THE HOLODISC DISTANCE TRANSFORM AND ITS APPLICATIONS IN IMAGE-ANALYSIS

Distance transforms are essential tools in image analysis for obtainin g the complete spectrum of eroded or dilated sets from a given object. Moreover, their remarkable topographical properties have led to numer ous developments in the field of shape segmentation and analysis. Most ly for algorithmical reasons, these operators have been derived from s quare or hexagonal distance transforms. In this paper, we show how the holodisc distance function makes it possible to access even more powe rful tools while taking advantage of a truly Euclidean metric.