Ara K. Yeramian1, Martin Haas1, Maxim Zaitsev1, Chris A. Cocosco1, Jan G. Korvink2, 3, Jrgen Hennig1
Specific Absorption Rate (SAR) estimation in parallel MR excitation is achieved through EM simulations of RF parallel excitation coils and 3D virtual models of human body parts in interest. To estimate SAR in the human head, we present a segmentation scheme to efficiently generate patient-specific head models, based on MP RAGE and T1-UTE acquired datasets. The multi-stage segmentation scheme comprises morphological and set operations, 2D and 3D component connectivity check, ITK-based hole filling, and two user-interactive stages. Segmentation output are nine head tissues: brain GM, WM, CSF, cortical bone, skull, skin, air cavities, the eye region, and soft tissue.