API Documentation

gaborExtract

The gaborExtract module is responsible for efficiently extracting Gabor means and stdevs from a random sampling of pixels, using multiple window sizes and Gabor orientations. These operations are GPU-accelerated on platforms with compatible GPUs. Features are stored in HDF5 files.

classifyFeatures

Given features extracted from the gaborExtract model, this module parses feature files and classifies sets of unknown pixel features as belonging to known pixel feature-sets using a k-nearest-neighbors approach.