Username: Password:
Menu 
VPA
Computer Vision And Pattern Analysis Laboratory Home Page  Home
People  People
Publications  Publications
Publications  Databases
Contact Information  Contact
Research
Supported Research Projects  Supported Research Projects
Research Activites  Research Activites
Research Groups
SPIS - Signal Processing and Information Systems Lab.SPIS - Signal Processing and Information Systems Lab.
Medical Vision and Analysis Group  Medical Research Activities
Biometrics Research Group  Biometrics Research Group
SPIS - Signal Processing and Information Systems Lab.MISAM - Machine Intelligence for Speech Audio and Multimedia.
Knowledge Base
  Paper Library
  VPA Lab Inventory
  Databases in VPALAB
  Recordings
Calendar
<<September 2017>>
Mo Tu We Th Fr Sa Su
1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30
Upcoming Events:
None

Template-based CTA X-ray angio rigid registration of coronary arteries in frequency domain with automatic X-Ray segmentation
Authors: Aksoy, Timur and Demirci, Stefanie and Degertekin, Muzaffer and Navab, Nassir and Unal, Gozde
Published in: Medical Physics, vol:40, number:10, October 2013
Publication year: 2013
Abstract: Purpose:

A key challenge for image guided coronary interventions is accurate and absolutely robust image registration bringing together preinterventional information extracted from a three-dimensional (3D) patient scan and live interventional image information. In this paper, the authors present a novel scheme for 3D to two-dimensional (2D) rigid registration of coronary arteries extracted from preoperative image scan (3D) and a single segmented intraoperative x-ray angio frame in frequency and spatial domains for real-time angiography interventions by C-arm fluoroscopy.

Methods:

Most existing rigid registration approaches require a close initialization due to the abundance of local minima and high complexity of search algorithms. The authors' method eliminates this requirement by transforming the projections into translation-invariant Fourier domain for estimating the 3D pose. For 3D rotation recovery, template Digitally Reconstructed Radiographs (DRR) as candidate poses of 3D vessels of segmented computed tomography angiography are produced by rotating the camera (image intensifier) around the DICOM angle values with a specific range as in C-arm setup. The authors have compared the 3D poses of template DRRs with the segmented x-ray after equalizing the scales in three domains, namely, Fourier magnitude, Fourier phase, and Fourier polar. The best rotation pose candidate was chosen by one of the highest similarity measures returned by the methods in these domains. It has been noted in literature that frequency domain methods are robust against noise and occlusion which was also validated by the authors' results. 3D translation of the volume was then recovered by distance-map based BFGS optimization well suited to convex structure of the authors' objective function without local minima due to distance maps. A novel automatic x-ray vessel segmentation was also performed in this study.

Results:

Final results were evaluated in 2D projection space for patient data; and with ground truth values and landmark distances for the images acquired with a solid phantom vessel. Results validate that rotation recovery in frequency domain is robust against differences in segmentations in two modalities. Distance-map translation is successful in aligning coronary trees with highest possible overlap.

Conclusions:

Numerical and qualitative results show that single view rigid alignment in projection space is successful. This work can be extended with multiple views to resolve depth ambiguity and with deformable registration to account for nonrigid motion in patient data.
 

Home Back