3-D Ultrasound Palmprint Recognition System Based on Principal Lines Extracted at Several Under Skin Depths

摘要

In recent years, several palmprint recognition procedures that use 3-D information on the palm surface have been proposed in order to overcome some limitations of 2-D palmprint. However, 3-D optical images provide information on the external skin surface only. Ultrasound waves have the capability to penetrate into human tissue and, therefore, can provide new kinds of 3-D information from acquired palmprints. In this paper, a 3-D ultrasound palmprint recognition system that accounts for principal line depth is proposed. It exploits several 2-D palmprints, extracted at different under skin depths with a classic procedure, opportunely combined to achieve a 3-D template. An ad hoc matching criterion is then defined in order to account for line depth information when comparing 3-D templates. Experiments of both verification and identification, performed on a homemade database, have shown that the 3-D recognition procedure exhibits very good recognition rates. The proposed 3-D technique also benefits from other advantages of ultrasound, including not being sensitive to many kinds of surface contaminations and its capability to detect liveness during the acquisition phase, which makes it very difficult to counterfeit.