PSO-based Membrane Clustering Algorithm And Its Application In Image Compression

As a branch of natural computing, membrane computing is a novel class of distributed parallel computing models. Membrane computing, which describes models called P systems, is inspired by the structure and functioning of living cells, as well as from the way the cells are organized in tissues or higher order structures.Membrane computing has several attractive features such as distribution, maximum parallelism and non-determinism. At present, the research on membrane computing has attracted many researchers.Clustering is a procedure of grouping a batch of real or abstract data objects into several classes or clusters. The most common criterion adopted in partitional clustering is minimizing some measures of dissimilarity in the samples within each cluster and maximizing the dissimilarity of different clusters, also known as the proverb that a feather flock together, so that to discover the inner structure of data set.Image compression known as compression coding, means to reduce the amount of data required to represent a digital image. And also means on the premise of no significant distortion, converting the image bitmap information into another expression which can reduce the amount of data. Image compression is one of the most useful and successful technology in the area of digital image processing and on business.This paper uses the tissue-like P system as the computing framework and discusses the PSO-Based Membrane clustering algorithm and its application in image compression. The main researches are shown as follows:(1) A clustering algorithm using the particle swarm optimization(PSO) mechanism under the framework of membrane computing is proposed, called PSO-MC algorithm. A tissue-like P system is designed to exploit the optimal cluster centers, where the velocity-position model of PSO is used to evolve the objects and inherent communication mechanism is used to share the objects between different elementary membranes. The proposed clustering algorithm is evaluated on two real-life data sets and is further compared with classical k-means algorithm, genetic algorithm(GA)-based clustering algorithm and particle swarm optimization(PSO)-based clustering algorithm respectively. Such comparisons reveal the superiority of the PSO-MC algorithm in terms of performance and robustness.(2)Apply the PSO-Based Membrane clustering algorithm to image compression based on vector quantization. The proposed method is compared with the existing vector quantization image compression method on several real-life gray images.