Latent fingerprint recognition: Challenges and advances

Abstract

Latent fingerprint recognition is desirable for forensics applications such as criminal investigation and forensic detection. Compared to traditional fingerprint recognition, automatic latent fingerprint recognition is more challenging because it must be able to deal with the smudgy, fragmentary and nonlinearly deformed fingerprints lifted from crime scenes and it has high liability for following legal procedures. In latent fingerprint recognition, latent fingerprint development draws little attention from pattern recognition due to that latent fingerprint development is traditionally regarded as a topic of forensic medical. In practice, latent fingerprint development is a pre-step of latent fingerprint recognition, whose result would cause dramatic influence to the performance of latent fingerprint recognition system. In this paper, we claim that latent fingerprint development is important to automatic latent fingerprint recognition system, and regard research of the relationship between latent fingerprint development and automatic latent fingerprint recognition system as a new topic. Moreover, we summarize the challenges, and provide critical analysis of the advances of automatic latent fingerprint recognition. © Springer International Publishing 2013.