Image classification using local binary pattern operators for static images
Oana Astrid Vatamanu, Mircea Jivulescu · 2013
This paper aims to present an image classification method using Local Binary Pattern techniques. Local Binary Pattern operator transforms an static image, at pixel level, into a matrix of labels. These labels - integer numbers - describe and characterise the original image at a much lower scale. The authors propose the use of labels as a global characteristic of an static image. These techniques can be applied to an image or to a group of images and the characterization is done through an array of values extracted by the algorithm. The application developed allows the characterization of an image or a set of images, determining the similarity between different images and the degree of belonging to a particular group. Vectors of values are required for more images and image groups and each vector is representing different textures and their classification. As a result it becomes possible that indexing images, take into account the content of the information present in the image.