Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 15 min ago
PreprintarXiv licencearXiv

Tri-Directional Sub-Voxel Displacement and Ipsilateral Territory Decoupling for 3D Bilateral Coronary Artery Tree Reconstruction in Cardiac CT Angiography

  • Zeyu Jia
  • Xin Ming
Open at sourcePDFDOICite

Abstract

Three-dimensional (3D) reconstruction of bilateral coronary trees (LCA and RCA) from cardiac computed tomography angiography (CCTA) is limited by 643 quantization, territory ambiguity, and sinotubular junction (STJ) ostium drift. We present a canonical SO(3) (ZIS, YPA, XLR) multi-scale representation and Riemannian arborescence framework. First, we map 3D CCTA volumes, centerlines, and thick-stem ostia into (ZIS, YPA, XLR) space, extending a 3D pyramid and tubular decoder with a 1283 → 643 boundary skip, a stop-gradient 1283 residual branch, a sub-voxel displacement head Δû(x), a dual-channel ipsilateral head (PLCA, PRCA), and a 323 ostium detector. Second, all $327,385$ active parameters ($398,843$ total) are trained in one 30-epoch AdamW loop (ε= 10−8) under decoupled L2-SP and aortic-wall consistency. Third, our Riemannian solver couples cylindrical Valsalva-STJ calibration, 1283 orifice contact rooting, contralateral inhibition, centripetal pruning, and 4-stage curvature-adaptive normal-plane pullback. On 14 unseen ImageCAS development volumes (28 trees; cases 201--250 isolated), the framework achieves 2.1652 ± 0.4686 mm CDCL (CDCLarc1mm = 2.2697 ± 0.4863 mm, p ≤ 0.00024; 57.40% / 78.20% F1@1/2mm; 4.6880 / 4.9298 mm ostia). Zero-shot evaluation on 50 unseen volumes (Cases 251--300, 100 trees) yields 2.9389 ± 1.2785 mm CDCL (p ≤ 7.24 × 10−7; ostia 5.1815 / 7.5442 mm), with 2.7696 ± 1.1919 mm pooled N=64 CDCL and 128/128 trees (β0 = 1, β1 = 0, 0.0000 mm detachment).

Related

  1. Quantification of plaque and detection of high-risk plaque characteristics based on coronary computed tomography angiography: comparison between superresolution deep learning reconstruction and hybrid iterative reconstruction
    Limiao Zou, Keting Xu, Xiaohuan Liu +26
    Quantitative imaging in medicine and surgeryPeer-reviewedCC BY-NC-NDFull textMedical imaging & radiologyEurope PMC
  2. Implant-Based Versus Autologous Reconstruction After Salvage Mastectomy for Ipsilateral Breast Tumor Recurrence in the Previously Irradiated Breast: A Consecutive Institutional Series and a Proposed Autologous-First Algorithm
    Jongmoo Park, Woosung Jang, Myeong Jae Kang +12
    CancersPeer-reviewedCC BY 4.0Full textSurgery & anaesthesiaCrossrefEurope PMC
  3. Cardiac-Specific AI Reconstruction Versus Hybrid Iterative Reconstruction for Emergency Coronary CT Angiography: A Paired Single-Center Study Including Triple-Rule-Out CT
    Sung Min Ko, Hyunjung Kim, Sung-Jin Cha +1
    Journal of Clinical MedicinePeer-reviewedCC BY 4.0Full textMedical imaging & radiologyCrossrefEurope PMC
  4. Treatment strategies for unilateral papillary thyroid carcinoma with ipsilateral cervical lymph node metastasis: a narrative review
    Gland SurgeryPeer-reviewedCC BY-NC-NDFull textGeneral medicineEurope PMC
  5. Repeat breast-conserving surgery for ipsilateral breast tumour recurrence: towards a paradigm shift—a narrative review
    Gland SurgeryPeer-reviewedCC BY-NC-NDFull textGeneral medicineEurope PMC

Cite

BibTeXRIS
BibTeX
@misc{jia2026tridirectional,
  title = {{Tri-Directional Sub-Voxel Displacement and Ipsilateral Territory Decoupling for 3D Bilateral Coronary Artery Tree Reconstruction in Cardiac CT Angiography}},
  author = {Zeyu Jia and Xin Ming},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.05900},
  eprint = {2610.05900},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.05900},
}
RIS
TY  - UNPB
TI  - Tri-Directional Sub-Voxel Displacement and Ipsilateral Territory Decoupling for 3D Bilateral Coronary Artery Tree Reconstruction in Cardiac CT Angiography
AU  - Zeyu Jia
AU  - Xin Ming
PY  - 2026
DA  - 2026/10/27
PB  - arXiv
DO  - 10.48550/arxiv.2610.05900
UR  - https://arxiv.org/abs/2610.05900
AB  - Three-dimensional (3D) reconstruction of bilateral coronary trees (LCA and RCA) from cardiac computed tomography angiography (CCTA) is limited by 64^3 quantization, territory ambiguity, and sinotubular junction (STJ) ostium drift. We present a canonical SO(3) (Z_IS, Y_PA, X_LR) multi-scale representation and Riemannian arborescence framework. First, we map 3D CCTA volumes, centerlines, and thick-stem ostia into (Z_IS, Y_PA, X_LR) space, extending a 3D pyramid and tubular decoder with a 128^3 → 64^3 boundary skip, a stop-gradient 128^3 residual branch, a sub-voxel displacement head Δû(x), a dual-channel ipsilateral head (P_LCA, P_RCA), and a 32^3 ostium detector. Second, all $327,385$ active parameters ($398,843$ total) are trained in one 30-epoch AdamW loop (ε= 10^−8) under decoupled L2-SP and aortic-wall consistency. Third, our Riemannian solver couples cylindrical Valsalva-STJ calibration, 128^3 orifice contact rooting, contralateral inhibition, centripetal pruning, and 4-stage curvature-adaptive normal-plane pullback. On 14 unseen ImageCAS development volumes (28 trees; cases 201--250 isolated), the framework achieves 2.1652 ± 0.4686 mm CD_CL (CD_CL^arc1mm = 2.2697 ± 0.4863 mm, p ≤ 0.00024; 57.40% / 78.20% F1@1/2mm; 4.6880 / 4.9298 mm ostia). Zero-shot evaluation on 50 unseen volumes (Cases 251--300, 100 trees) yields 2.9389 ± 1.2785 mm CD_CL (p ≤ 7.24 × 10^−7; ostia 5.1815 / 7.5442 mm), with 2.7696 ± 1.1919 mm pooled N=64 CD_CL and 128/128 trees (β_0 = 1, β_1 = 0, 0.0000 mm detachment).
ER  -