Queue and priority queue based algorithms for computing the quasi-distance transform

Abstract

The quasi-distance transform introduced by Beucher shows interesting properties for various tasks in image processing such as segmentation, filtering and images simplification. Despite its simple formulation, a naive and direct implementation of the transform leads to poor results in terms of computational time. This article proposes a new algorithm for computing the quasi-distance, based on a front propagating approach by means of queues and hierarchical queues. Quantitative analysis of the running time are provided, and show an exponential downscale of the complexity compared to the original algorithm. © 2010 Springer-Verlag.