Quality Induced Fingerprint Identification using Extended Feature Set

Automatic fingerprint identification systems use level-1 and level-2 features for fingerprint identification. However, forensic examiners utilize inherent level-3 details along with level-2 features. Existing level-3 feature extraction algorithms are computationally expensive to be used for identification. This paper presents a novel algorithm for fast level-3 feature extraction and identification. The algorithm starts with computing local image quality score using redundant discrete wavelet transform. A fast curve evolution algorithm is then used to extract four level-3 features namely, pores, ridge contours, dots, and incipient ridges. Along with level-1 and level-2 features, these level-3 features are used in a Delaunay triangulation based indexing algorithm. Finally, quality-based likelihood ratio is used to further improve the identification performance. Experiments conducted on a high resolution fingerprint database containing rolled, slap and latent images indicate that the algorithm offers significant benefits for fast fingerprint identification.