Abstract: A fingerprint is an impression left on the plain surface of an object by ridges of human finger. They are considered to be as sole keys that we can carry anywhere. In court of law, latent fingerprints serve as an important source of forensic evidence. In several applications, automatic matching of latent fingerprints to exemplar fingerprints is important. However, latent impressions are of poor quality, feature extraction and latent matching becomes a puzzling problem. This is due to difficult background noise in latent fingerprints. A top-down feedback from an exemplar is proposed which is used to refine the extracted feature from latents for enhancing accuracy. Ridge orientation and ridge frequency are the refined features from exemplars. Rolled and plain or slap fingerprints are collectively called exemplars. After feedback, the refined features are used to re-match the latent resulting top K candidate list. It is returned by the baseline matcher which is then resorted.
Keywords: Latent Fingerprint Matching, Exemplars, AFIS, Top-Down Feedback
Title: Latent Fingerprint Matching: Performance Gain via Top-Down Feedback
Author: Nadhiya Nazeer Khan, Smitha C Thomas
International Journal of Computer Science and Information Technology Research
ISSN 2348-1196 (print), ISSN 2348-120X (online)
Research Publish Journals
