{
  "abstract": "Three-dimensional generation methods commonly synthesize shapes through implicit fields, point sets, volumetric representations, or collections of Gaussian primitives. Although these representations can describe complex geometry, generating a fractured object introduces an additional challenge: the final geometry is the result of a sequence of structural events that create cracks, divide material regions, expose internal surfaces, and produce a variable number of fragments. This paper proposes Fracture-Event-Driven 3D Generation (FED3D), a candidate framework that represents three-dimensional fracture as a structured generative process jointly involving fracture-event sequences, fragment partitions, and fracture-surface geometry. Given an intact object, material descriptors, initial defects, and optional loading conditions, the framework generates a fracture-event graph whose nodes represent geometric states and whose edges encode crack initiation, propagation, branching, intersection, and separation. A geometry decoder then constructs a variable-cardinality set of fragments while enforcing partition consistency, interface correspondence, and material-boundary constraints. A probabilistic event model represents alternative plausible fracture outcomes, while optional mechanics-based losses encourage compatibility with energy, loading, and material constraints. Unlike a formulation that predicts only a final fracture pattern, FED3D explicitly treats the causal organization of fracture events as a generative variable. The proposed framework is intended for rock fragmentation, shattered glass, impact damage, geological structures, and topology-changing three-dimensional content creation. This paper defines the architecture, mathematical formulation, training objectives, and experimental protocol. Its effectiveness and novelty remain hypotheses requiring implementation and empirical validation.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": null,
  "author_total": 1,
  "authors": [
    {
      "name": "Jincheng Zhang",
      "orcid": "0009-0005-1860-0009"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/6asg7cwdjvz58.bib",
  "cited_by": 0,
  "doi": "10.5281/zenodo.23636172",
  "fulltext_url": null,
  "id": "6asg7cwdjvz58",
  "issns": [],
  "landing_url": "https://zenodo.org/records/23636172",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [],
  "openalex": null,
  "pdf_url": "https://zenodo.org/api/records/23636172/files/Fracture-Event-Driven%203D%20Generation.pdf/content",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-25",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/6asg7cwdjvz58.ris",
  "sources": [
    {
      "first_seen": "2026-11-08T12:21:22Z",
      "native_id": "23636172",
      "source": "zenodo",
      "url": "https://zenodo.org/records/23636172"
    }
  ],
  "subjects": [],
  "title": "Fracture-Event-Driven 3D Generation: A Generative Framework for Crack Propagation, Fragmentation, and Topology-Changing Shape Synthesis",
  "title_text": "Fracture-Event-Driven 3D Generation: A Generative Framework for Crack Propagation, Fragmentation, and Topology-Changing Shape Synthesis",
  "type": "preprint",
  "updated_at": "2026-11-08T12:21:22Z",
  "url": "https://research.draconitic.com/paper/6asg7cwdjvz58",
  "venue": "Zenodo",
  "versions": []
}