CSDL Home IEEE Transactions on Pattern Analysis & Machine Intelligence 1997 vol.19 Issue No.05 - May
Issue No.05 - May (1997 vol.19)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/34.589204
<p><b>Abstract</b>—The parallel watershed transformation used in gray scale image segmentation is reconsidered in this paper on the basis of the component labeling problem. The main idea is to break the sequentiality of the watershed transformation and to correctly delimit the extent of all connected components locally, on each processor, simultaneously. The internal fragmentation of the catchment basins, due to domain decomposition, into smaller subcomponents is ulteriorly solved by employing a global connected components operator. Therefore, in a pyramidal structure of master-slave processors, internal contours of adjacent subcomponents within the same component are hierarchically removed. Global final connected areas are efficiently obtained in <it>log</it><sub>2</sub><it>N</it> steps on a logical grid of <it>N</it> processors. Timings and segmentation results of the algorithm built on top of the <it>Message Passing Interface</it> (MPI) and tested on the Cray T3D are brought forward to justify the superiority of the novel design solution compared against previous implementations.</p>
Watersheds, connected components, image segmentation, parallel computing, efficient algorithms.
Alina N. Moga, Moncef Gabbouj, "Parallel Image Component Labeling With Watershed Transformation", IEEE Transactions on Pattern Analysis & Machine Intelligence, vol.19, no. 5, pp. 441-450, May 1997, doi:10.1109/34.589204