Research on Deep Learning-Based Semantic Segmentation Algorithms for Subcortical Brain Structures
FD-FCN is a new 3D patch-based network designed for fast and accurate segmentation of subcortical brain structures in T1 MRI scans. It achieves an 89.81% Dice overlap, outperforming existing methods by at least 3.66% in accuracy and significantly speeding up processing time to just 53 seconds per image.