Segmentation Of Medical Images Based On Hierarchical Evolutionary And Bee Algorithms
Hamed Azami, Milad Azarbad, Saeid Sanei

Dynamic or adaptive thresholding strategy is of high interest in pattern recognition, signal and image processing. In this article a powerful method using a combination of multilevel thresholding algorithm, bee algorithm (BA), and the hierarchical evolutionary algorithm (HEA) is proposed for segmentation of magnetic resonance images (MRIs). The HEA can be viewed as a modified variant of basic genetic algorithm (GA). The proposed method is based on the BA and, in fact, is an unsupervised clustering method depending on an automatic multilevel thresholding approach. One advantage of the proposed method is that the number of clusters in the given image does not require to be known previously. The results show that the accuracy of the proposed algorithm is very excellent (about 97%).