Fuzzy clustering algorithms based on resolution and their application in image compression

This paper presents an idea of clustering resolution. On the basis of the idea, fuzzy clustering algorithms based on resolution are deduced, which naturally comprise a set of clustering algorithms. Thus, e-means algorithm and fuzzy c-means algorithms are actually special examples in the set. As an application for codebook design in image compression based on vector quantization, fuzzy clustering algorithms based on multiresolution are developed, which are almost prior to conventional algorithms in all aspects. (C) 2002 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.