Model Based LV-Reconstruction in Bi-Plane X-Ray Angiography

Werner Backfrieder, Roland Swoboda, Martin Carpella, Clemens Steinwender, Christian Gabriel, Franz Leisch

Research output: Chapter in Book/Report/Conference proceedingsConference contributionpeer-review

6 Citations (Scopus)


Interventional x-ray angiography is state of the art in diagnosis and therapy of severe diseases of the cardio-vascular system. Diagnosis is based on contrast enhanced dynamic projection images of the left ventricle. A new model based algorithm for three dimensional reconstruction of the left ventricle from bi-planar angiograms was developed. Parametric super ellipses are deformed until their projection profiles optimally fit measured ventricular projections. Deformation is controlled by a simplex optimization procedure. A resulting optimized parameter set builds the initial guess for neighboring slices. A three dimensional surface model of the ventricle is built from stacked contours. The accuracy of the algorithm has been tested with mathematical phantom data and clinical data. Results show conformance with provided projection data and high convergence speed makes the algorithm useful for clinical application. Fully three dimensional reconstruction of the left ventricle has a high potential for improvements of clinical findings in interventional cardiology.

Original languageEnglish
Title of host publicationMedical Imaging 2005: Image Processing
PublisherSPIE Press
Number of pages9
Publication statusPublished - 2005
EventSPIE Medical Imaging 2005 - San Diego, United States
Duration: 12 Feb 200517 Feb 2005

Publication series

NameProgress in Biomedical Optics and Imaging - Proceedings of SPIE
ISSN (Print)1605-7422


ConferenceSPIE Medical Imaging 2005
Country/TerritoryUnited States
CitySan Diego


  • Cardiac angiography
  • Deformable models
  • Fully 3d reconstruction
  • Model based reconstruction
  • Myocardial infarction
  • Tomography


Dive into the research topics of 'Model Based LV-Reconstruction in Bi-Plane X-Ray Angiography'. Together they form a unique fingerprint.

Cite this