Search results for key=TSB2004a : 1 match found.

In the last years there has been a boom in digital real-world image databases, for instance, in medical imagery and consumer digital photography. In Information Retrieval, the field of Content-Based Image Retrieval (CBIR) aims to retrieve images from a repository, and rank the retrieved images by computing the similarity between the images in the database and a query image. A basic process in CBIR is the extraction of features from the digitized representation of an image on a computer. These features are intended to represent features that coincide with human visual experience. We explore the human visual experience of an image and discuss what an image is, and how it relates to the visual experience. We show that an image consist of indivisible information carrying elements that make up the whole of the image. We argue that a CBIR system must have an intermediate level for reasons of efficiency and effectiveness. Then we apply the technical philosophical notion of supervenience to explain the nature of the relation between information carrying elements, the intermediate level and the image. At the end of the paper, we offer our ontology of an image in a CBIR system.