A parallel, O(N) algorithm for unbiased, thin watershed

Abstract

The watershed transform is a powerful tool for morphological segmentation. Most common implementations of this method involve a strict hierarchy on gray tones in processing the pixels composing an image. Those dependencies complexify the efficient use of modern computational architectures. This paper aims at answering this problem by introducing a new way of simulating the waterflood that alleviates the sequential nature of hierachical queue propagation. Simultaneous and disorderly growth is made possible using this method. higher speed is reached and bigger data volume can be processed. Experimental results show that the algorithm is accurate and produces a thin, well centered watershed line.