Third Level Identification with Hyperspectral Imaging in Fingerprints

Fingerprints are accepted as the key evidence in reaching the crime and the criminal. Identification from fingerprints is done at three levels. At the first level; fingerprint groups and subgroups, at the second level; fingerprint line characteristics, and at the third level, details such as pores, papilla structure, width, mold area and fractures. With the developing technology, algorithms are developed in which fingerprint direction and angle, combinations of placement between papillary characteristics and papillary flows are included in the evaluation process. In the study, 1-n queries were made in a dataset created from 800 fingerprints by manually and automatically extracted features with the blob detection algorithm, by randomly generating the latent fingerprints on which the pore details were displayed on the fingerprint line characteristics with hyperspectral imaging. In the fingerprint query, the latent fingerprints and the plain tenprints in the dataset were ranked using a score-based approach using the iterative closest point algorithm. According to the results obtained, it was observed that the score of the first candidate was higher and the scores of the candidates who came after it, were lower with the increase in the number of pores marked on the papilla lines and the accurate marking of these pores.