dHCP Brain Atlas 2017 (pairwise)
Neonatal brain atlas constructed from 275 term-born subjects of the dHCP cohort in 2017 using the "pairwise" approach with SVFFD transformation model, an inverse-consistent symmetric registration energy, and Log-Euclidean mean deformations. This version of the atlas construction workflow has been used to obtain the results for Schuh (2017) (PhD thesis, ch. 6.2).
The initial workflow steps implement the "Unbiased global normalization" as described in chapter 6.1 of Schuh (2017). The thus obtained affine output transformations were also used for the non-rigid atlas construction using the more recent iterative "groupwise" approach (cf. dhcp-v3.0).
The used method is described in:
- A Schuh, M Murgasova, A Makropoulos, C Ledig, SJ Counsell, JV Hajnal, P Aljabar, D Rueckert, Construction of a 4D Brain Atlas and Growth Model using Diffeomorphic Registration, 3rd MICCAI Workshop on Spatiotemporal Image Analysis for Longitudinal and Time-Series Image Data (STIA), LNCS 8682, pp. 27-37, Boston, MA (2014).
- A Schuh, Computational models of the morphology of the developing neonatal human brain, PhD thesis, ch. 6.2 (2017).