Two-dimensional thresholding by new evolution mechanism

Computational cost is the main drawback for two-dimensional (2-D) thresholding methods. GAs can accelerate the calculation, but still fail to satisfy most of applications. By introducing an implantation, which replaces the selection of GAs, an evolution mechanism was proposed. An excellent schema is implanted in the low performance strings to improve both the performance and diversity of population. Thus, the search efficiency will be improved in each iteration cycle, resulting in the acceleration of the threshold selecting. Experimental results showed that the proposed method could reduce the computation cost in 2-D Renyi´s entropy thresholding