Image Segmentation using Exchange Market Algorithm

Image segmentation is the basic step for assisting experts to determine immense required information from the image for real time applications. In this research work the most promising objective functions such as Kapur Otsu and Minimum Cross Entropy (MCE) are used for precise image segmentation. In this work the similarity detection based multilevel thresholding technique is used to achieve the target. The objective is attained through powerful robust Exchange Market Algorithm (EMA) aided with the objective functions. The three teams of EMA in stable and unstable market situations and primarily the role of team B and C following team A of EMA plays a vital role to achieve balanced exploration and exploitation. Thus the segmented details assist the experts for various real time applications. The proposed method using EMA based MLT is applied and tested with four different threshold values m = 2 3 4 5 for gray images and the color images are tested at 456 and 7 threshold levels. Various performance metrics such as low CPU time high PSNR with low RMSE high SSIM and uniformity measure validates the performance of the proposed technique.