Meeting Banner
Abstract #1240

Automatic Segmentation of Brain Tissues for Newborn MRI in Longitudinal Study

Feng Shi1, Yong Fan1, Songyuan Tang1, Katie Cleary2, Martin Styner2, John Gilmore2, Weili Lin1, Dinggang Shen1

1Department of Radiology and BRIC, University of North Carolina, Chapel Hill, NC, USA; 2Department of Psychiatry, University of North Carolina, Chapel Hill, NC, USA

Tissue segmentation in newborn MRI is challenging compared to that of late time brain images (even at one- or two-year-old), due to low signal contrast and different tissue development in neonatal stage. Considering the human cortical convolution pattern is generated during gestation and remains very similar in the whole life and segmentation on later time image is relatively easy, a dedicated longitudinal newborn MRI segmentation framework is proposed by taking the later time image as subject-specific atlases to guide the segmentation of the newborn brain. Segmentation accuracy is validated qualitatively by visual inspection and quantitatively by comparison with manual segmentations.