Title :
Locating an affine/projective invariant identifier patch on an image
Author :
Araabi, B.N. ; Kehtarnavaz, N. ; Yeary, M. ; Hillman, G. ; Würsig, B.
Author_Institution :
Dept. of Electr. & Comput. Eng., Tehran Univ., Iran
Abstract :
Single view recognition is encountered in many photo-identification applications. Often a query image is retrieved by searching a database of previously identified images. Important markings and patterns are often located in a particular area of the object. This paper presents a method to locate an invariant patch on an image using few points in correspondence and few invariant ratios. Computation of ratios is simple and intuitive both geometrically and analytically. Once ratios are computed for an appropriate multilateral patch, the same ratios can be used on other images to locate the same identifier patch. In particular, this method is deployed to obtain the same area of interest on spotted dolphins
Keywords :
biocommunications; image recognition; image retrieval; visual databases; affine/projective invariant identifier patch; correspondence points; database searching; invariant ratios; multilateral patch; photo-identification; query image retrieval; single view recognition; spotted dolphins; Biomedical imaging; Cameras; Dolphins; Image analysis; Image databases; Image recognition; Image retrieval; Information retrieval; Seals; Spatial databases;
Conference_Titel :
Image Analysis and Interpretation, 2002. Proceedings. Fifth IEEE Southwest Symposium on
Conference_Location :
Sante Fe, NM
Print_ISBN :
0-7695-1537-1
DOI :
10.1109/IAI.2002.999902