FEATURE-EXTRACTION TECHNIQUES FOR EXPLORATORY VISUALIZATION OF VECTOR-VALUED IMAGERY

This paper addresses the exploratory visualization of multispectral im age data. In such data, each component of the vector pixel corresponds to a different imaging modality or a different combination of imaging parameters, and may provide different levels of contrast sensitivity between different regions of the underlying image. We address the prob lem of presenting this multidimensional data to human observers by syn thesizing a display matched to their visual capabilities. Specifically , we seek to determine a data-adaptive linear projection of the vector data to one dimension that produces a grayscale image providing maxim um discrimination between the different regions of the underlying obje ct. The approach is equivalent to the extraction of the best linear fe ature of the vector field. Several new feature-extraction criteria tha t take into account both the spatial and multivariate structures of th e data are proposed and illustrated by simulations on test images.