Reference-Point Detection for Latent Fingerprint Images based on Spectrum Analysis

摘要

We propose a reference-point detection algorithm for latent fingerprint images based on spectrum analysis in spatial-frequency domain. The proposed method enhances fingerprint's spectra and uses these spectra to generate a spectral crossing field. This spectral crossing field can reveal a hidden reference-point by finding its maximum peak. By testing with the NIST27 latent fingerprint database, our experimental results show that the proposed method outperforms a recent singular point detection algorithm and a commercial off-the-shelf algorithm in terms of precision and identification accuracy.