Generalized Distance Transforms using the FEED-class Algorithm

Abstract

The Fast Exact Euclidean Distance (FEED) transform is generalized to support intensity values and gray-scale images. The generalized FEED (gFEED) class algorithms support both Euclidean and squared Euclidean distances. The algorithms are tested on datasets by Schouten and van den Broek and on newly developed datasets. Tests show a O(m*n) time complexity on squared Euclidean distances and a O(m * n)(m+ n) time complexity on Euclidean distances.