The Latent Fingerprint Analysis by “VeriFinger 12.0” Software

Abstract

Recently, fingerprint verification by experts combined with the automated fingerprint identification system (AFIS) became the standard method for fingerprint verification in Thailand. Office of Police Forensic Science has determined that the minutiae must match at least 10 points to confirm that the fingerprints belong to the same person. This study aims to assess the performance of the VeriFinger 12.0 software for detecting latent fingerprints with sufficient minutiae before sending it to the experts for verification by Mini AFIS. The study population consisted of 6 volunteers. The fingerprint samples were collected from all five fingers by stamping them on two materials, including smooth surface (glass) and rough surface (lacquered wood), and then collecting latent fingerprints with a dusting method. Then, the samples were analyzed with VeriFinger 12.0 and Mini AFIS software and calculated to compare the passive fingerprint verification performance with the standard method (Mini AFIS with experts). The results showed that VeriFinger 12.0 defined minutiae point matching the standard method accounting for 84.9 %, 15.1% of minutiae points that were not detected according to the standard method, and 49.4 % of minutiae were misconfigured or different from the traditional way. The correlation between VeriFinger 12.0 with experts versus the standard method and Mini AFIS, the Spearman correlation of % correct, % missed, and % false were 0.738 to 0.774, which were statistically significant (p<0.001). The results showed that VeriFinger 12.0 with experts versus the standard method and Mini AFIS provides coherent information. In conclusion, VeriFinger 12.0 software’s performance in latent fingerprint verification was similar to the standard method and Mini AFIS software.