Fast Palmprint Identification Using Orientation Pattern Hashing

Abstract

A palmprint identification system recognizes a query palmprint image by searching for its nearest neighbor from among all the templates in a database. When applied on a large-scale identification system, it is often necessary to speed up the nearest-neighbor searching process. In this paper, by viewing the palmprint feature as a high-dimension binary vector, we present a palmprint identification method using orientation pattern hashing. We propose three properties required by the hash function and demonstrate that the orientation pattern has all of these properties. Under some simple assumptions we give the parameter selection method for fast and accurate palmprint identification. Experimental results on the Hong Kong large scale database (9667 palms) show that the proposed method is over 16 times faster than brute force searching, while its accuracy is slightly higher. Evaluations on the CASIA palmprint database (600 palms) plus a synthetic database (100,000 palms) show a speedup of 6.8 over brute force searching and a negligible loss of accuracy.