Outline for 2nd lecture on segmentation.
SLT -- Monday, 1 May 1995

A. Uniformity as having the same planar approximation

B. Split-and-Merge algorithm
  (there was not enough time to cover it on Friday)

C. Segmentation in terms of clustering.
  Clustering -- terminology, objectives in terms of
   minimizing within-cluster variance, maximizing intercluster distances.

D. ISODATA algorithm.

E. Pyramid Linking algorithm of Burt, Hong and Rosenfeld.