Rectification of distortion in single rolled fingerprint

Recognizing low quality fingerprints is important while identifying a person in watch-list and fraudulent scams which is equally important to authenticate legitimate user. Flat or dab fingerprints and rolled fingerprints are being acquired from people by many biometric agencies across the world from their respective governments. But at times un-identification of legitimate users and fault recognition of fraudulent persons are the major issues in biometric identification and verification. Though fingerprint acquisition systems are working accurately, the issue arises during verification or authentication of a person if there is a distortion in the fingerprint (whether flat or rolled). Distortion may be due to intentional torque applied by the deceitful person or the acquisition procedure. To curb this issue, we are proposing novel approach, which is totally a combination of off-line, and online registration methods which includes old method of rolled fingerprints registration (card-ink based), a two phase approach. In the first phase clear rolled fingerprints are taken from the people through fingerprint acquisition to store them in database as minutiae files. In the second phase card-ink based rolled fingerprints which are likely distorted are taken through web-camera and processed by different image processing filters for distortion rectification. Then minutiae file of the rectified rolled fingerprint is going to be compared with the existing minutiae files in the database. The results obtained have been promising on the JNTUA Rolled Fingerprint Database.