An efficient data mining technique with applications in multivariate image analysis

The "curse of dimensionality" complicates the analysis of multiparameter images. The application of a data mining technique, as a tool to resolve the image puzzle, is the subject of this study. This well developed technique is a sequential planing procedure, based on the minimal spanning tree of the data. It is armed with a Hopfield neural network implementation and, being computationally efficient, appears to be very attractive for image understanding tasks. Integrated in a GUI environment, it can be a useful prelude to the subsequent image analysis.