Automatic Neonatal Brain Segmentation for Nine Tissue Types Simultaneously

Petronella Anbeek¹, Britt van Kooij², Floris Groenendaal², Linda S. de Vries², Max A. Viergever¹, Manon J. Benders²

This paper proposes a fully automated method for segmentation of brain tissue in neonatal MR imaging. The method is based on K-nearest neighbor classification and segments nine tissue types simultaneously: (un)myelinated white matter, cortical gray matter, basal ganglia, cerebro-spinal fluid, ventricles, brainstem, cerebellum, myelinated white matter in the cerebral peduncle, and the posterior limb of the internal capsule. Manual segmentations were used for training and evaluation of the results. High accuracy is reached for the segmentation results of all larger tissue types.