Page MenuHomePhabricator

Feature processor for classification
Closed, WontfixPublic


Would be useful to have a class that takes a number of images and computes a selection of features on them, returning a map of all including the inputs to later extract feature vectors for classification.
Maybe start with a base class that just handles input (maybe conversion to specific data type if desired) and output and has an abstract ComputeFeatures method that child classes can then override with the desired pipeline of filters, like an ITK ProcessObject...