Pixel-Label-Based Segmentation of Cross-Sectional Brain MRI Using Simplified SegNet Architecture-Based CNN

Table 4

Comparison of deep learning approaches for brain structure segmentation.

Authors	CNN style	Dimension	Accuracy	Data
Zhang et al. [13]	Patchwise	2D	DSC 83.5% (CSF), 85.2% (GM), 86.4% (WM)	Private data (10 healthy infants)
Nie et al. [14]	Semantic-pixelwise	2D	DSC 85.5% (CSF), 87.3% (GM), 88.7% (WM)	Private data (10 healthy infants)
de Brebisson et al. [15]	Patchwise	2D/3D	Overall DSC 72.5% ∓ 16.3%	MICCAI 2012-multi-atlas labeling
Moeskops et al. [16]	Patchwise	2D/3D	Overall DSC 73.53%	MICCAI 2012-multi-atlas labeling
Proposed method	Pixel-label semantic (SegNet CNN)	2D	DSC 72.2% (CSF), 74.6% (GM), 81.9% (WM)	OASIS cross-sectional MRI