Abstract

The image segmentation is a technique of image processing which divides image into segments. The many proposed image segmentation techniques, k-Means clustering has been one of the basic image segmentation techniques. The advantages of k-Means are easy calculation, the number of small iteration, and one of the most commonly used clustering algorithm. but, The main problem in this algorithm is sensitive to selection initial cluster center. In this research, we present two approaches method which are used to execute image. It is PSO and k-Means. k-Means integrated with Particle Swarm Optimization (PSO) to improve the accuracy. The purpose of this research to find the effect of PSO towards k-Means in order to get the best selection initial cluster center. This research has been implemented using matlab and taking image dataset from weizzmann institute. The Result of our experiment, we have different result RMSE of k-Means PSO. Euclidean has less RMSE value than Manhattan. The difference RMSE between Euclidean PSO and Manhattan PSO only four point. but if we compare by processing time we have significant difference.

Authors

Elindra Ambar Pambudi, Pulung Nurtantio Andono, Ricardus Anggi Pramunendar
Universitas Dian Nuswantoro, Indonesia

Keywords

Vision Computing, Image Processing, Segmentation, Swarm Intelligence, Computer Science