Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Surface-Based Registration with a Particle Filter and Randy E. Ellis1
 

Summary: Surface-Based Registration with a Particle Filter
Burton Ma1
and Randy E. Ellis1
School of Computing, Queen's University at Kingston, Canada K7L 3N6
Abstract. We propose the use of a particle filter as a solution to the rigid shape-
based registration problem commonly found in computer-assisted surgery. This
approach is especially useful where there are only a few registration points cor-
responding to only a fraction of the surface model. Tests performed on patient
models, with registration points collected during surgery, suggest that particle
filters perform well and also provide novel quality measures to the surgeon.
1 Introduction
Preoperative 3D medical images, such as CT and MRI scans, can be registered in-
traoperatively to a patient's anatomy by estimating a transformation from surfaces in
image coordinates to anatomical points in patient coordinates for use in image-guided
surgery. Two notable limitations of current algorithms are: (a) most algorithms are non-
incremental and process an additional anatomical point by reconsidering the entire set
of anatomical points gathered during surgery; and (b) most algorithms report errors such
as root-mean-square (RMS) or target registration (TRE) but do not report the probable
distribution of errors. Particle filters offer an incremental method for computing proba-
bility distributions of the rotational and translational components of a rigid registration.

  

Source: Abolmaesumi, Purang - School of Computing, Queen's University (Kingston)

 

Collections: Computer Technologies and Information Sciences