Skeleton-based Region Competition for Automated Gray Matter and White Matter Segmentation of Human Brain MR Images

摘要

Image segmentation is an essential process for quantitative analysis. Segmentation of brain tissues in magneticresonance (MR) images is very important for understanding the structural-functional relationship for variouspathological conditions, such as dementia vs. normal brain aging. Different brain regions are responsible for certainfunctions and may have specific implication for diagnosis. Segmentation may facilitate the analysis of different brainregions to aid in early diagnosis. Region competition has been recently proposed as an effective method for imagesegmentation by minimizing a generalized Bayes/MDL criterion. However, it is sensitive to initial conditions – the“seeds”, therefore an optimal choice of “seeds” is necessary for accurate segmentation. In this paper, we present a newskeleton-based region competition algorithm for automated gray and white matter segmentation. Skeletons can beconsidered as good “seed regions” since they provide the morphological a priori information, thus guarantee a correctinitial condition. Intensity gradient information is also added to the global energy function to achieve a precise boundarylocalization. This algorithm was applied to perform gray and white matter segmentation using simulated MRI imagesfrom a realistic digital brain phantom. Nine different brain regions were manually outlined for evaluation of theperformance in these separate regions. The results were compared to the gold-standard measure to calculate the truepositive and true negative percentages. In general, this method worked well with a 96% accuracy, although theperformance varied in different regions. We conclude that the skeleton-based region competition is an effective methodfor gray and white matter segmentation.