PrepareTrainingData

QMRITools`

QMRITools`

PrepareTrainingData

PrepareTrainingData[inFolder,outFolder]

prepares the data in de inFolder for training a neural network for segmentation and outputs in outFolder.

PrepareTrainingData[{labFolder,datFolder},outFolder]

does the same but the labels are stored in labFolder and data is stored in datFolder.

Details

The following options can be given:

	LabelTag	"label"	LabelTag is an option for PrepareTrainingData. It defines the tag used in the filenames of the label data.
	DataTag	"data"	DataTag is an option for PrepareTrainingData. It defines the tag used in the filenames of the data.
	OutputTag	""	OutputTag is an option for PrepareTrainingData. It defines the tag that will be added to the output training data.
	InputLabels	Automatic	InputLabels is an option for PrepareTrainingData. Can be set to a list of integers corresponding to the labels to be used from the given segmentation.
	OutputLabels	Automatic	OutputLabels is an option for PrepareTrainingData. Can be set to a list of integers. The used label number will be replaced by these numbers.
	TrainVoxelSize	Automatic	TrainVoxelSize is an option for PrepareTrainingData. It defines the voxel size of the training data. When set to Automatic the voxel size is that of the data.
	CleanUpSegmentations	True	CleanUpSegmentations is an option for PrepareTrainingData. If set to True the segmentations are cleaned up by removing holes reducing to one volume and smoothing.
	TestRun	False	TestRun is an option for PrepareTrainingData. If set to True the data is not saved only analyzed.
	SegmentationsPerSlice	All	SegmentationsPerSlice is an option for PrepareTrainingData. It defines how many segmentations should be present in a slice for it to be used in the training data.