CONTACTLESS BIOMETRIC IDENTIFICATION USING 3D FINGER KNUCKLE PATTERNS

Abstract

Study on finger knuckle patterns has attracted increasing attention for the automated biometric identification. However, finger knuckle pattern is essentially a 3D biometric identifier and the usage or availability of only 2D finger knuckle databases in the literature is the key limitation to avail full potential from this biometric identifier. This paper therefore introduces (first) contactless 3D finger knuckle database in public domain, which is acquired from 130 different subjects in two-session imaging using photometric stereo approach. This paper investigates on the 3D information from the finger knuckle patterns and introduces a new feature descriptor to extract discriminative 3D features for more accurate 3D finger knuckle matching. An individuality model for the proposed feature descriptor is also presented. Comparative experimental results using the state-of-the-art feature extraction methods on this challenging 3D finger knuckle database validate the effectiveness of our approach. Although our feature descriptor is designed for 3D finger knuckle patterns, it is also attractive for other hand-based biometric identifiers with similar patterns such as the palmprint and fingerprint. This observation is validated from the outperforming results, using the state-of-the-art pixel-wise 3D palmprint and 3D fingerprint feature descriptors, on other publicly available datasets.  

Existing System

Subspace learning, spectral features

Proposed System

Multiple 2D finger knuckle images are firstly acquired under different illuminations and the acquisition is automatically synchronized using with respective illumination using a computer. The acquired images are then preprocessed and automatically segmented to extract region of interest images. These segmented images, acquired under different illumination illuminations, are then used for estimating surface nor mal vectors. Unlike other photometric stereo based biometric imaging system, the complex process of integrating surface normal vectors for recovering the depth images is not required in our system. The 3D finger knuckle features are then directly extracted from the surface normal vectors of 3D finger knuckle images. The 2D finger knuckle image, although noisy as each of them is acquired under partial illumination, used to recover 3D finger knuckle images can also be utilized to improve match accuracy for the system and is also investigated in our work. The m match scores between the probe and gallery pairs are then respectively computed for 3D and 2D finger knuckle images. The final decision to assign an unknown user to either genuine or impost imposter class is made using the combinate match score and its comparison with the decision threshold threshold.

Architecture

BLOCK DIAGRAM