Incorporating independent component analysis to Q-ball imaging for diffusion orientation distribution reconstruction

M Jing; T M McGinnity; S Coleman; H Zhang

doi:10.1109/IEMBS.2010.5626529

Incorporating independent component analysis to Q-ball imaging for diffusion orientation distribution reconstruction

Annu Int Conf IEEE Eng Med Biol Soc. 2010:2010:2706-9. doi: 10.1109/IEMBS.2010.5626529.

Authors

M Jing¹, T M McGinnity, S Coleman, H Zhang

Affiliation

¹ Intelligent Systems Research Centre, School of Computing and Intelligent Systems, University of Ulster, Magee Campus, UK. m.jing@ulster.ac.uk

PMID: 21096204
DOI: 10.1109/IEMBS.2010.5626529

Abstract

In this paper, we investigate the incorporation of independent component analysis (ICA) with Q-ball imaging (QBI) to extract information on the diffusion orientation distribution function (ODF) from an inner voxel. In our approach, the ICA algorithm is applied to a mixture of ODFs which are constructed based on the analytical QBI solution. The numerical simulation results demonstrate that the proposed ICA framework can not only successfully separate the diffusion ODF from the noisy diffusion data, but also achieves better performance compared with a QBI solution when the data has a low signal to noise ratio (SNR).

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Algorithms
Brain / pathology*
Computer Simulation
Diffusion
Humans
Models, Neurological
Models, Statistical
Models, Theoretical
Neurons / pathology*
Principal Component Analysis
Signal Processing, Computer-Assisted