Efficient watershed segmentation algorithm suitable for parallel implementation

An important aspect of designing a parallel algorithm is exploitation of the data locality for minimization of the communication overhead. Aiming at this goal, we propose here a reformulation of a global image operation called the watershed transformation. The method lies among various approaches for image segmentation and performs by labeling connected components. Both serial and parallel programming models are presented and evaluated when running on SUNR and DECR Alpha AXP workstations, and a Cray T3D, respectively