An efficient watershed transform computation method

In this paper, we propose a watershed computation method that is faster than and is devoid of certain drawbacks experienced by a traditional immersion-based watershed algorithm. Flooding of catchment basins in our algorithm always starts from precomputed regional minima and conditional neighborhood comparisons while processing the eight neighboring pixels of a labeled center pixel. This invariably results in thin continuous watershed lines. Better time complexity of the proposed algorithm relative to the traditional algorithm stems from the fact that the labels of various neighboring pixels are determined at the same time. The results of running both the proposed and the traditional algorithm on different images clearly establish the superiority of the proposed algorithm.