Data Clustering Methods Applied to Image Segmentation
Image processing, image segmentation, textures, Gaussian blending, clustering.
Due to the great technological advance, the computer is increasingly being used to extract information in images. Segmentation is one of the most main and complex process in image processing. This paper presents a new technique of images automatic segmentation based on textures through a clustering application method. This method is composed of a set of distinct techniques and divided into two steps. The first step uses vector quantization to group the data set (pixels) in auxiliary centers and the second step group the auxiliary centers that have more common properties in a single group. The main differential of this proposal is the use the method Quadratic Renyi’s Negentropy Separation (QRNS) which creates a hierarchical structure between groups found. Validation of the method is performed with different images and its performance compared to other methods.