Summary of state of arts

01 Dec 2018

The convNet for medical image segmentation

• DeepNAT(Wachinger et al., 2017)
• VoxResNAT(Chen et al., 2017)
• MeshNET (Fedorov et al., 2017)
• U-Net

Multi-atlas Segmentation

First, the n most similar atlases to the query image are selected and registered to the query image; Second, the same registration transformations are applied to the labels of the n atlases, and these labels are propagated to produce n segmentations of the query image; Finally the segmentations are combined using a fusion strategy.

• PICSL
• Freesurfer

Image Segmentation

• FCN