An algorithm and quantitative evaluation framework for registration of multi-modal brain MRI
Omar Ocegueda 1 , Eleftherios Garyfallidis 2 , Maxime Descoteaux 2 , and Mariano Rivera 1
Computer Science Department, Centro de
Investigacin en Matemticas, Guanajuato, Guanajuato,
Connectivity Imaging Lab (SCIL), Computer Science
department, Universit de Sherbrooke, Sherbrooke,
We present a new algorithm for multi-modal symmetric
diffeomorphic image registration and propose a
validation protocol, based on existing manually
annotated datasets, to quantitatively evaluate
multi-modal image registration algorithms. Our
validation protocol reveals that the Cross Correlation
(CC) metric may be severely affected in the multi-modal
case even though it has proven to be one of the most
robust and accurate metrics for mono-modal registration.
Our algorithm is based on the Symmetric Normalization
(SyN) algorithm. It compares favourably with SyN with CC
(in the multi-modal case) and is very competitive with
SyN with Mutual Information.
This abstract and the presentation materials are available to members only;
a login is required.