Adaptation of K - Means Algorithm for Image Segmentation

Image segmentation based on an adaptive K-means clustering algorithm is presented. The proposed method tries to develop K-means algorithm to obtain high performance and efficiency. This method proposes initialization step in K-means algorithm. In addition, it solves a model selection number by determining the number of clusters using datasets from images by frame size and the absolute value between the means, and additional steps for convergence step in K-means algorithm are added. Moreover, in order to evaluate the performance of the proposed method, the results of the proposed method, standard K-means and recently modified K-means are compared. The experimental results showed that the proposed method provides better output. Keywords— Initialization, Modify K-Means, Segmentation, Standard K-Means.