Maria Murgasova1, Latha Srinivasan2, Ioannis S. Gousias2, Paul Aljabar1, Joseph V. Hajnal2, A David Edwards2, Daniel Rueckert1
1Computing, Imperial College London, London, United Kingdom; 2Imaging Sciences, Imperial College London
Probabilistic atlases have been established in the literature as a standard tool for enhancing the intensity-based classification of brain MRI. The rapidly growing neonatal brain requires an age-specific spatial probabilistic atlas to guide the segmentation process. In this paper we describe a method for dynamically creating a probabilistic atlas for any chosen stage of neonatal brain development. We present an atlas created from the segmentations of 153 subjects providing prior tissue probability maps for six structures - cortex, white matter, subcortical gray matter, brainstem and cerebellum, for ages of 29 to 44 weeks of gestation.