Characterization of 3-D coronary tree motion from MSCT angiography

  • Authors:

  • Guanyu Yang;Jian Zhou;Dominique Boulmier;Marie-Paule Garcia;Limin Luo;Christine Toumoulin

  • Affiliations:

  • Leiden Univ. Medical Center, The Netherlands and Southeast Univ., China and Lab. Traitement du Signal et de l'Image, Inst. Natl. de la Sante et de la Rec. Medicale, Centre de Rec. en Inf. Biomed. ...;Dept. of Biomedical Eng., Univ. of California, Davis and Lab. of Image Sci. and Techn., Southeast Univ., China and Inst. National de la Sante et de la Rec. Medicale and Centre de Rec. en Inf. Biom ...;Laboratoire Traitement du Signal et de l'Image, Institut National de la Santé et de la Recherche Médicale U642, Université de Rennes 1, Rennes, France and Département de Cardio ...;Laboratoire Traitement du Signal et de l'Image, Institut National de la Santé et de la Recherche Médicale U642 and Centre de Recherche en Information Biomédicale Sino-français, ...;Laboratory of Image Science and Technology, Southeast University, Nanjing, China Centre de Recherche en Information Biomédicale Sino-français, Université de Rennes 1, Rennes, France;Laboratoire Traitement du Signal et de l'Image, Institut National de la Santé et de la Recherche Médicale U642 and Centre de Recherche en Information Biomédicale Sino-français, ...

  • Venue:
  • IEEE Transactions on Information Technology in Biomedicine
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper describes a method for the characterization of coronary artery motion using multislice computed tomography (MSCT) volume sequences. Coronary trees are first extracted by a spatial vessel tracking method in each volume of MSCT sequence. A point-based matching algorithm, with feature landmarks constraint, is then applied to match the 3-D extracted centerlines between two consecutive instants over a complete cardiac cycle. The transformation functions and correspondence matrices are estimated simultaneously, and allow deformable fitting of the vessels over the volume series. Either point-based or branch-based motion features can be derived. Experiments have been conducted in order to evaluate the performance of the method with amatching error analysis.