Multilevel Image Thresholding Using Bat Algorithm Based on Otsu

In this paper, optimal thresholds for multi-level thresholding in image segmentation are gained by maximizing Otsu’s between-class variance using bat algorithm (BA). The performances of the proposed algorithm are demonstrated by considering four benchmark images. The performance assessment is carried using peak-to-signal ratio (PSNR) and root mean square error (RMSE). The experiment results show that the more threshold, the better the segmentation effect.