{
  "abstract": "Three-dimensional generative models have achieved substantial progress through diffusion models, implicit neural representations, point-cloud generation, and Gaussian-based scene representations. However, these approaches generally construct geometry through learned distributions, explicit geometric operations, or combinations of learned generation and physical optimization. This paper proposes Surface-Tension-Driven 3D Generative Modeling (ST3D-GM), a conceptual framework in which surface-tension-inspired energy minimization becomes a central mechanism for generating three-dimensional shapes. Rather than treating surface tension exclusively as a physical property to be simulated or a constraint applied after generation, ST3D-GM uses evolving interfaces as computational entities that compete, deform, merge, separate, and reorganize under an objective combining interfacial energy, geometric constraints, and target-shape guidance. The framework represents candidate shapes using a differentiable three-dimensional interface field and iteratively updates the field to produce geometries that balance surface regularity, volume preservation, interface interactions, and target compatibility. A multi-interface extension introduces competition between candidate geometric components, while a generative prior can preserve semantic plausibility during energy-driven evolution. We formulate the energy functional, describe the optimization procedure, present a modular algorithm, and propose an experimental protocol for evaluating geometric quality, target fidelity, energy reduction, and computational efficiency. The proposed framework is a research hypothesis rather than a validated generative model; its novelty and effectiveness require systematic comparison with existing physical simulation and physics-guided 3D generation methods.",
  "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/5n7p00tdmtrjc.bib",
  "cited_by": 0,
  "doi": "10.5281/zenodo.23635978",
  "fulltext_url": null,
  "id": "5n7p00tdmtrjc",
  "issns": [],
  "landing_url": "https://zenodo.org/records/23635978",
  "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/23635978/files/Surface-Tension-Driven%203D%20Generative%20Modeling.pdf/content",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-25",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/5n7p00tdmtrjc.ris",
  "sources": [
    {
      "first_seen": "2026-11-08T12:21:50Z",
      "native_id": "23635978",
      "source": "zenodo",
      "url": "https://zenodo.org/records/23635978"
    }
  ],
  "subjects": [],
  "title": "Surface-Tension-Driven 3D Generative Modeling: A Framework for Energy-Minimizing Interface Competition and Dynamic Geometric Formation",
  "title_text": "Surface-Tension-Driven 3D Generative Modeling: A Framework for Energy-Minimizing Interface Competition and Dynamic Geometric Formation",
  "type": "preprint",
  "updated_at": "2026-11-08T12:21:50Z",
  "url": "https://research.draconitic.com/paper/5n7p00tdmtrjc",
  "venue": "Zenodo",
  "versions": []
}