Feng Shi1, Yong Fan1, Jie-Zhi Cheng1, Lawrence L. Wald2,3, Guido Gerig4, Weili Lin1, Dinggang Shen1
1Department of Radiology and BRIC, University of North Carolina, Chapel Hill, NC, USA; 2A.A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, USA; 3Harvard-MIT Division of Health Sciences and Technology, USA; 4Scientific Computing and Imaging Institute, University of Utah, USA
Tissue segmentation in neonatal brain MR image is a great challenge due to small brain size and low tissue contrast. To address this problem, a dedicated phased array neonatal head coil was utilized for brain imaging. 8 images of each modality (T1 or T2) were acquired with their corresponding coil sensitivity profiles. High quality T1 and T2 images were reconstructed with high tissue contrast and clear cortical convolutions. A novel hybrid atlas based MRI segmentation algorithm was developed, by combining the fuzzy population atlas and traced cortical convolutions. Results illustrate segmentation accuracy and potential for cortical surface analysis.