Image Segmentation through Clustering Based on Natural Computing Techniques

Abstract

Natural Computing (NC) is a novel approach to solve real life problems inspired in the life itself. A diversity of algorithms had been proposed such as evolutionary techniques, Genetic Algorithms and Particle Swarm Optimization (PSO). These approaches, together with fuzzy and neural networks, give powerful tools for researchers in a diversity of problems of optimization, classification, data analysis and clustering. Clustering methods are usually stated as methods for finding the hidden structure of data. A partition of a set of N patterns in a p-dimensional feature space must be found in a way that those patterns in a given cluster are more similar to each other than the rest. Applications to clustering algorithms range from engineering to biology (Xu & Wunsch II, 2005; Xu & Wunsch, 2008; Jain et al., 1999). Image segmentation techniques are based on Pattern Recognition concepts and such a task aims to identify behavior in a data set. In the context of image segmentation, the data set represents image data, coded as follows: the light intensity value (the pixel data) represents a pattern, an item in the data set, and the color information is represented by columns (the feature vectors). Clustering techniques represent the non-supervised pattern classification in groups (Jain et al., 1999). Considering the image context, the clusters correspond to some semantic meaning in the image, which is, objects. More than simple image characteristics, these grouped semantic regions represent information; and image segmentation is applicable in an endless list of areas and applications, for example: computer-aided diagnosis (CAD) being used in the detection of breast cancer on mammograms (Doi, 2007), outdoor object recognition, robot vision, content-based image, and marketplace decision support. Among the many methods for data analysis through clustering and unsupervised image segmentation is: Nearest Neighbor Clustering, Fuzzy Clustering, and Artificial Neural Networks for Clustering (Jain et al., 1999). Such bio and social-inspired methods try to solve the related problems using knowledge found in the way nature solves problems. Social inspired approaches intend to solve problems considering that an initial and previously defined weak solution can lead the whole population to find a better or a best so far solution. This chapter presents concepts and experimental results of approaches to data clustering and image segmentation using (NC) approaches. The main focus are on Evolutionary Computing, which is based on the concepts of the evolutionary biology and individual-topopulation adaptation, and Swarm Intelligence, which is inspired in the behavior of individuals, together, try to achieve better results for a complex optimization problem.