Robust Multi-site MR Data Processing: Iterative Optimization of Bias Correction, Tissue Classification, and Registration

Frontiers in Neuroinformatics

Eun Young eKim¹  Hans J Johnson¹ 

[1] University of Iowa;

关键词: Registration;    segmentation;    multi-center studies;    tissue classification;    Inhomogeneity Correction;   

DOI  :  10.3389/fninf.2013.00029

来源: DOAJ

【 摘 要 】

A robust multi-modal tool, for automated registration,bias correction, and tissue classification, has beenimplemented for large-scale heterogeneousmulti-site longitudinal MR data analysis.This work focused on improving thean iterative optimization frameworkbetween bias-correction, registration, and tissueclassification inspired from previous work.The primary contributions are robustness improvementsfrom incorporation of following four elements:1) utilize multi-modal and repeated scans,2) incorporate high-deformable registration,3) use extended set of tissue definitions, and4) use of multi-modal aware intensity-context priors.The benefits of these enhancements were investigatedby a series of experiments with both simulatedbrain data set (BrainWeb) and by applyingto highly-heterogeneous data from a 32 site imaging studywith quality assessments through the expert visual inspection.The implementation of this tool is tailored for,but not limited to, large-scale data processingwith great data variation with a flexible interface.In this paper, we describe enhancements to a jointregistration, bias correction, and the tissue classification,that improve the generalizability and robustness forprocessing multi-modal longitudinal MR scanscollected at multi-sites.The tool was evaluated by using both simulated andsimulated and human subject MRI images.With these enhancements,the results showed dramatic improvementin accuracy and robustnessfor large-scale heterogeneous MRI processing.

【 授权许可】

Unknown