An Efficient Approach for Latent Fingerprint Recognition

The online fingerprints by biometric system is not widely used now a days and there is less scope as user is friendly with the system. This paper represents a framework and applying the latent fingerprints obtained from the crime scene. These prints would be matched with our database and we identify the criminal. For this process we have to get the fingerprints of all the citizens. This technique may reduce the crime to a large extent. Latent prints are different from the patent prints. These fingerprints are found at the time of crime and these fingerprints are left accidentally. By this approach we collect these fingerprints by chemicals, powder, lasers and other physical means. Sometimes, fingerprints have a broken curve and it is not so clear due to low pressure. We apply the M_join algorithm to join the curve to achieve better results. Thus, our proposed approach eliminates the pseudo minutiae and joins the broken curves in fingerprints.