Utilizing the anatomical fiducals framework to identify other salient brain regions and automatic localization of anatomical fiducials using neural networks
Convert3D
Convert3D:
- .fcsv -> threshold image -> landmark distance map (could be considered probability map)
- distance map used for training
Convert3D:
- brainmask.nii -> 3D patches sampled at x voxels
- matching of distance maps and anatomical imaging patches is crucial for proper training