{
  "abstract": "Contemporary image generation methods commonly synthesize images through learned latent representations, iterative denoising, autoregressive prediction, or combinations of these mechanisms. An alternative perspective is to organize image synthesis around the spatial construction of visual elements that collectively cover a target image domain. This paper proposes Tessellation-Based Image Generation (TBIG), a conceptual and algorithmic framework in which an image is constructed through the placement, adaptation, and coordination of geometric or semantic visual units. Unlike approaches that use tiling primarily to produce seamless textures or repeated patterns, TBIG formulates spatial tiling as an explicit organizational mechanism for image synthesis. The framework represents an image as a collection of cells with geometric boundaries, appearance attributes, and neighborhood relationships. A generation controller selects and modifies cells according to geometric feasibility, boundary compatibility, semantic consistency, visual quality, and computational cost. The framework supports regular and irregular tessellations, hierarchical subdivision, adaptive cell deformation, and optional reinforcement-learning-based action selection. A unified objective function is introduced to balance coverage, overlap, boundary consistency, semantic fidelity, and generation cost. An iterative algorithm and an experimental protocol are presented to make the proposal implementable and empirically testable. The paper does not claim that tessellation, image tiling, patch-based synthesis, or adaptive geometric subdivision is new in isolation. Instead, it establishes a research framework for investigating how these mechanisms can be integrated into a general image-generation process. The effectiveness, scalability, and potential advantages of the proposed framework remain to be established experimentally.",
  "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/6q524k6w524z8.bib",
  "cited_by": 0,
  "doi": "10.5281/zenodo.23555823",
  "fulltext_url": null,
  "id": "6q524k6w524z8",
  "issns": [],
  "landing_url": "https://zenodo.org/records/23555823",
  "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/23555823/files/Tessellation-Based%20Image%20Generation.pdf/content",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-22",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/6q524k6w524z8.ris",
  "sources": [
    {
      "first_seen": "2026-11-05T12:20:46Z",
      "native_id": "23555823",
      "source": "zenodo",
      "url": "https://zenodo.org/records/23555823"
    }
  ],
  "subjects": [],
  "title": "Tessellation-Based Image Generation: A Framework for Constraint-Guided Visual Synthesis Through Dynamic Spatial Tiling",
  "title_text": "Tessellation-Based Image Generation: A Framework for Constraint-Guided Visual Synthesis Through Dynamic Spatial Tiling",
  "type": "preprint",
  "updated_at": "2026-11-05T12:20:46Z",
  "url": "https://research.draconitic.com/paper/6q524k6w524z8",
  "venue": "Zenodo",
  "versions": []
}