{
  "abstract": "Contemporary image generation methods commonly synthesize visual content through learned latent representations, iterative denoising, autoregressive prediction, or transformations of existing visual information. This paper proposes Adaptive Brickwork Image Generation (ABIG), a construction-oriented framework in which an image is progressively assembled from spatially organized visual units called bricks. Inspired by the principles of brick masonry, the framework represents image synthesis as a sequence of placement, alignment, subdivision, merging, replacement, and repair operations. Unlike a formulation based exclusively on fixed rectangular patches, ABIG permits variable-sized and potentially irregular visual units, explicit relationships between neighboring units, staggered boundary configurations, and reversible construction decisions.\n\nThe framework consists of a brick representation module, a construction-state encoder, a candidate placement generator, a structural constraint evaluator, a visual renderer, and a construction controller. The controller selects operations according to their expected contribution to visual quality, semantic consistency, structural coherence, and computational efficiency. Reinforcement learning is proposed as one possible mechanism for learning construction policies, while optimization-based and heuristic controllers provide simpler alternatives. A multi-objective formulation is introduced to balance image quality, construction cost, boundary artifacts, and repair complexity. This paper formalizes the proposed framework, presents a construction algorithm, specifies an experimental protocol, and discusses implementation challenges. Because the framework has not yet been empirically validated, its potential advantages and comparative performance remain hypotheses to be tested. The paper also distinguishes the proposal from existing patch-based synthesis and sequential constructive generation research, including prior work explicitly titled Brick-by-Brick.",
  "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/3g8gxt8js7ewm.bib",
  "cited_by": 0,
  "doi": "10.5281/zenodo.23555813",
  "fulltext_url": null,
  "id": "3g8gxt8js7ewm",
  "issns": [],
  "landing_url": "https://zenodo.org/records/23555813",
  "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/23555813/files/Adaptive%20Brickwork%20Image%20Generation.pdf/content",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-22",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/3g8gxt8js7ewm.ris",
  "sources": [
    {
      "first_seen": "2026-11-05T12:20:46Z",
      "native_id": "23555813",
      "source": "zenodo",
      "url": "https://zenodo.org/records/23555813"
    }
  ],
  "subjects": [],
  "title": "Adaptive Brickwork Image Generation: A Framework for Sequential Visual Construction with Dynamic Blocks, Structural Constraints, and Reinforcement Learning",
  "title_text": "Adaptive Brickwork Image Generation: A Framework for Sequential Visual Construction with Dynamic Blocks, Structural Constraints, and Reinforcement Learning",
  "type": "preprint",
  "updated_at": "2026-11-05T12:20:46Z",
  "url": "https://research.draconitic.com/paper/3g8gxt8js7ewm",
  "venue": "Zenodo",
  "versions": []
}