Cuckoo search with differential evolution mutation and Masi entropy for multi-level image segmentation

Since the beginning of the twenty-first century, the Cuckoo Search (CS) algorithm has emerged as one of the robust, flexible, fast, and easily implementable techniques for the global search to solve many complex problems over continuous spaces. CS operates like other Nature-Inspired Algorithms (NIOA) whose effectiveness significantly depends on the exploration and exploitation phases. CS already proofs its efficiency in solving real-world optimization problems in various application domains. In this study, the author tries to enhance the efficiency of the CS by incorporating six different mutation strategies of Differential Evolution (DE). The performance of the proposed CS variants has been investigated over Multi-level thresholding based image segmentation field as it is considered one of the dominant image segmentation techniques of the recent era. It is known that computation of the optimal set of thresholds is significantly influenced by the considered objective function, and it can be trapped into local optima. On the other hand, the computational time of Multi-level thresholding increases exponentially when the number of threshold points increases. To overcome these problems, this study introduces CS variants over this segmentation field, where Masi entropy is maximized to find the optimal threshold points. The experiment has been conducted on various color pathology images. The results of such a comparative study provide valuable insight and information to develop efficient CS variants using optimal or adaptive mutation strategies of DE.

Automated summary

The study aims to enhance the efficiency of the Cuckoo Search algorithm in image segmentation by incorporating mutation strategies of Differential Evolution. Results of the experiment provide valuable insight for developing efficient CS variants using optimal or adaptive mutation strategies of DE.