Segmentation techniques for overlapped latent fingerprint matching

Abstract

Image processing is the technique used for analyzing the images by converting them into a digital form so that analysis on the images can be done for some purpose. Forensic department uses the images for identifying the suspects based on their biometric information. The biometric information can be anything such as, fingerprints, iris or may be heart beat. Fingerprint is one of the largely and widely used biometric in major areas. The fingerprint information collected at the crime scene is useful for identifying the victim who have been committed the crime. In case of the rolled or plain fingerprint, it is easy for the analyst to find out the one who is found guilty. But it is not the case in latent. The latent fingerprints are noisy, blurred and smudgy. For which the technique known as Descriptor Based Hough Transform is used. By using the minutiae information the fingerprint can be matched with the rolled or plain fingerprints and identified after fingerprints are segmented from the background.