[工程科技]RotationInvariantTextureRetrievalwithGaussianizedSteerablePyramids_”submittedto.pdfVIP

Rotation-invariant texture retrieval with Gaussianized steerable pyramids George Tzagkarakis , Baltasar Beferull-Lozano , Associate Member, IEEE, and Panagiotis Tsakalides , Member, IEEE Department of Computer Science, University of Crete Institute of Computer Science - FO.R.T.H. ICS-FORTH, P.O. Box 1385, 711 10 Heraklion, Crete, Greece Phone: +30 2810 391670, Fax: +30.81 - 391601 E-mail: gtzag@ics.forth.gr, tsakalid@ics.forth.gr School of Computer and Comm. Sciences Audiovisual Communications Laboratory Swiss Federal Institute of Technology (EPFL) CH1015, Lausanne, Switzerland E-mail: Baltasar.Beferull@epfl.ch Submitted to: IEEE Trans. on Image Processing (April 2005) Revised: September 2005 EDICS: 5-OTHE 1 Abstract This paper presents a novel rotation-invariant image retrieval scheme based on a transformation of the texture information via a steerable pyramid. First, we fit the distribution of the subband coefficients using a joint alpha-stable sub-Gaussian model to capture their non-Gaussian behavior. Then, we apply a normalization process in order to Gaussianize the coefficients. As a result, the feature extraction step consists of estimating the covariances between the normalized pyramid coefficients. The similarity between two distinct texture images is measured by minimizing a rota