Effective Image Segmentation using Modified K-Means Technique

In general, images are segmented based on some similarity characteristics. This technique is very useful in medical, satellite, multi-focus, image processing applications. Once images are segmented, it is easy to process the important regions in the images. Clustering can be implemented in many ways. The most popular unsupervised clustering algorithm is a K-means clustering algorithm. This is used to make several clusters. In this work, to begin with [11] K-means clustering algorithm is applied to the original image. In the second step edge detection is used to segment the regions effectively. Experiments are performed on five images. Experimental results are indicating that the modified K-Means algorithm is giving better results. To examine the performance of the present algorithm, the proposed work have analyzed the performance metrics like accuracy, precision, recall, F1 score, and sensitivity.