MakeTrainData[{data1,data2,n}]
makes multi contrast training data by blending the co-registered 3D datasets data1 and data2 in n linear steps from data1 to data2. Each blend is normalized and the result is masked. The output is a 4D dataset where each blend is a channel.
MakeTrainData[{{data1,data2,n},..}]
does the same for multiple pairs and joins all blends as channels, e.g. {{outph, inph, 6}, {wat, fat, 5}}. The output can be used as data for PrepareTrainingData, during training a random channel is selected for each sample which acts as contrast augmentation.
MakeTrainData
MakeTrainData[{data1,data2,n}]
makes multi contrast training data by blending the co-registered 3D datasets data1 and data2 in n linear steps from data1 to data2. Each blend is normalized and the result is masked. The output is a 4D dataset where each blend is a channel.
MakeTrainData[{{data1,data2,n},..}]
does the same for multiple pairs and joins all blends as channels, e.g. {{outph, inph, 6}, {wat, fat, 5}}. The output can be used as data for PrepareTrainingData, during training a random channel is selected for each sample which acts as contrast augmentation.