Efficient Dynamic Image Retrieval Using the Á Trous Wavelet Transformation

Abstract

The increasing use of digital image in applications ranging from remote sensing to medical applications and industrial control systems results in a demand for well-suited and efficient techniques for their storage, management and retrieval. The state-of-the-art approach for image retrieval considers a priori extracted features, which are compared to query information supplied by a user in the form of, for example, a list of keywords or the corresponding features of a sample image or sketch. In this paper an alternative. object-based approach for image retrieval is presented. This allows the user to specify and to search for certain regions of interest in images. The marked regions are represented by wavelet coefficients and searched in all image sect,ious during query runtime. All other image elements are ignored, thus a detailed search can be realisd. The resulting computational effort can be overcome by utilisation of parallel architectures. An example for a cluster-based image database is discussed in the last part of this paper.