{
  "abstract": "Contemporary image generation methods commonly construct images through learned latent representations, iterative denoising, autoregressive prediction, or direct optimization of image parameters. This paper proposes Foam-Based Image Generation (FBIG), a framework in which an image is constructed through the dynamic evolution of a variable population of foam cells. Inspired by the formation and organization of physical foams, the framework represents visual content using interacting cells that can nucleate, expand, compete for spatial occupancy, establish shared boundaries, merge, rupture, and disappear. Unlike methods that use bubbles solely as visual effects, FBIG treats the foam-cell population and its evolving interfaces as computational variables that participate directly in image construction. A differentiable rendering formulation maps cell states and boundary configurations to an image, while target-dependent feedback guides the evolution of cell positions, sizes, colors, and topological relationships. The proposed objective combines pixel reconstruction, perceptual similarity, boundary regularization, spatial competition, and structural complexity. A discrete-action formulation is also introduced to support reinforcement-learning-based control of cell-level operations. The framework is intended to investigate whether dynamic foam organization can provide a useful alternative representation for image synthesis, reconstruction, and adaptive visual refinement. This paper establishes the conceptual framework, mathematical formulation, algorithmic procedure, and experimental protocol. It does not claim empirical superiority or validated performance; these remain questions for future experiments. The relationship between the proposed framework and existing work on foam simulation, differentiable Voronoi representations, and neural cellular automata is explicitly discussed.",
  "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/cmmfzvfz0d6nv.bib",
  "cited_by": 0,
  "doi": "10.5281/zenodo.23556032",
  "fulltext_url": null,
  "id": "cmmfzvfz0d6nv",
  "issns": [],
  "landing_url": "https://zenodo.org/records/23556032",
  "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/23556032/files/Foam-Based%20Image%20Generation.pdf/content",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-22",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/cmmfzvfz0d6nv.ris",
  "sources": [
    {
      "first_seen": "2026-11-05T12:20:44Z",
      "native_id": "23556032",
      "source": "zenodo",
      "url": "https://zenodo.org/records/23556032"
    }
  ],
  "subjects": [],
  "title": "Foam-Based Image Generation: A Dynamic Framework for Bubble Formation, Spatial Competition, and Boundary-Guided Visual Synthesis",
  "title_text": "Foam-Based Image Generation: A Dynamic Framework for Bubble Formation, Spatial Competition, and Boundary-Guided Visual Synthesis",
  "type": "preprint",
  "updated_at": "2026-11-05T12:20:44Z",
  "url": "https://research.draconitic.com/paper/cmmfzvfz0d6nv",
  "venue": "Zenodo",
  "versions": []
}