An ultrasound technique for extracting 3D palmprints has been recently proposed. A drawback of that method relies in the small volume acquired, which is limited by the aperture of the array. To overcome this drawback, an improved method able to acquire a wider volume that includes the base of the fingers is presented in this work. The 3D image of the human palm is acquired by exploiting the ULtrasound Advanced Open Platform (ULA-OP) as ultrasound imaging system and a commercial high frequency linear array, which is moved by an automated scanning system based on a numeric controlled pantograph. The volume of interest is acquired through three linear parallel scans. The adjacent scans are partially overlapped to guarantee an effective continuity. To make the acquisition process very fast, the whole 3D image is acquired in a single record; the overlapped volumes are then realigned offline and fused using a convex combination, which weights follow a raised cosine law. Some experimental example achieved with the proposed technique are presented and their suitability to palmprint recognition process is highlighted and discussed. © 2014 IEEE.
Wide 3D ultrasound palmprint for biometric recognition
IULA, Antonio;
2014-01-01
Abstract
An ultrasound technique for extracting 3D palmprints has been recently proposed. A drawback of that method relies in the small volume acquired, which is limited by the aperture of the array. To overcome this drawback, an improved method able to acquire a wider volume that includes the base of the fingers is presented in this work. The 3D image of the human palm is acquired by exploiting the ULtrasound Advanced Open Platform (ULA-OP) as ultrasound imaging system and a commercial high frequency linear array, which is moved by an automated scanning system based on a numeric controlled pantograph. The volume of interest is acquired through three linear parallel scans. The adjacent scans are partially overlapped to guarantee an effective continuity. To make the acquisition process very fast, the whole 3D image is acquired in a single record; the overlapped volumes are then realigned offline and fused using a convex combination, which weights follow a raised cosine law. Some experimental example achieved with the proposed technique are presented and their suitability to palmprint recognition process is highlighted and discussed. © 2014 IEEE.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
