A Multi-Agent based Approach for Fuzzy Clustering of Large Image Data

Nashwa M. Abdelghaffar · · ; HewaydaM. S. Lotfy ; Khamis, Soheir 


Data clustering usually requires extensive computations of similarity measures between dataset members and cluster centers, especially for large datasets. Image clustering can be an intermediate process in image retrieval or segmentation, where a fast process is critically required for large image databases. This paper introduces a new approach of multi-agents for fuzzy image clustering (MAFIC) to improve the time cost of the sequential fuzzy c-means algorithm (FCM). The approach has the distinguished feature of distributing the computation of cluster centers and membership function among several parallel agents, where each agent works independently on a different sub-image of an image. Based on the Java Agent Development Framework platform, an implementation of MAFIC is tested on 24-bit large size images. The experimental results show that the time performance of MAFIC outperforms that of the sequential FCM algorithm by at least four times, and thus reduces the time needed for the clustering process.

Other data

Keywords Image clustering Fuzzy c-means Multi-agent system
Issue Date 17-Nov-2014
Publisher Springer
Journal Journal of Real-Time Image Processing 
URI http://research.asu.edu.eg/123456789/1117
DOI 10.1007/s11554-014-0473-3

File Description SizeFormat 
10.1007_s11554-014-0473-3.pdf1.72 MBAdobe PDFView/Open
Recommend this item

CORE Recommender


Items in Ain Shams Scholar are protected by copyright, with all rights reserved, unless otherwise indicated.