10,474 shapes. On the everyday-object subset, DGL achieves 31.0% part accuracy: the category-averaged share of pieces with Chamfer distance below 0.01. This restricted test leaves high-piece reassembly difficult.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.