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 preserves the locality of data to be processed. We propose a region growth algorithm based on arrowing graphs that is strictly linear despite the valuation domain of input images. Simultaneous and disorderly growth is made possible by using a synchronization mechanism coded directly on the weight of nodes. Experimental results show that the algorithm is accurate and by far outperforms common watershed algorithms.
Document type :
Conference papers
Complete list of metadatas

Cited literature [16 references]  Display  Hide  Download

Contributor : Théodore Chabardès <>
Submitted on : Tuesday, February 9, 2016 - 5:47:09 PM
Last modification on : Monday, November 12, 2018 - 11:01:40 AM
Long-term archiving on : Saturday, November 12, 2016 - 3:40:10 PM


Files produced by the author(s)


Distributed under a Creative Commons Attribution 4.0 International License


  • HAL Id : hal-01266889, version 2


Theodore Chabardes, Petr Dokládal, Matthieu Faessel, Michel Bilodeau. A parallel, O(n), algorithm for unbiased, thin watershed . IEEE International Conference on Image Processing, Sep 2016, Phoenix, United States. ⟨hal-01266889v2⟩



Record views


Files downloads