The Menze group publishes a deep learning-based approach to segment the claustrum in human brain MRI scans
Automated claustrum segmentation in human brain MRI using deep learning
Despite the central role of the claustrum in the structure and function of the mammalian forebrain, few in vivo human studies of this delicate, sheet-like structure exist. Here, the Menze group and their collaborators present a Deep Learning-based multi-view approach for segmenting the claustrum in T1-weighted MRI scans. The algorithm enables robust automatic claustrum segmentation and therefore holds potential for facilitating MRI-based human claustrum studies.
See Li et al., Hum Brain Mapp.
Access the software and models via: https://github.com/hongweilibran/claustrum_multi_view