Chapter

Medical Image Computing and Computer-Assisted Intervention - MICCAI 2003

Volume 2879 of the series Lecture Notes in Computer Science pp 788-795

Computing 3D Non-rigid Brain Registration Using Extended Robust Point Matching for Composite Multisubject fMRI Analysis

  • Xenophon PapademetrisAffiliated withChild Study Center, Yale University New Haven
  • , Andrea P. JackowskiAffiliated withChild Study Center, Yale University New Haven
  • , Robert T. SchultzAffiliated withChild Study Center, Yale University New Haven
  • , Lawrence H. StaibAffiliated withDepartments of Elec. Engineering, Yale University New HavenDiag. Radiology, Yale University New Haven
  • , James S. DuncanAffiliated withDepartments of Elec. Engineering, Yale University New HavenDiag. Radiology, Yale University New Haven

* Final gross prices may vary according to local VAT.

Get Access

Abstract

In this paper we present extensions to the Robust Point Matching framework to improve its ability to handle larger point sets with greater computational efficiency. While in the past this methodology has only been used to register either two-dimensional or small synthetic three-dimensional data-sets we demonstrate its first successful application on large real 3D data-sets. We apply this methodology to the problem of forming composite activation maps from functional magnetic resonance images. In particular we demonstrate the superior performance of this algorithm to a pure intensity-based registration in the specific area of the fusiform gyrus. We also demonstrate that the robustness of this method can be useful in the case where part of the brain is missing as a result of incorrect image slice specification.