Parallel Marker-Based Image Segmentation with Watershed Transformation

The parallel watershed transformation used in gray-scale image segmentation is here augmented to perform with the aid of a priori supplied image cues called markers. The reason for introducing markers is to calibrate a resilient algorithm to oversegmentation. In a hybrid fashion, pixels are first clustered based on spatial proximity and gray-level homogeneity with the watershed transformation. Boundary-based region merging is then effected to condense nonmarked regions into marked catchment basins. The agglomeration strategy works with a weighted neighborhood graph representation of the oversegmented image. The throughput of a parallel Bor?vka-like minimum spanning forest (MSF) operator, applied on the considered graph, embodies the desired image partition, reasoning that all regions in a tree fuse into a homogeneous area containing a unique marker. Two figures of merit of the parallel algorithm are worth mentioning: the local detection of the catchment basins conforming to the watershed principle (which strongly depends on the history of the regions' growth) and the parallel computation of the Bor?vka-like MSF which merges, at the same time, partial regions, produced by the local labeling, and nonmarked regions to marked basins. Both modules are designed with great concurrency, locality, and reduced software engineering cost, emerging into a scalable algorithm.

[1]  Fernand Meyer,et al.  Minimum Spanning Forests for Morphological Segmentation , 1994, ISMM.

[2]  Moncef Gabbouj,et al.  A Parallel Watershed Algorithm Based on Rainfalling Simulation , 1995 .

[3]  Robert E. Tarjan,et al.  Finding Minimum Spanning Trees , 1976, SIAM J. Comput..

[4]  Moncef Gabbouj,et al.  A parallel marker based watershed transformation , 1996, Proceedings of 3rd IEEE International Conference on Image Processing.

[5]  Moncef Gabbouj,et al.  A Parallel Watershed Algorithm Based on the Shortest Path Computation , 1995 .

[6]  Robert E. Tarjan,et al.  Data structures and network algorithms , 1983, CBMS-NSF regional conference series in applied mathematics.

[7]  Viktor K. Prasanna,et al.  Parallel Architectures and Algorithms for Image Component Labeling , 1992, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Moncef Gabbouj,et al.  Parallel Image Component Labeling With Watershed Transformation , 1997, IEEE Trans. Pattern Anal. Mach. Intell..

[9]  Joseph JáJá,et al.  An Introduction to Parallel Algorithms , 1992 .

[10]  George Karypis,et al.  Introduction to Parallel Computing , 1994 .

[11]  Pierre Soille,et al.  Mathematical Morphology and Its Applications to Image Processing , 1994, Computational Imaging and Vision.

[12]  Message Passing Interface Forum.  MPI: A message - passing interface standard , 1994 .

[13]  Ronald L. Graham,et al.  On the History of the Minimum Spanning Tree Problem , 1985, Annals of the History of Computing.

[14]  Paul T. Jackway,et al.  Gradient watersheds in morphological scale-space , 1996, IEEE Trans. Image Process..

[15]  Fernand Meyer,et al.  Topographic distance and watershed lines , 1994, Signal Process..

[16]  Michael J. Quinn,et al.  Designing Efficient Algorithms for Parallel Computers , 1987 .

[17]  Moncef Gabbouj,et al.  Parallel Watershed Algorithm Based on Sequential Scanning , 1995 .

[18]  Wojciech Rytter,et al.  Efficient parallel algorithms , 1988 .

[19]  Edward R. Dougherty,et al.  Mathematical Morphology in Image Processing , 1992 .

[20]  S. Sitharama Iyengar,et al.  Introduction to parallel algorithms , 1998, Wiley series on parallel and distributed computing.

[21]  Moncef Gabbouj,et al.  An efficient watershed segmentation algorithm suitable for parallel implementation , 1995, Proceedings., International Conference on Image Processing.