Temporal analysis of fingerprint impressions

Fingerprint analysis has been performed on static images since the establishment of friction-ridge patterns as a reliable human trait for identification. The image of a fingerprint can vary considerably, however, as a result of such factors as skin elasticity and pressure on the imaging surface. With today's live-scan imaging devices, changes in appearance are notable as a finger presses against the scanning during acquisition. During the brief imaging period, fingerprint ridges stretch and compress, while some portions of the fingertip display varying levels of visibility and quality. Imaging systems select a single static image frame, thereby ignoring information that is absent from a particular image but present in others. This paper presents a novel approach that employs temporal analysis to a sequence of images captured during enrollment. Minutiae are extracted and tracked throughout the sequence to build a composite template that contains more genuine minutia than are found from simple static analysis. It was discovered, through temporal analysis, that 18.2% of minutiae changed type throughout a single impression sequence. The proposed method led to an increase in the number of minutiae detected by 25% and an improvement by 9.3% in the rank-1 identification rate of a set of partial prints, when using an imposter database of temporal composite templates over static templates.