A Novel Bi-Level Artificial Bee Colony Algorithm and its Application to Image Segmentation

Image segmentation requires optimum multilevel threshold values obtained from the image in order to partition it into multiple regions. Estimating these thresholds poses a great challenge. In this paper, we propose a novel swarm intelligence technique, namely Bi-level Artificial Bee Colony (BABC) algorithm, to obtain the optimum thresholds by using the Tsallis Entropy as an objective function. BABC is used, along with a Sinusoidal Evaluation of Fitness Function (SEFF), to ensure that all the threshold values of the image are examined before arriving at the best possible solution. Experimental results show the promising performance of BABC for image segmentation as compared to other optimization algorithms like Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and Bacterial Foraging (BF) Algorithm.