
Cantata(Subform)             Khoros              Cantata(Subform)

     CLASSIFICATION/CLUSTERING ALGORITHMS

     These routines are used to perform supervised or unsuper-
     vised classification.

     SHORT DESCRIPTION OF ROUTINES

     K-means -           Perform the Kmeans clustering algorithm
                         on an image.

     IS02 -              Perform the IS02 clustering algorithm on
                         an image.

     Non-parametric -    Perform N-dimensional vector quantiza-
                         tion or classification.

     Minimum Distance -  Perform Minimum Distance classification
                         on an image.

     Weighed Minimum distance detector -
                         Perform a weighted minimum distance
                         detection or classification.

     Labeling -          Perform region labeling on a multiband
                         or cluster image.

     LRF-training -      Trains for the weights of the Localized
                         Receptive Field classifier.

     LRF-classifier -    Classifies an image using the Localized
                         Receptive Field classifier.

     Quantization Error -
                         Computes the RMS error between a quan-
                         tized image and the original image.

                         Copyright 1991, University of New Mex-
                         ico. All rights reserved.

Khoros                    Release: 1.0                          1

