{
  "abstract": "Computed tomography (CT) and magnetic resonance (MR) imaging provide complementary information for the brain, but acquiring both modalities is not always possible. The cycle-consistent adversarial network (CycleGAN) enables unpaired CT-to-MR synthesis, but because the standard L1-norm cycle?consistency penalty grows linearly with the residual magnitude, the network tends to focus on large residuals. This paper introduces an L0-norm-based cycle?consistency term, in which the L0-norm is approximated by a smooth and bounded function. This approximation not only increases sensitivity to small residuals but also effectively suppresses the influence of large residuals. The proposed model and an otherwise identical L1-norm baseline were trained on 102 images and evaluated on 9 test images. The proposed model achieves a lower mean squared error (0.02746 versus 0.03464, a 20.7% reduction), a higher mean peak signal-to-noise ratio (15.80 versus 14.77 dB), and a higher mean structural similarity index (0.5689 versus 0.5407). It also has a lower maximum MSE and higher minimum PSNR and SSIM than the baseline. These results indicate that modifying the cycle-consistency penalty can improve reconstruction accuracy on this dataset.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2610.32491",
  "author_total": 4,
  "authors": [
    {
      "name": "Yunxiao Zhang"
    },
    {
      "name": "Shouqiang Du"
    },
    {
      "name": "Zexian Xu"
    },
    {
      "name": "Chuanlong Xu"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/fa708t3a165ef.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2610.32491",
  "fulltext_url": null,
  "id": "fa708t3a165ef",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2610.32491",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.CV"
    ],
    [
      "arxiv",
      "math.OC"
    ],
    [
      "openalex",
      "1707|Generative Adversarial Networks and Image Synthesis"
    ]
  ],
  "openalex": "W7223649965",
  "pdf_url": "https://arxiv.org/pdf/2610.32491",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-23",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/fa708t3a165ef.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:35:19Z",
      "native_id": "2610.32491",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2610.32491"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:math.OC",
      "subfield": "optimisation-control"
    }
  ],
  "title": "L0-Norm-Based CycleGAN for Unpaired Brain CT-to-MR Synthesis",
  "title_text": "L0-Norm-Based CycleGAN for Unpaired Brain CT-to-MR Synthesis",
  "type": "preprint",
  "updated_at": "2026-10-30T12:35:19Z",
  "url": "https://research.draconitic.com/paper/fa708t3a165ef",
  "venue": "arXiv",
  "versions": []
}