Palmprint Extraction from 3D Ultrasound Images Collected Through Compounded Plane Waves

Multimodal biometric systems are becoming more and more popular because they allow to improve the overall performance of unimodal ones. Ultrasound systems have the peculiarity of collecting 3D image that contains information on different biometric characteristics. In the present work, a technique for collecting volumetric images of the palm based on an advanced ultrasound scanner (ULA OP 256) and compound plane waves is proposed and experimentally evaluated. Good quality images were achieved by transmitting 17 steered plane waves in the range ±3.15°. Image field depth was sufficiently high to extract both palmprint and vein patterns. Recognition capabilities are evaluated through palmprint verification and identification experiments on a homemade database composed of 216 samples from 47 users, obtaining very good results (EER=0.26% and identification rate=100%). It has also been experimentally verified that the proposed system allows to sensibly speed up the acquisition process without image quality loss. Further potentialities of the proposed approach are finally discussed.